The Legionella Lqs-LvbR Regulatory Network Controls Temperature-Dependent Growth Onset and Bacterial Cell Density

ABSTRACT

Legionella species are facultative intracellular pathogens that cause a life-threatening pneumonia termed Legionnaires’ disease. Legionella pneumophila employs the Lqs-LvbR (Legionella quorum sensing-Legionella virulence and biofilm regulator) network to regulate virulence and motility, but its role for growth in media is ill-defined. Here, we report that compared to the L. pneumophila reference strain JR32, a ΔlqsR mutant showed a reduced lag phase at 30°C and reached a higher cell density at 45°C, while the ΔlqsA, ΔlqsS, and ΔlqsT mutants showed a longer lag phase and reached a lower cell density. A ΔlvbR mutant resumed growth like the parental strain at 30°C but exhibited a substantially reduced cell density at 45°C. Thus, LvbR is an important cell density regulator at elevated temperatures. Environmental and clinical L. pneumophila strains grew in N-(2-acetamido)-2-aminoethanesulfonic acid (ACES)-buffered yeast extract (AYE) medium after distinct lag phases with similar rates at 30°C, reached different cell densities at the optimal growth temperature of 40°C, and no longer grew at 50°C. Legionella longbeachae reached a rather low cell density at 40°C and did not grow at and beyond 45°C. Genes encoding components of the Lqs-LvbR network were present in the genomes of the environmental and clinical L. pneumophila isolates, and upon growth at 30°C or 45°C, the P_lqsR, P_lqsA, P_lqsS, and P_lvbR promoters from strain JR32 were expressed in these strains with distinct patterns. Taken together, our results indicate that the Lqs-LvbR network governs the temperature-dependent growth onset and cell density of the L. pneumophila reference strain JR32 and possibly also of environmental and clinical L. pneumophila isolates.

IMPORTANCE Environmental bacteria of the genus Legionella are the causative agents of the severe pneumonia Legionnaires’ disease, the incidence of which is on the rise worldwide. Legionella pneumophila and Legionella longbeachae are the clinically most relevant species. The opportunistic pathogens are inhaled through contaminated aerosols and replicate in human lung macrophages with a mechanism similar to that in their natural hosts, free-living amoebae. Given their prevalence in natural and technical water systems, an efficient control of Legionella spp. by physical, chemical, or biological means will reduce the incidence of Legionnaires’ disease. Here, we show that the Legionella quorum sensing (Lqs) system and the pleiotropic transcription factor LvbR govern the temperature-dependent growth onset and cell density of bacterial cultures. Hence, the growth of L. pneumophila in water systems is determined not only by the temperature and nutrient availability but also by quorum sensing, i.e., density- and signaling molecule-dependent gene regulation.

INTRODUCTION

Environmental bacteria of the genus Legionella are the causative agents of Legionnaires’ disease (1, 2). This life-threatening pneumonia almost exclusively affects elderly people, can occur in epidemics of several hundreds of cases, and is globally on the rise. A total of 65 Legionella species have been identified to date (3), of which at least 24 have been linked to human disease (1). However, more than 90% of the Legionnaires’ disease cases diagnosed are caused by Legionella pneumophila, which together with Legionella longbeachae accounts for ca. 95% of all Legionella infections (1).

L. pneumophila parasitizes free-living amoebae and other protozoa (4–6), wherein it forms a unique replication-permissive compartment, the Legionella-containing vacuole (LCV) (7–12). L. pneumophila is a facultative intracellular bacterium that can also grow extracellularly in complex and defined media as well as in biofilms (13, 14). For intra- and extracellular growth, L. pneumophila requires micronutrients such as iron (15, 16). The analysis of the carbon and energy sources of L. pneumophila revealed that this fastidious, obligate aerobic bacterium metabolizes mainly amino acids (such as serine) but also uses carbohydrates and polysaccharides (reviewed in reference 17). Accordingly, L. pneumophila utilizes glucose (18, 19) and inositol (20), or glycerol (21) as well as glycerolipids (22), and possibly polysaccharides (glycogen, starch, cellulose, chitin) (23). In general, L. pneumophila employs a bipartite metabolism (24), since amino acids (e.g., serine) are preferentially used catabolically as energy supply and carbohydrates (e.g., glucose) or glycerol is utilized primarily for anabolic processes.

L. pneumophila employs a biphasic life cycle, where growth and virulence are reciprocally linked: growing bacteria are nonvirulent, while nongrowing bacteria are virulent, motile, and stress resistant (transmissive) (17, 25). The biphasic life cycle is controlled by the RNA-binding, posttranscriptional global regulator CsrA (carbon storage regulator A) by the two-component systems LetAS, PmrAB, and CpxRA (26), as well as by the Lqs (Legionella quorum sensing) system (27, 28). The Lqs system promotes bacterial density-dependent gene regulation and produces, detects, and responds to the low-molecular-weight signaling molecule LAI-1 (Legionella autoinducer-1, 3-hydroxypentadecane-4-one) (29, 30). The components of the Lqs system are encoded in a cluster (lqsA–lqsR–hdeD–lqsS) (31, 32) and at a distant site in the L. pneumophila genome (lqsT) (33), and they are all expressed from distinct promoters (34).

The constituents of the Lqs system comprise the pyridoxal-5-phosphate-dependent autoinducer synthase LqsA (35), the membrane-bound homologous sensor histidine kinases LqsS and LqsT (33, 35), and the prototypic response regulator LqsR (36) (Fig. 1). The latter contains a canonical phosphate receiver domain, dimerizes upon sensor kinase-mediated phosphorylation, and harbors an output domain structurally related to nucleotide-binding proteins (37, 38). Intriguingly, the Lqs system is linked to c-di-GMP signaling through the pleiotropic transcription factor LvbR (Legionella virulence and biofilm regulator), which negatively regulates hnox1 expression (39, 40) and itself is negatively regulated by LqsS (35, 39) (Fig. 1). The Hnox1-Lpg1057 system comprises the diguanylate cyclase inhibitor Hnox1 and the GGDEF/EAL domain-containing diguanylate cyclase Lpg1057 (41, 42).

FIG 1.

The Lqs-LvbR-cyclic di-GMP regulatory network. The environmental bacterium L. pneumophila employs a complex regulatory network to integrate and respond to extra- and intracellular signals. The Lqs (Legionella quorum sensing) system produces, detects, and responds to the endogenous signaling molecule LAI-1 (Legionella autoinducer-1; 3-hydroxypentadecane-4-one). The Lqs system comprises the autoinducer synthase LqsA, the homologous sensor kinases LqsS and LqsT, and the cognate response regulator LqsR. Lpg1057 is a diguanylate cyclase, which produces the intracellular second messenger cyclic di-GMP (c-di-GMP). Hnox1 is a receptor of nitric oxide (NO) and an inhibitor of the diguanylate cyclase Lpg1057. The transcription factor LvbR (Legionella virulence and biofilm regulator) negatively regulates hnox1 expression, and lvbR expression is negatively regulated by LqsS. Thus, LvbR links the Lqs system and c-di-GMP signaling. The Lqs-LvbR regulatory network controls bacterial virulence, motility, natural competence, expression of a genomic fitness island, planktonic growth, and biofilm architecture.

The Lqs-LvbR regulatory network controls pathogen-host cell interactions, bacterial motility, filament production, natural competence for DNA uptake, and biofilm architecture as well as the expression of a 133-kb genomic “fitness island” (39, 40, 43) (Fig. 1). The Lqs system and synthetic LAI-1 regulate not only flagellum production and motility of L. pneumophila (30) but also the migration of eukaryotic cells, and accordingly, the bacterial signaling compound promotes interkingdom signaling (44). Recent studies revealed that the Lqs system also determines phenotypic heterogeneity and the occurrence of functionally different L. pneumophila subpopulations in infected amoebae and under sessile/biofilm conditions (45–47). In this work, we show that the Lqs-LvbR system governs the temperature-dependent growth onset and cell density of an L. pneumophila reference strain and possibly also of environmental and clinical L. pneumophila isolates.

RESULTS

The Lqs-LvbR network regulates L. pneumophila growth and density in AYE medium.

The L. pneumophila Lqs-LvbR regulatory network controls a number of bacterial traits, including virulence and motility, natural competence, and biofilm architecture (40), but its role for planktonic bacterial growth and temperature dependency of replication has not been studied in detail. To address this question, we compared growth of the parental strain JR32 with the ΔlqsR, ΔlqsA, ΔlqsS, ΔlqsT, ΔlqsS-ΔlqsT, or ΔlvbR mutant strains in N-(2-acetamido)-2-aminoethanesulfonic acid (ACES)-buffered yeast extract (AYE) medium at different temperatures. We chose a set of temperatures (30°C, 40°C, and 45°C) that flanks the laboratory standard (and optimal) growth temperature of 37°C to define the extent of temperature dependency.

At 30°C, all strains grew with similar maximal rates, but their lag phases varied greatly (Fig. 2A). While the ΔlqsS-ΔlqsT and ΔlvbR mutant strains showed the same lag phase as the parental strain JR32, the ΔlqsR mutant exhibited a shortened lag phase, and the ΔlqsA, ΔlqsS, and ΔlqsT mutants showed a prolonged lag phase. The growth initiation phenotype of ΔlqsR (shortened lag phase) was fully complemented (Fig. S1A) (36), while the phenotype of ΔlqsA (prolonged lag phase) was partially reverted by expressing plasmid-borne lqsA (Fig. S1A). Thus, the lag phase length is regulated by these genes. All lqs mutant strains as well as the lvbR mutant used have been complemented previously for other phenotypes, including defects in uptake by phagocytes, intracellular growth, and cytotoxicity, as well as hypercompetence for DNA uptake (33, 35, 36, 39).

FIG 2.

The Lqs-LvbR network regulates L. pneumophila growth and density in AYE medium and MDM. (A to C) L. pneumophila JR32, ΔlvbR, ΔlqsR, ΔlqsA, ΔlqsS, ΔlqsT, or ΔlqsS-ΔlqsT mutant strains were grown in AYE medium for 23 to 24 h at 37°C and inoculated in AYE medium at an initial OD₆₀₀ of 0.1. Growth of strains at (A) 30°C, (B) 40°C, or (C) 45°C was monitored over time by measuring OD₆₀₀ with a microplate reader. (D to F) L. pneumophila JR32, ΔlvbR, ΔlqsR, ΔlqsA, ΔlqsS, ΔlqsT, or ΔlqsS-ΔlqsT mutant strains were grown in MDM for 28 to 29 h at 37°C and inoculated in MDM at an initial OD₆₀₀ of 0.1. Growth of strains at (D) 30°C, (E) 40°C, or (F) 45°C was monitored over time by measuring OD₆₀₀ with a microplate reader. Growth curves shown are means and standard deviations of biological triplicates and representative of at least four independent experiments.

Upon growth at 40°C, the strains showed essentially the same pattern as at 30°C, except that the differences between the lag phase lengths were smaller (Fig. 2B). Furthermore, all strains exhibited a biphasic growth pattern. Interestingly, however, at 45°C, some strains grew with different rates and reached vastly different final cell densities (optical density at 600 nm [OD₆₀₀] of ca. 0.2 to 0.8) (Fig. 2C). At 45°C, the ΔlqsR, ΔlqsA, ΔlqsS, ΔlqsT, and ΔlqsS-ΔlqsT mutant strains grew with a doubling time similar to that of the parental strain JR32 (ca. 2.1 to 2.2 h), while the ΔlvbR mutant showed a longer doubling time (ca. 2.6 h) and, therefore, grew significantly more slowly than the parental strain (Fig. S2). Moreover, at 45°C, the ΔlqsR and ΔlqsS-ΔlqsT mutant strains reached a higher cell density than JR32, and the ΔlqsA, ΔlqsS, and ΔlqsT mutants reached a lower cell density. Strikingly, the final cell density of the ΔlvbR mutant was less than one-third of the parental strain JR32 (Fig. 2C). The pronounced cell density phenotype of the ΔlvbR mutant strain was complemented to the full extent by the corresponding plasmid-borne gene and, thus, is caused by the absence of this gene (Fig. S1B). In summary, compared to the parental strain JR32, the lqs and the ΔlvbR mutant strains showed different lag phase lengths upon growth initiation in AYE medium at 30°C, and they reached different final cell densities at 45°C. In particular, the transcription factor LvbR and the autoinducer synthase LqsA positively regulated the cell density of L. pneumophila at elevated temperatures.

The Lqs-LvbR network regulates L. pneumophila growth and density in minimal defined medium.

Next, we assessed the role of the Lqs-LvbR regulatory network for growth in minimal defined medium (MDM) (21). To this end, we compared growth of the parental strain JR32 with the ΔlqsR, ΔlqsA, ΔlqsS, ΔlqsT, ΔlqsS-ΔlqsT, or ΔlvbR mutant strains in MDM at 30°C, 40°C, and 45°C. At 30°C, all strains grew with similar maximal rates, but—similar to growth in AYE medium—the lag phases of the ΔlqsA, ΔlqsS, and ΔlqsT mutants were prolonged (Fig. 2D). Hence, the differences in the lag phases of the mutant strains are observed in the nutrient-rich and complex AYE medium, as well as in the nutrient-poorer and defined MDM. The ΔlqsR and ΔlqsS-ΔlqsT as well as the ΔlvbR mutant strains showed the same lag phase as the parental strain JR32.

Upon growth at 40°C, the strains showed a similar pattern, but the lag phase of only the ΔlqsA and ΔlqsS mutants was prolonged (Fig. 2E). Moreover, at this temperature, the final cell density of the ΔlqsA, ΔlqsS, ΔlqsT, and ΔlqsS-ΔlqsT mutants was increased, while the ΔlqsR and ΔlvbR mutants grew to a density similar to that of strain JR32. At 45°C, growth of all strains was dramatically reduced, if not abolished (Fig. 2F), suggesting that the strains cope less successfully with higher temperatures under nutrient-poor conditions than in AYE medium. Taken together, in MDM, the ΔlqsA, ΔlqsS, and ΔlqsT mutants initiated growth later than the parental strain JR32 at 30°C and—like the other lqs mutants as well as the parental strain—did not grow at all at 45°C. The differences in the lag phases of the mutant strains were observed in the nutrient-rich AYE medium, as well as in the nutrient-poorer MDM, and accordingly, this trait seems to be independent of the nutrients available.

Temperature-dependent growth of environmental and clinical Legionella isolates.

With the aim to assess the detailed growth characteristics and a role for the Lqs-LvbR system for environmental and clinical Legionella strains, we initially started out with a total of 19 isolates, comprising 15 L. pneumophila strains (serogroups [sg] 1, 5, and 6), 2 L. longbeachae strains, and 2 L. micdadei strains (Fig. S3). Growth of these Legionella strains was quantitatively assessed by measuring the optical density at 600 nm (OD₆₀₀) in AYE medium at 37°C with a 96-well microplate reader. Using this approach, we observed the following robust patterns. (i) All L. pneumophila strains grew with a similar doubling time (ca. 2.3 to 3.2 h) with the exception of the more slowly growing sg 6 clinical strain 526 (4.7 h), but the lag phase of the strains was different from and, in some cases, shorter than that of the sg 1 reference strain JR32. (ii) The doubling time of the L. longbeachae strains (2.3 h) was similar to that of the L. pneumophila strains, but the final density was lower. (iii) The doubling time of the L. micdadei strains (3.3 to 3.6 h) was considerably higher than that of the L. pneumophila or L. longbeachae strains, and their final density was 3 to 4 times lower.

Based on the observed growth patterns, we picked representative strains for a detailed analysis of growth characteristics, i.e., the L. pneumophila environmental isolates 500 and 529 (both sg 1) and 525 (sg 6) and the clinical isolate 509 (sg 1), as well as the clinical reference strains L. pneumophila JR32 (sg 1) and L. longbeachae NSW150. Like L. pneumophila, L. micdadei harbors the Lqs system (48); yet, due to the poor growth of the L. micdadei strains in AYE medium, this species was not studied further. L. longbeachae lacks the Lqs system (49), and thus, we included this species for comparison. The selected Legionella strains grew with a larger doubling time and to a lower density in diluted AYE medium at 37°C (Fig. S4), indicating that the growth characteristics indeed depend on nutrient availability.

Following the growth rate of the selected L. pneumophila and L. longbeachae strains in AYE medium over a temperature range of 18°C to 50°C revealed that the growth optimum for all strains is approximately 40°C, except for strain 529, which grew optimally at 45°C (Fig. 3A). L. pneumophila strain 509 and L. longbeachae NSW150 were the most temperature sensitive, and at or beyond 50°C (L. pneumophila) or 45°C (L. longbeachae) the strains did not grow anymore. At the same time, the doubling times decreased from ca. 15 to 20 h at 18°C to ca. 2 h at 40°C (Fig. S5A). Intriguingly, the growth rate curve was not symmetrical: while the growth rates increased linearly in the temperature range of 18 to 40°C, the growth rate sharply dropped to zero between 40 and 50°C (Fig. 3A).

FIG 3.

Temperature-dependent growth of environmental and clinical Legionella isolates. L. pneumophila strains JR32, 500, 509, 525, and 529 and L. longbeachae strain NSW150 were diluted in (A) AYE medium or (B) MDM to an initial OD₆₀₀ of 0.2 and grown at the temperatures indicated (18 to 55°C). Bacterial growth was determined over time by measuring OD₆₀₀ using a microplate reader. From the resulting growth curves, the growth rates were calculated using a nonlinear regression model, considering only exponential growth phases defined in semilogarithmic plots. Data shown are means and standard deviation of growth rates from three independent experiments performed in biological triplicates each.

The growth rates of the L. pneumophila strains were ca. 3-fold lower at a given temperature in MDM, but the temperature range was similar to that of growth in AYE medium (Fig. 3B). Under these conditions, the doubling times decreased from ca. 38 to 52 h at 18°C to ca. 5 to 8 h at 40°C for most strains (Fig. S5B). L. pneumophila strain 525 (sg 6) grew exceptionally slowly in MDM with a doubling time of ca. 210 h at 18°C and ca. 25 h at 40°C, and L. longbeachae NSW150 did not grow at all. Taken together, the doubling times of selected L. pneumophila and L. longbeachae strains decreased ca. 5- to 10-fold between growth at 18°C and the optimum of 40°C, and the doubling times of the L. pneumophila strains were ca. 3-fold shorter upon growth in AYE medium compared to those upon growth in MDM.

Growth onset and cell density of environmental and clinical Legionella strains in AYE medium.

To compare growth of the selected environmental and clinical Legionella strains to that of the reference strain JR32 and isogenic lqs-lvbR mutants (Fig. 2), we further assessed in detail the growth characteristics at 30°C, 40°C, or 45°C (Fig. 4). Upon growth in AYE medium at 30°C, the doubling time of the 5 L. pneumophila strains was similar (ca. 4.0 to 4.6 h; Table 1), but the lag phase length was different (Fig. 4A), similar to the initially observed pattern (Fig. S3). While the strains 500, 525, and 529 showed a lag phase shorter than that of strain JR32, the lag phase of strain 509 was longer.

FIG 4.

Growth onset and cell density of environmental and clinical Legionella strains in AYE medium and MDM. (A to C) L. pneumophila strains JR32, 500, 509, 525, 529 and L. longbeachae strain NSW150 were grown in AYE medium for 23 to 24 h at 37°C and inoculated in AYE medium at an initial OD₆₀₀ of 0.2. Growth of Legionella strains at (A) 30°C, (B) 40°C, or (C) 45°C was monitored over time by measuring OD₆₀₀ with a microplate reader. (D to F) L. pneumophila strains JR32, 500, 509, 525, and 529 and L. longbeachae strain NSW150 were grown in minimal defined medium (MDM) for 28 to 29 h at 37°C and inoculated in MDM at an initial OD₆₀₀ of 0.2. Growth of strains at (D) 30°C, (E) 40°C, or (F) 45°C was monitored over time by measuring OD₆₀₀ with a microplate reader. Growth curves shown are means and standard deviations of biological triplicates and representative of six independent experiments.

TABLE 1.

Doubling times of environmental and clinical Legionella isolates in AYE medium or MDM

Strain	td 30°Ca	td 40°Ca	td 45°Ca
AYE medium
JR32	4.06 ± 0.03	2.22 ± 0.02	2.51 ± 0.04
500	3.99 ± 0.03	2.72 ± 0.08	3.05 ± 0.10
509	4.28 ± 0.02	2.41 ± 0.01	3.73 ± 0.37
525	4.58 ± 0.05	2.44 ± 0.03	2.62 ± 0.01
529	4.45 ± 0.03	2.33 ± 0.01	2.00 ± 0.02
NSW150	2.86 ± 0.06	2.16 ± 0.09	NG
MDM
JR32	7.93 ± 0.09	5.00 ± 0.04	7.32 ± 0.21
500	10.29 ± 0.08	8.61 ± 0.12	15.57 ± 0.70
509	9.08 ± 0.16	7.31 ± 0.02	19.40 ± 0.73
525	31.61 ± 0.44	25.08 ± 0.31	ND
529	8.86 ± 0.10	7.27 ± 0.02	9.70 ± 0.39
NSW150	NG	NG	NG

^aDoubling times (t_d) are indicated in hours. Data shown are means and (±) standard deviation from doubling times derived from three independent cultures. NG, no growth observed.

Upon growth at 40°C, the doubling time of the 5 L. pneumophila strains was reduced compared to that of growth at 30°C (ca. 2.2 to 2.7 h; Table 1), the lag phase length differences were smaller, and the final cell density of L. pneumophila strain 500 was reduced (Fig. 4B). At 45°C, the L. pneumophila strains 525, 529, and JR32 grew fastest with doubling times of 2.0 to 2.6 h (Table 1), while the L. pneumophila strains 500 and 509 grew more slowly and to a much lower density (Fig. 4C). Intriguingly, the L. pneumophila strains showed a biphasic growth pattern, in particular at 40°C and to a somewhat lesser extent at 45°C. Finally, L. longbeachae NSW150 grew similarly to L. pneumophila at 30°C, less densely at 40°C, or not at all at 45°C (Fig. 4A to C, Table 1). Taken together, environmental and clinical L. pneumophila and L. longbeachae strains grew similarly in AYE medium after distinct lag phases at 30°C, and the strains reached different cell densities at the optimal temperature of 40°C and in particular at an elevated temperature of 45°C.

Growth onset and cell density of environmental and clinical Legionella strains in minimal defined medium.

Next, we assessed the growth characteristics of the 6 selected Legionella strains in MDM at 30°C, 40°C, or 45°C. Upon growth in MDM at 30°C the L. pneumophila strains JR32, 509, and 529 grew with the shortest doubling time (7.9 to 9.1 h; Table 1), and strain JR32 reached the highest density (Fig. 4D). The L. pneumophila strains 500 (sg 1) and 525 (sg 6) grew much less densely, and L. longbeachae NSW150 did not grow at all. Compared to the growth patterns observed in the nutrient-rich AYE medium, the L. pneumophila strain 509 grew with a comparable pattern in MDM (log phase elongated compared to JR32), while the other environmental and clinical strains showed a considerably different growth pattern and also grew less densely.

Upon growth at 40°C, the growth characteristics of the strains were largely the same as those at 30°C (Fig. 4E), but the L. pneumophila strains JR32, 509, and 529 grew more quickly with doubling times of 5.0 to 7.3 h (Table 1). At 45°C, the L. pneumophila strains JR32 and 529 grew with doubling times similar to what was observed at 30°C, and the other strains grew barely or not at all (Fig. 4F). However, under these conditions, the final cell density of the strains JR32 and 529 was considerably lower than in AYE medium, suggesting that the strains cope less successfully with higher temperatures under nutrient-poor conditions. Similar to growth in AYE medium, many L. pneumophila strains also showed a bi- if not triphasic growth in MDM, in particular the strains JR32, 500, 509, and 529 at 40°C. Taken together, the growth characteristics of the L. pneumophila sg 1 strains JR32, 500, 509, and 529 in MDM were qualitatively similar to those in AYE medium, but the final cell densities were much lower. L. pneumophila 525 (sg 6) barely grew, and L. longbeachae NSW150 did not grow at all, suggesting that these strains were lacking essential nutrients in the minimal medium.

Presence of lqs and lvbR genes in environmental and clinical Legionella isolates.

In order to possibly implicate the lqs and lvbR genes in the growth characteristics of the selected environmental and clinical Legionella strains, we first assessed the presence of the genes in these strains. To this end, the regions comprising P_lqsR-lqsR, P_lqsA-lqsA, P_lqsS-lqsS, or P_lvbR-lvbR were amplified by PCR in the L. pneumophila strains JR32 (positive control), 500, 509, 525, and 529 and the L. longbeachae strain NSW150 (negative control) (Fig. 5A). Using genomic DNA of these strains as a template, fragments comprising P_lqsR-lqsR (1,727 bp), P_lqsA-lqsA (1,936 bp), and P_lqsS-lqsS (1,984 bp) were amplified from all L. pneumophila strains tested but not from L. longbeachae NSW150, in agreement with the absence of the lqs genes in the genome of the latter (49).

FIG 5.

Presence of lqsR, lqsA, lqsS, lvbR, and hnox1 gene regions in environmental and clinical Legionella isolates. The presence of (A) P_lqsR-lqsR, P_lqsA-lqsA, or P_lqsS-lqsS or (B) P_lvbR-lvbR, lvbR_1-201, LvbR_BS, or P_hnox1-hnox1 in the genomes of L. pneumophila strains JR32 (positive control), 500, 509, 525, and 529 or L. longbeachae strain NSW150 (negative control) was assessed by PCR amplification of the corresponding genes (and putative promoters) and/or intergenic regions indicated. PCR products were analyzed by agarose gel electrophoresis and visualized with MaestroSafe staining: P_lqsR-lqsR (1,727 bp), P_lqsA-lqsA (1,936 bp), P_lqsS-lqsS (1,984 bp), P_lvbR-lvbR (948 bp), lvbR (264 bp), lvbR_1-201 (201 bp), LvbR_BS (347 bp), P_hnox1-hnox1 (1,344 bp), and LvbR_BS-hnox1 (969 bp). Schemes depicting the lqs or the lvbR-hnox1 genomic region and the PCR fragments amplified are shown below the gels.

The fragments P_lvbR-lvbR (948 bp), lvbR (264 bp), and P_hnox1-hnox1 (1,344 bp) were amplified robustly in the L. pneumophila strains JR32, 500, 509, and 529 but only very weakly in L. pneumophila 525 (lvbR; 264 bp) and not at all in L. longbeachae NSW150 (Fig. 5B). However, the fragments lvbR_1-201 (201 bp), LvbR_BS (LvbR binding site [39]; 347 bp), and LvbR_BS-hnox1 (969 bp) were robustly amplified in all L. pneumophila strains tested, including strain 525, indicating that the lvbR as well as the lvbR-hnox1 intergenic region and hnox1 are indeed present in the genome of the latter strain. Strain 525 harbors an LvbR_BS region, which is shorter than that in the other strains tested, and an lvbR gene, which appears to be a more distantly related homologue. Sequencing of the LvbR_BS intergenic region and part of the lvbR and hnox1 genes of the strains JR32, 500, 509, 525, and 529 revealed the presence of almost identical lvbR and hnox1 gene fragments and an LvbR_BS intergenic region of strain 525, which is indeed 80 nucleotides shorter (Fig. S6). L. longbeachae NSW150 likely does not harbor an lvbR homologue, and the most closely related protein is a putative YdnC-like transcription factor (29% identity of 55/178 amino acids; E value 7 × e⁻⁴). Taken together, these results revealed that the lqsR, lqsA, lqsS, and lvbR genes are present in all L. pneumophila strains analyzed in detail in this study.

Regulation of lqs and lvbR genes in environmental and clinical Legionella isolates.

Next, we tested the regulation of the lqs and lvbR genes in environmental and clinical L. pneumophila isolates. The genome sequences for these strains are not available, and therefore, we used the corresponding promoters of strain JR32. Accordingly, we transformed the strains JR32, 500, 509, 525, and 529 with reporter constructs comprising transcriptional fusions of different promoters from strains JR32 and gfp (P_lqsR-gfp, P_lqsA-gfp, P_lqsS-gfp, P_lvbR-gfp, or P_hnox1-gfp) and quantified GFP production by fluorescence upon growth at 30°C (Fig. 6). This approach revealed that all environmental and clinical L. pneumophila isolates tested expressed P_lqsR-gfp (Fig. 6A), P_lqsA-gfp (Fig. 6B), P_lqsS-gfp (Fig. 6C), P_lvbR-gfp (Fig. 6D), and P_hnox1-gfp (Fig. 6E). The expression patterns of P_lqsR, P_lqsS, and P_lvbR, as well as of P_lqsA and P_hnox1, were similar in the environmental and clinical L. pneumophila isolates. Interestingly, some environmental and clinical L. pneumophila isolates expressed P_lvbR with high intensity, in contrast to the parental strain JR32, which barely expressed P_lvbR (Fig. 6D), as observed previously (39). All environmental and clinical strains harbor lqsS (Fig. 5A) and presumably produce LqsS, the negative regulator of LvbR in strain JR32 (35, 39). Therefore, another mechanism likely accounts for the observation that these strains indeed express P_lvbR.

FIG 6.

Regulation of lqsR, lqsA, lqsS, lvbR, and hnox1 promoters in environmental and clinical Legionella isolates. L. pneumophila strains JR32, 500, 509, 525, and 529 harboring (A) P_lqsR-gfp (pRH037), (B) P_lqsA-gfp (pRH038), (C) P_lqsS-gfp (pRH051), (D) P_lvbR-gfp (pRH023), or (E) P_hnox1-gfp (pRH026) reporter constructs (promoters of strain JR32) were grown in AYE medium for 23 to 24 h at 37°C and inoculated in AYE medium at an initial OD₆₀₀ of 0.2. Strains were incubated at 30°C, and GFP fluorescence was measured over time (relative fluorescence units, RFU) using a microplate reader. Data shown are means and standard deviations of biological triplicates and representative of three independent experiments.

The expression of P_lqsR, P_lqsS, and P_lvbR in the environmental and clinical L. pneumophila isolates peaked at different times (Fig. 6), which precisely correspond to the different lag phase lengths of the strains and, accordingly, to their transition from the exponential to the stationary-growth phase (Fig. 7). On the other hand, P_lqsA and P_hnox1 showed a similar but more complex expression profile, which was not linked exclusively to the transition from the exponential to the stationary-growth phase (Fig. 7).

FIG 7.

Growth of environmental and clinical L. pneumophila isolates harboring promoter reporters. L. pneumophila strains JR32, 500, 509, 525, and 529 containing (A) P_lqsR-gfp (pRH037), (B) P_lqsA-gfp (pRH038), (C) P_lqsS-gfp (pRH051), (D) P_lvbR-gfp (pRH023), or (E) P_hnox1-gfp (pRH026) reporter constructs (promoters of strain JR32) were grown in AYE medium for 23 to 24 h at 37°C and inoculated in AYE medium at an initial OD₆₀₀ of 0.2. Bacterial growth at 30°C was followed over time by measuring OD₆₀₀ with a microplate reader. Data shown are means and standard deviations of biological triplicates and representative of three independent experiments.

The L. pneumophila strains JR32, 500, 509, 525, and 529 harboring the empty control plasmid pTS10 showed very low fluorescence upon growth at 30°C (Fig. S7A). These strains showed a growth profile similar to that of the strains harboring no plasmid (Fig. 4A) or, with the exception of strain 525 (sg 6), harboring a promoter reporter plasmid (Fig. 7A to E). Strain 525 harboring any of the promoter reporter plasmids showed a prolonged lag phase. Thus, the promoter reporter plasmids do not affect the growth of most of the environmental and clinical strains tested but delay the growth of the environmental strain 525.

The strains JR32, 500, 509, 525, and 529 harboring the P_lqsR-gfp, P_lqsA-gfp, P_lqsS-gfp, or P_lvbR-gfp reporter constructs were also tested for green fluorescent protein (GFP) fluorescence upon growth at 45°C (Fig. S8). Under these conditions, all environmental and clinical L. pneumophila isolates robustly expressed P_lqsR-gfp (Fig. S8A), P_lqsA-gfp (Fig. S8B), or P_lvbR-gfp (Fig. S8D), while P_lqsS-gfp showed a lower fluorescence (Fig. S8C), which was in the range of the control plasmid (pTS10) (Fig. S7B). Notably, the environmental strains 529 and 525 showed the highest expression of P_lqsR-gfp and P_lvbR-gfp, or P_lqsA-gfp, respectively. The parental strain JR32 barely expressed P_lvbR upon growth at 45°C (Fig. S8D), as observed upon growth at 30°C (Fig. 6D) and reported previously (39). In general, at 45°C, promoter expression reached the highest levels in stationary-growth phase and remained high (Fig. S8). This is in contrast to what was observed at 30°C, where the expression of P_lqsR-gfp, P_lqsS-gfp, and P_lvbR-gfp peaked sharply at the onset of the stationary phase (Fig. 6). At 45°C, all strains harboring a reporter construct grew rapidly and similarly, with the exception of the clinical strain 509, which reached lower cell densities (Fig. S8), and the growth pattern of all strains was similar to that of the strains harboring the control plasmid (Fig. S7B) or no plasmid (Fig. 4C).

In summary, upon growth at 30°C or 45°C, all environmental and clinical L. pneumophila isolates tested expressed P_lqsR, P_lqsA, P_lqsS, and P_lvbR of strain JR32. At 30°C, P_lqsR, P_lqsS, and P_lvbR were expressed with a single peak in the postexponential growth phase, and at 45°C the environmental strains 529 and 525 showed the highest expression of P_lqsR-gfp and P_lvbR-gfp, or P_lqsA-gfp, respectively. These results suggest that the Lqs-LvbR network is active in the environmental and clinical L. pneumophila strains and that components of the network are regulated in a temperature-dependent manner.

DISCUSSION

In this study, we analyzed in detail the growth characteristics and the role of the Lqs-LvbR network for temperature-dependent growth rate and cell density of an L. pneumophila reference strain and mutant strains, as well as those of selected environmental and clinical L. pneumophila and L. longbeachae isolates. For the L. pneumophila reference strain, we implicated components of the Lqs system in growth onset, doubling time, and cell density (Fig. 2). Hence, the Lqs system regulates not only L. pneumophila virulence, motility, competence, and a genomic fitness island but also growth characteristics of planktonic bacteria (Fig. 1). Compared to the parental strain JR32, a ΔlqsR mutant strain showed a reduced lag phase at 30°C in AYE medium and reached a higher cell density at 45°C (Fig. 2). The ΔlqsA, ΔlqsS, and ΔlqsT mutant strains exhibited a longer lag phase and reached a lower cell density. Intriguingly, the components of the Lqs system regulated growth resumption of sessile L. pneumophila in a similar manner: compared to the parental strain JR32, the percentage of ΔlqsR mutant bacteria resuming growth was higher, while the percentage of ΔlqsA, ΔlqsS, or ΔlqsT mutants was lower (46).

In addition to the Lqs system, we found that the pleiotropic transcription factor LvbR played a major role for doubling time and cell density of L. pneumophila (Fig. 2). In AYE medium, the ΔlvbR mutant strain resumed growth like the parental strain at 30°C (Fig. 2A) but exhibited a dramatically reduced cell density at 45°C (Fig. 2C). In MDM, the ΔlvbR mutant strain behaved similarly to the parental strain JR32 (Fig. 2D to F). Hence, preferentially under nutrient-rich conditions and at elevated temperatures, the LvbR transcription factor functions as an important regulator of L. pneumophila cell density.

Of note, growth resumption and final cell density of the lqs and lvbR mutant strains are not correlated with their virulence and transmission phenotypes. Both the ΔlqsR and the ΔlvbR mutant strains are severely impaired for virulence (36, 39), yet growth resumption characteristics and final cell density at the temperatures tested were different (Fig. 2). Inversely, the ΔlqsA mutant does not show a severe virulence phenotype compared to the parental strain JR32 (35), yet growth resumption and final cell density of the two strains were distinct (Fig. 2). Finally, the ΔlqsS, ΔlqsT, and ΔlqsS-ΔlqsT double mutant strains are all impaired for virulence (33, 35), yet the individual sensor kinase mutants reached a lower cell density than the parental strain at 45°C, while the ΔlqsS-ΔlqsT double mutant reached a higher cell density. A possible explanation for the latter observation is that LqsS and LqsT regulate many genes in a complex and reciprocal manner (33), and hence, the ΔlqsS-ΔlqsT double mutant might cancel out or revert phenotypes of the single mutants. In any case, the effects of L. pneumophila Lqs-LvbR components on planktonic growth resumption and final cell density are apparently not directly linked to the virulence phenotypes of the corresponding mutant strains.

The Lqs system and the α-hydroxyketone autoinducer LAI-1 are not strictly conserved among Legionella species, but α-hydroxyketone-mediated signaling and cell-cell communication seem to be widely distributed among environmental bacteria (27). A recent bioinformatics analysis for the presence of the lqs cluster in 58 Legionella species genomes revealed three categories of species (48): 19 harbored a complete lqs cluster, 20 did not possess lqsA but maintained the receptor lqsR and/or the sensor kinase lqsS, and 19 did not have any of the lqs genes. These results are in agreement with the notion that the Lqs system functions in intra- as well as interspecies communication. In agreement with a broad distribution of the Lqs-LvbR network, the components are present not only in the reference strain JR32 but also in the genomes of the environmental and clinical L. pneumophila strains analyzed in detail here (Fig. 5).

The P_lqsR, P_lqsA, and P_lqsS promoters from strain JR32 were expressed in the environmental and clinical L. pneumophila strains upon growth at 30°C (Fig. 6) and 45°C (Fig. S8). Interestingly, the environmental isolates 500 and 529 (both sg 1) expressed P_lqsR to an even higher degree than JR32 (4 to 5 times higher GFP fluorescence intensity). P_lvbR was also highly expressed in these strains but barely expressed in JR32 (Fig. 6 and Fig. S8), confirming previous findings (39). In strain JR32, LvbR is negatively regulated by LqsS (35, 39) (Fig. 6D). All environmental and clinical strains tested harbored lqsS (Fig. 5A) and lvbR (Fig. 5B, Fig. S6), and they expressed P_lqsS and P_lvbR from strain JR32 (Fig. 6CD). Accordingly, lvbR does not seem to be negatively regulated by lqsS in these strains. While the local lvbR-hnox1 region is similar in strain JR32 and the environmental and clinical strains (Fig. S6), other regulatory processes might be different. In strain JR32, lvbR localizes to a high-plasticity 133- kb genomic “fitness island” harboring 125 genes, of which 52 are negatively regulated by lqsS (35). Accordingly, the fitness island as a whole might be absent or differently regulated in the environmental and clinical strains, thus accounting for the different regulatory patterns observed.

In addition to the Lqs-LvbR network, other regulatory systems are implicated in the survival of L. pneumophila in aqueous environments, such as the “stringent response” and the alternative sigma factor RpoS (σ^S/σ³⁸) (50), as well as the two-component system (TCS) LetAS (51). The stringent response is triggered upon production of the second messenger guanosine 3′,5′-bispyrophosphate (ppGpp) by the RelA and SpoT synthases (52–54), which in turn are activated by amino acid starvation (55) and inhibition of fatty acid biosynthesis (54). In response to ppGpp, the stationary-phase sigma factor RpoS controls the expression of L. pneumophila virulence and transmission (52, 56), and the LetAS TCS also regulates these traits (28, 57–59). RpoS and, to a lesser extent, LetA control the production of LqsR in stationary-growth phase (36), and thus, the stringent response is linked to the Lqs system. For planktonic growth of L. pneumophila in broth, minimal media, and aqueous environments, this implicates that nutrient starvation likely upregulates the Lqs system and, hence, promotes density-dependent regulation.

The LetAS TCS has been implicated in the regulation of virulence and the switch from the replicative to the transmissive phase (28, 57–59). LetA also controls the production of LqsR (36), and therefore, this pathway might account for the role of some of the lqs genes in regulating the lag phase length upon growth at 30°C (Fig. 2A). CsrA, an RNA-binding global repressor of transmission traits, regulates LqsR by directly binding its mRNA (60), thereby revealing a direct link between the production of LqsR and the growth phase switch of L. pneumophila. Intriguingly, the peak expression of P_lqsR, P_lqsS, and P_lvbR in the environmental and clinical L. pneumophila strains upon growth at 30°C (Fig. 6) correlated with the transition from the exponential- to the stationary-growth phase (Fig. 7). Hence, the growth phase switch regulated by the LetAS TCS and CsrA might also control the expression of lqsR, lqsS, and lvbR in the environmental and clinical L. pneumophila strains.

The ΔlvbR mutant reached only a low final cell density in AYE medium upon growth at 45°C (Fig. 2C), but the expression of P_lvbR (and P_lqsR or P_lqsA) did not strictly correlate to the final cell density of the environmental and clinical L. pneumophila isolates (Fig. S8). How LvbR and the Lqs system determine the differences in the final cell density is not well understood. The Lqs-LvbR network is a pleiotropic regulatory network affecting the expression of hundreds of genes and a plethora of bacterial traits (39, 40). Comparative transcriptomics revealed possible mechanistic clues regarding the Lqs-LvbR-dependent cell density regulation. In the ΔlqsR as well as the ΔlvbR mutant strains, a number of metabolic genes are upregulated. Examples include lpg0063/aroH-lpg0068 (aromatic amino acid biosynthesis), lpg0802-lpg0811 (fatty acid oxidation, cell wall biosynthesis, cell cycle), lpg1650-lpg1652 (inositol catabolism), lpg1705-lpg1708 (arginine catabolism), and lpg2241-lpg2246 (riboflavin biosynthesis, dicarboxylate transport) (39). Hence, genes implicated in the metabolism of amino acids, carbohydrates, and lipids are regulated by LvbR and LqsR, which might underlie the cell density regulation observed upon growth at 45°C in AYE medium (Fig. 2C) and at 40°C in MDM (Fig. 2E).

In L. pneumophila strain JR32, LvbR links the Lqs system to c-di-GMP signaling (39, 40), and therefore, it is likely that the second messenger c-di-GMP controls temperature-dependent growth rate and cell density of the pathogen. L. pneumophila JR32 harbors 22 genes predicted to encode proteins with domains implicated in c-di-GMP synthesis, hydrolysis, and recognition (61, 62). However, only a few components of the c-di-GMP regulatory network have been characterized, including the Hnox1-Lpg1057 system, comprising the diguanylate cyclase inhibitor Hnox1 and the GGDEF/EAL domain-containing diguanylate cyclase Lpg1057 (41, 42) (Fig. 1). It is currently not known which components of the c-di-GMP regulatory network play a role in the temperature-dependent control of L. pneumophila growth rate and cell density under given conditions.

Recently, we analyzed the role of the Lqs-LvbR network during growth of sessile L. pneumophila on surfaces and in biofilm using single cell techniques (46). We found that sessile L. pneumophila exhibits phenotypic heterogeneity and adopts growing and nongrowing (“dormant”) states in biofilms and microcolonies formed in AYE medium. The establishment of phenotypic heterogeneity was controlled by the temperature, the Lqs system, and LvbR. The Lqs system and LvbR were found to determine the ratio between growing and nongrowing (“persister”) sessile subpopulations, as well as the frequency of growth resumption (“resuscitation”) and microcolony formation of individual bacteria. The microcolony growth rate was the same for the parental strain JR32 and lqs mutant strains, similar to what we observed for planktonic growth in AYE medium at 30°C or 40°C (Fig. 2A and B). LvbR regulated the growth resumption of sessile L. pneumophila (46) but not the lag phase length of planktonic L. pneumophila growing at 30°C or 40°C (Fig. 2A and B), and therefore, substantial differences between the LvbR-dependent growth regulation of sessile and planktonic bacteria exist.

We also assessed the role of the temperature and nutrient availability for growth of Legionella spp. (Fig. 3 and 4, Fig. S3 to S5). Compared to L. pneumophila, L. longbeachae appeared to be more susceptible to elevated temperatures in AYE medium (Fig. 3 and 4). While L. pneumophila strains grew up to 47°C and ceased to grow at 50°C, L. longbeachae reached a lower final cell density upon growth at only 40°C and ceased to grow already at 45°C (Fig. 3 and 4). The most temperature-sensitive L. pneumophila strains were the environmental isolate 500 and the clinical isolate 509 (both sg 1), which at 45°C in AYE medium reached a 2- to 3-fold-lower cell density compared to the reference strain JR32 (Fig. 4C). The final cell density of these strains did not correlate with the differences in the lag phase length, since compared to JR32, strain 500 or 509 showed a shortened or extended lag phase, respectively, at 30°C. In general, the temperature-dependent differences in the final cell density of the environmental and clinical isolates became even more pronounced in MDM (Fig. 4D to F). Only strain JR32 and the environmental isolate 529 grew under these conditions at 30°C, 40°C, and 45°C, albeit to a low density, while all other L. pneumophila strains no longer grew at 45°C (Fig. 4D to F).

The environmental and clinical L. pneumophila strains all harbor the lqsR, lqsA, lqsS, and lvbR genes in their genomes (Fig. 5), and thus, the Lqs-LvbR network might play a role for the temperature-dependent growth patterns but not for the growth differences observed among the strains. The P_lqsR, P_lqsA, P_lqsS, and P_lvbR promoters from strain JR32 were expressed in the environmental and clinical L. pneumophila strains upon growth at 30°C (Fig. 6) and 45°C (Fig. S8). The promoters were expressed with distinct intensities and patterns at the two different temperatures, clearly implicating the temperature as an extrinsic (environmental) cue for the Lqs-LvbR network. Overall, the results documented in this study are in agreement with a general role for quorum sensing (Lqs) and c-di-GMP signaling (LvbR) for growth onset, doubling time, and final cell density of L. pneumophila under different conditions. Future studies will identify the “downstream” components implicated in the temperature-, Lqs-, and LvbR-dependent regulation of L. pneumophila growth under different extra- and intracellular conditions.

MATERIALS AND METHODS

Growth of Legionella strains.

The Legionella strains used in this study are listed in Table 2. Legionella strains were grown on CYE agar plates for 3 days (63), in liquid cultures with N-(2-acetamido)-2-aminoethanesulfonic acid (ACES)-buffered yeast extract (AYE) medium (64) for 23 to 24 h (stationary phase), or in minimal defined medium (MDM) (21) for 28 to 29 h (stationary phase) at 37°C on a wheel (80 rpm). MDM is an ACES-buffered defined medium containing 10 amino acids, including cysteine and 6 mM serine (as carbon and energy source), and salts, including ferric iron (21). To obtain growth curves, Legionella strains were diluted in Eppendorf tubes from cultures grown in AYE medium or MDM to an initial OD₆₀₀ of 0.1 or 0.2 with AYE medium or MDM, transferred into 96-well microplates (100 μL/well, polystyrene, 353072, Falcon), and incubated at the temperatures indicated in the figure legends while orbitally vigorously shaking. Bacterial growth was monitored over time in triplicates by measuring the optical density at 600 nm (OD₆₀₀) using a microplate reader (Synergy H1 or Cytation5 Hybrid Multi-Mode Reader, BioTek). L. pneumophila strains harboring empty control plasmids (pTS10 or pNT28) or complementation plasmids (pAK18, pTS02, pTS04) were cultivated and analyzed in AYE medium supplemented with chloramphenicol (Cm; 5 μg/mL).

TABLE 2.

Legionella strains and plasmids used in this study

Strain or plasmida	Relevant propertiesb	References/sourceb
L. pneumophila
JR32a	L. pneumophila Philadelphia-1, serogroup (sg) 1, salt-sensitive isolate of AM511, clinical isolate, reference strain	65
AK01 (ΔlqsT)	JR32 lqsT::Km	33
AK02 (ΔlqsS-ΔlqsT)	JR32 lqsS::Km lqsT::Gm	33
AK03 (ΔlvbR)	JR32 lvbR::Km	39
NT02 (ΔlqsA)	JR32 lqsA::Km	35
NT03 (ΔlqsR)	JR32 lqsR::Km	36
NT05 (ΔlqsS)	JR32 lqsS::Km	35
478	L. pneumophila, sg 6, clinical isolate	NRCL (#80970); this study
480	L. pneumophila, sg 5, clinical isolate	NRCL (#7980); this study
481	L. pneumophila, sg 1, clinical isolate	NRCL (#10891); this study
485	L. pneumophila, sg 1, clinical isolate	NRCL (#10943); this study
488	L. pneumophila, sg 5, environmental isolate	NRCL (#7797); this study
489	L. pneumophila, sg 6, environmental isolate	NRCL (#7775); this study
493	L. pneumophila, sg 1, environmental isolate	NRCL (#11045); this study
500a	L. pneumophila, sg 1, environmental isolate	NRCL (#10909); this study
509a	L. pneumophila, sg 1, clinical isolate	NRCL (#760); 29, 66
514	L. pneumophila, sg 1, clinical isolate	NRCL (#883); 29, 66
525a	L. pneumophila, sg 6, environmental isolate	NRCL (#1365); 66
526	L. pneumophila, sg 6, clinical isolate	NRCL (#1374); 66
529a	L. pneumophila, sg 1, environmental isolate	NRCL (#1488); 29, 66
534	L. pneumophila, sg 1, environmental isolate	NRCL (#1563); 29, 66
L. longbeachae
NSW150a	L. longbeachae, clinical isolate, reference strain	49
496	L. longbeachae, clinical isolate	NRCL (#8565); this study
L. micdadei
501	L. micdadei, clinical isolate	NRCL (#634); this study
503	L. micdadei, clinical isolate	NRCL (#668); this study
Plasmids
pAK18	pMMB207C-gfp (constitutive)-PlvbR-lvbR, Cm	39
pCM009	pMMB207C-PflaA-gfp (ASV), Cm	30
pNT28	pMMB207C-gfp (constitutive), Cm	36
pRH023	pMMB207C-PlvbR-gfp (ASV), Cm	39
pRH026	pMMB207C-Phnox1-gfp (ASV), Cm	39
pRH037	pMMB207C-PlqsR-gfp (ASV), Cm	This study
pRH038	pMMB207C-PlqsA-gfp (ASV), Cm	This study
pRH051	pMMB207C-PlqsS-gfp (ASV), Cm	This study
pTS02	pMMB207C-RBS, Ptac-lqsA, Cm	29
pTS04	pMMB207C-RBS, Ptac-lqsR, Cm	36
pTS10	pMMB207C-RBS, Cm	36

^aEnvironmental and clinical Legionella isolates analyzed in detail in this study.

^bAbbreviations: NRCL, Swiss National Reference Center for Legionella; Km, kanamycin resistance; Gm, gentamicin resistance; Cm, chloramphenicol resistance.

The experiments and the statistics were performed in biological triplicates, as defined by individual bacterial cultures growing in separate wells of a microtiter plate. The biological triplicates were repeated at least twice in independent experiments, defined as bacterial cultures originating from different precultures, which yielded practically identical results. In most cases, the experiments were independently repeated 3 to 5 times. The resulting growth curves were analyzed with the software GraphPad Prism (version 5.04). Growth rates (maximum) and doubling times (minimum) were calculated using a curve fit with a nonlinear regression model, considering only the fastest exponential growth as defined in semilogarithmic plots.

Plasmid construction.

Cloning was performed according to standard protocols, and the primers used for PCRs are listed in Table 3. Plasmids were amplified and isolated from Escherichia coli TOP10 (Invitrogen) using commercially available kits (Macherey-Nagel). Reporter plasmids pRH037, pRH038, and pRH051 containing a transcriptional P_lqsR-, P_lqsA-, or P_lqsS-gfp (ASV) fusion were constructed by amplifying the insert by PCR using the primer pairs oRH188/189, oRH190/191, or oRH223/224, respectively, and genomic JR32 DNA as a template. PCR products were cloned with the NEBuilder HiFi DNA assembly reaction (NEB) into the SacI and XbaI sites of pCM009 (30), thereby replacing P_flaA. The resulting constructs were verified by DNA sequencing.

TABLE 3.

Oligonucleotides used in this study

Oligonucleotide	Sequence 5′-3′a	Comments
oRH051	TCAAAAAGCATATGTGCTATGC	LvbR_BS (fo) (39)
oRH068	CTCATTTGGTGTGACTTCTAATCC	lvbR 1-201 (re), sequencing
oRH069	CTATGGTCTTCAGCATAAAACAAATC	LvbR_BS (re) (39), sequencing
oRH104	CCACATGAATTGTTCCATCAATGC	sequencing
oRH178	CAAAACAATGCATCCCGCTGC	PlqsR-lqsR (fo)
oRH179	GAGTGCCTGGTCAGACTTGC	PlqsR-lqsR (re)
oRH188	GGAAACAGAATTCGAGCTCACTGCAATAAGCGCCTGCC	PlqsR (fo) (into SacI and XbaI site)
oRH189	CATATGTATATCTCCTTCTTAAATCTAGAGGCTCCTCCTGAGCAAAACG	PlqsR (re) (into SacI and XbaI site)
oRH190	GGAAACAGAATTCGAGCTCCCCCTTGATTGCAGAAGC	PlqsA (fo) (into SacI and XbaI site)
oRH191	CATATGTATATCTCCTTCTTAAATCTAGACCATAACTCAGATTCTTTTCCC	PlqsA (re) (into SacI and XbaI site)
oRH199	ATACAAGTACCTTTTGCTGGC	PlqsA-lqsA (fo)
oRH200	CAATACCTGGAAGAGTTAGATGC	PlqsA-lqsA (re)
oRH201	GAAACATGTATTCACCAAAGGC	PlvbR-lvbR (fo)
oRH202	ATTAATTTTCCTTATCCTGCAAGC	PlvbR-lvbR (re)
oRH208	ATGAAAAAACAAACCGACTTAATTG	lvbR ORF (fo)
oRH209	TTATGAGGCTGAGTTTTTAGTTATTTTC	lvbR ORF (re)
oRH219	AATAACCCTTTCCCTCCTGC	PlqsS-lqsS (fo)
oRH220	CGCTATCTCAGTAAAAACGGC	PlqsS-lqsS (re)
oRH221	ATTTTCCAGCAATAATTCGGC	Phnox1-hnox1 (fo)
oRH223	GGAAACAGAATTCGAGCTCATCGCCAGAGGTTATCATGG	PlqsS (fo) (into SacI and XbaI site)
oRH224	CATATGTATATCTCCTTCTTAAATCTAGATTCTTATCCTTAAATCCAGCTAAAAAACG	PlqsS (re) (into SacI and XbaI site)

^aRegions overlapping with destination vector are underlined.

Genomic DNA extraction, PCR amplification, and sequencing.

Genomic DNA was extracted from Legionella strains cultured for 18 h in AYE medium using the GenElute bacterial genomic DNA kit (NA2100, Sigma-Aldrich), quantified, and used in equal amounts as the template for PCRs. Genes and putative promoter regions were amplified by standard PCR protocols using the primer pairs oRH178/179 (P_lqsR-lqsR), oRH199/200 (P_lqsA-lqsA), oRH219/220 (P_lqsS-lqsS), oRH201/202 (P_lvbR-lvbR), oRH208/209 (lvbR), oRH208/068 (lvbR_1-201), oRH051/069 (LvbR_BS), oRH221/202 (P_hnox1-hnox1), or oRH221/051 (LvbR_BS-hnox1), and the PCR fragments were subsequently analyzed by agarose gel electrophoresis. DNA bands were visualized with MaestroSafe staining (Labgene Scientific SA) and imaged using a FluorChem SP imager (Alpha Innotech). For sequence analysis, the lvbR-hnox1 gene regions were amplified from genomic DNA by PCR using primer pair oRH104/068. PCR products were analyzed by agarose gel electrophoresis, purified (gel and PCR cleanup kit, Macherey-Nagel), and subsequently sequenced by the use of primers oRH068, oRH069, or oRH104 (Sanger sequencing performed by Microsynth, Balgach, Switzerland). An alignment was created with the resulting sequences by the program CLC Main Workbench 7.

GFP reporter assays.

L. pneumophila JR32, 500, 509, 525, or 529 strains harboring P_lqsR-, P_lqsA-, P_lqsS-, P_lvbR-, or P_hnox1-gfp (ASV) fusion reporter plasmids (pRH037, pRH038, pRH051, pRH023, or pRH026) or empty pMMB207C plasmid (pTS10) were grown for 23 to 24 h at 37°C in AYE medium supplemented with chloramphenicol (Cm; 5 μg/mL). Strains were subsequently inoculated from these cultures at an initial OD₆₀₀ of 0.2 in AYE medium/Cm into a black clear bottom 96-well plate (100 μL/well, polystyrene, 353219, Falcon) and incubated at 30°C or 45°C while orbitally shaking. GFP production and bacterial growth were monitored in triplicates by measuring fluorescence (excitation, 485 nm; emission, 528 nm; gain, 50) and the OD₆₀₀ with a microplate reader (Synergy H1 or Cytation5 Hybrid Multi-Mode Reader, BioTek). Blank value (AYE medium) was subtracted from all values, and numbers are expressed as relative fluorescence units (RFU) or OD₆₀₀.