The DotL protein, a member of the TraG-coupling protein family, is essential for Viability of Legionella pneumophila strain Lp02

Abstract

Legionella pneumophila is able to survive inside phagocytic cells by an internalization route that bypasses fusion of the nascent phagosome with the endocytic pathway to allow formation of a replicative phagosome. The dot/icm genes, a major virulence system of L. pneumophila, encode a type IVB secretion system that is required for intracellular growth. One Dot protein, DotL, has sequence similarity to type IV secretion system coupling proteins (T4CPs). In other systems, coupling proteins are not required for viability of the organism. Here we report the first example of a strain, L. pneumophila Lp02, in which a putative T4CP is essential for viability of the organism on bacteriological media. This result is particularly surprising since the majority of the dot/icm genes in Lp02 are dispensable for growth outside of a host cell, a condition that does not require a functional Dot/Icm secretion complex. We were able to isolate suppressors of the Delta dotL lethality and found that many contained mutations in other components of the Dot/Icm secretion system. A systematic analysis of dot/icm deletion mutants revealed that the majority of them (20 of 26) suppressed the lethality phenotype, indicating a partially assembled secretion system may be the source of Delta dotL toxicity in the wild-type strain. These results are consistent with a model in which the DotL protein plays a role in regulating the activity of the L. pneumophila type IV secretion apparatus.

FIG. 1.

(Top) DotL shows sequence similarity to members of the T4CP family. DotL has extensive similarity to a number of putative type IVB secretion system ATPases, including an uncharacterized ORF in Coxiella burnetii [ORF (C.b.)], the TrbC protein of the IncI plasmid R64 [TrbC (R64)], and a TrbC orthologue on the pPT23A plasmid of Pseudomonas syringae strains [ORF (pPT23A)]. DotL has similarity to plasmid T4CPs (MobB, TraD, TraJ, TrwB, and TraG from the plasmids CloDF13, F, pKM101, R388, and RP4, respectively) and to T4CPs from adapted conjugation systems found in pathogens (HP0524 from Helicobacter pylori, VirD4 from the Ti plasmid of Agrobacterium tumefaciens, and a VirD4 orthologue from Rickettsia prowazekii). Most strains of L. pneumophila contain at least one additional T4CP, LvhD4, which is part of a second type IV secretion system (51). The dendrogram was generated by using CLUSTAL W alignment. (Bottom) The DotL protein is localized to the inner membrane of L. pneumophila. Extracts of wild-type L. pneumophila were separated into cytoplasmic and membrane fractions by high-speed centrifugation. The membrane fractions were then further separated into inner membrane versus outer membrane fractions by extraction with the detergent Triton X-100. Duplicate samples were run on two 7.5% acrylamide gels; the first gel was transferred to a polyvinylidene difluoride membrane and probed with anti-DotL serum (lanes 1 to 5), whereas the second gel was stained with Coomassie blue for total protein (lanes 6 to 10). Lanes 1 and 6 are total cell lysates, lanes 2 and 7 are soluble cytoplasmic fractions, lanes 3 and 8 are total membrane, lanes 4 and 9 are Triton X-100 soluble (inner membrane), and lanes 5 and 10 are Triton X-100 insoluble (outer membrane). All samples were loaded proportionally except for lanes 8, 9, and 10, which were overloaded in order to detect the protein profile (lane 8 is 3-fold, lane 9 is 2-fold, and lane 10 is 25-fold overloaded relative to lanes 1 to 7). The quality of the fractionation procedure can be determined by monitoring the localization of the major outer membrane protein, MOMP, on the Coomassie blue-stained gel (lane 10) (44). A DotL breakdown product detected by Western analysis is indicated with an asterisk (lane 2).

FIG. 2.

Assay for ability of L. pneumophila to tolerate the ΔdotL mutation. A merodiploid consisting of a wild-type copy of dotL and a dotL deletion was constructed by integration of the suicide plasmid pJB1001. This plasmid contains an origin, from the R6K plasmid, that is unable to replicate in L. pneumophila strains lacking the replication protein π (30). pSR47S also contains the selectable marker, Kan^r, and a counterselectable marker, sacB, which confers sensitivity to sucrose. The kanamycin marker was used to select for a single crossover generating a merodiploid strain containing both dotL⁺ and ΔdotL (step A). Recombination between duplicated sequences in the heterozygote was selected by growth on 5% sucrose (step B). If dotL is a nonessential gene, both dotL⁺ and the ΔdotL will be obtained (step C). If dotL is an essential gene, then only wild-type dotL will be recovered.

FIG. 3.

dotL is an essential gene on bacteriological media. dotL and dotB merodiploids were constructed and recombinants selected as described in Fig. 2 were analyzed by Southern analysis (as described in Materials and Methods). The top panel shows a Southern blot of recombinants derived from a parental dotL/ΔdotL merodiploid strain probed with a 700-bp SalI fragment of DNA adjacent to the dotL gene. Lane 1 is the wild-type strain Lp02; lane 2 is JV1003, a ΔdotL/dotL merodiploid; and lanes 3 to 16 are JV1003 plated on CYET plus 5% sucrose. All 14 strains that were selected for sucrose resistance in this fashion retained the wild-type version of dotL. In contrast, the bottom panel is a similar experiment in which sucrose resistant recombinants derived from a dotB/ΔdotB merodiploid strain were probed with a dotB-region specific probe. Lane 1 is the wild-type strain Lp02; lane 2 is JV941, a dotB/ΔdotB merodiploid; and lanes 3 to 16 are JV941 selected on CYET plus 5% sucrose. In this case, two distinct types of recombinants are observed, a finding consistent with either dotB or ΔdotB, indicating that dotB is not an essential gene.

FIG. 4.

Mini-Tn10 insertions in multiple dot/icm genes suppress the lethality caused by loss of dotL. The dotL/ΔdotL merodiploid strain, JV1003, was mutagenized with mini-Tn10, and viable strains harboring ΔdotL were directly selected on sucrose-chloramphenicol-containing plates. Shown are the L. pneumophila dot/icm regions I and II (63). dot/icm genes are indicated with filled arrows, whereas flanking genes that are not required for intracellular growth are designated by open arrows (ORFs 1 to 9). Region I contains an 8-kb intervening region, which contains apparent housekeeping genes, separating the two dot/icm loci. dotV is separated from the rest of the dot/icm genes in region II by 20 kb. Mini-Tn10 insertions that suppressed the ΔdotL lethality were found in dotA, dotG, dotI, dotO, dotV, and icmF, and the sites of insertions are indicated with vertical arrows.

FIG. 5.

The dotL gene can be deleted in a strain lacking dotA. Southern blot analysis of ΔdotL recombinants in a ΔdotA background. A 700-bp SalI fragment from pJB1001 encoding DNA flanking dotL on the chromosome was used as a probe to determine the status of dotL in these strains. Lane 1 is the wild-type strain Lp02; lane 2 is JV1005, a ΔdotL/dotL merodiploid in a dotA mutant background; and lanes 3 to 16 are JV1005 resolved on CYET plus 5% sucrose.

FIG. 6.

Deletion of most dot/icm genes can suppress the lethality caused by deletion of dotL. A dotL/ΔdotL::CmR merodiploid was constructed in a variety of different dot/icm backgrounds. Sucrose-resistant recombinants were selected and then screened for the ΔdotL allele by resistance to chloramphenicol. ΔdotL recombinants could not be obtained from the dotL/ΔdotL::Cm^r merodiploid strain JV1003. The presence of a wild-type copy of dotL on a low-copy vector allowed the isolation of ΔdotL recombinants (JV1003 plus dotL). Inactivation of 20 of 23 dot/icm genes suppressed the loss of dotL. In addition, deletion of citA/tphA, a housekeeping gene found near the dot/icm genes, did not allow loss of dotL (42). In contrast, dotL is not essential in a related L. pneumophila strain, JR32. The data shown reflects the average number of ΔdotL recombinants recovered from scoring 50 events from four independent experiments.

FIG. 7.

dotL is required for growth of the L. pneumophila strain JR32 in U937 cells. A number of L. pneumophila strains were assayed for their ability to replicate inside U937 cells over 3 days. The top panel includes Lp02 (wild type) and Lp03 (a dotA mutant derivative of Lp02) as controls. The bottom panel includes JR32 containing the vector pJB908, a JR32ΔdotL strain containing the dotL⁺ complementing clone pJB1014, and a JR32ΔdotL strain containing the vector pJB908. The data shown are the average of triplicate samples and are representative of two independent experiments.

FIG. 8.

The lvhB operon is not responsible for viability of the JR32 ΔdotL strain. (Top) The presence of the lvhB/lvhD operon in a variety of L. pneumophila strains was assayed by Southern analysis with a probe that contains the entire operon: lane 1 is the Philadelphia-1 progenitor of JR32, lane 2 is JR32, lane 3 is the Philadelphia-1 progenitor of Lp01 and Lp02, lane 4 is Lp01, and lane 5 is Lp02. (Bottom) A dotL/ΔdotL merodiploid of JR32 can be resolved to the ΔdotL, indicating it is not an essential gene. Two independently derived JR32 strains lacking the lvhB operon, JV1630 and JV1631, still allow deletion of dotL, whereas dotL is essential for viability in Lp02. The data shown reflect the average number of ΔdotL recombinants recovered from scoring 50 events from four independent experiments.

FIG. 9.

Model for potential DotL regulation of the Dot/Icm translocator. (A) In wild-type L. pneumophila strains, the Dot/Icm proteins form a secretion apparatus in the membrane, which is used to export substrate(s). DotL is shown interacting with the complex as a hexameric gate based on homology to the hexameric T4CP, TrwB (21). Translocated substrates would be exported through the complex after interacting with DotL on the cytoplasmic face of the inner membrane. (B) During conditions in which L. pneumophila is not actively secreting substrates, the export apparatus would be closed via DotL and potentially substrates such as LidA (indicated as a ball) (13). (C) In the absence of DotL, the secretion pore might remain constitutively open and the cell would die, possibly due to cell lysis. (D) Inactivation of the Dot/Icm complex would suppress the ΔdotL lethality since an unregulated pore would no longer exist.