ABSTRACT
Type II protein secretion (T2S) by Legionella pneumophila is required for intracellular infection of host cells, including macrophages and the amoebae Acanthamoeba castellanii and Hartmannella vermiformis. Previous proteomic analysis revealed that T2S by L. pneumophila 130b mediates the export of >25 proteins, including several that appeared to be novel. Following confirmation that they are unlike known proteins, T2S substrates NttA, NttB, and LegP were targeted for mutation. nttA mutants were impaired for intracellular multiplication in A. castellanii but not H. vermiformis or macrophages, suggesting that novel exoproteins which are specific to Legionella are especially important for infection. Because the importance of NttA was host cell dependent, we examined a panel of T2S substrate mutants that had not been tested before in more than one amoeba. As a result, RNase SrnA, acyltransferase PlaC, and metalloprotease ProA all proved to be required for optimal intracellular multiplication in H. vermiformis but not A. castellanii. Further examination of an lspF mutant lacking the T2S apparatus documented that T2S is also critical for infection of the amoeba Naegleria lovaniensis. Mutants lacking SrnA, PlaC, or ProA, but not those deficient for NttA, were defective in N. lovaniensis. Based upon analysis of a double mutant lacking PlaC and ProA, the role of ProA in H. vermiformis was connected to its ability to activate PlaC, whereas in N. lovaniensis, ProA appeared to have multiple functions. Together, these data document that the T2S system exports multiple effectors, including a novel one, which contribute in different ways to the broad host range of L. pneumophila.
INTRODUCTION
The aquatic bacterium Legionella pneumophila is the agent of Legionnaires' disease (1). The environmental persistence of L. pneumophila is largely dependent upon its ability to infect and grow in protozoa, notably amoebae belonging to the Acanthamoeba and Hartmannella genera (2–4). Human infection occurs after the inhalation of L. pneumophila-containing droplets that are generated from a variety of devices (5). In the lung, L. pneumophila multiplies in alveolar macrophages although persistence likely also involves growth in lung epithelial cells and extracellular survival (6, 7). Much of the ecology and pathogenesis of L. pneumophila is mediated by secreted factors (8–11). For secreting proteins into the extracellular milieu and/or target host cells, the organism uses both type II secretion (T2S) and the type IV secretion systems, large membrane-spanning machines that are composed of more than 10 component proteins (6, 9).
T2S promotes the growth, ecology, and virulence of many Gram-negative bacteria (12). T2S substrates are translocated across the inner membrane by the Sec or Tat pathway, and then a pseudopilus may act to push the proteins through a dedicated outer membrane pore (13, 14). On many occasions, we along with others have observed that L. pneumophila mutants lacking the T2S apparatus are severely impaired for infection of Acanthamoeba castellanii and Hartmannella vermiformis, indicating that proteins secreted via T2S are required for infection of protozoan hosts (15–21). The magnitude of the defects is similar in infections of the two types of amoebae. The T2S mutants are not impaired for entry into the amoebae, indicating that T2S is promoting bacterial resistance to intracellular killing and/or facilitating replication itself (22). Others of our studies have shown that T2S also promotes the persistence of L. pneumophila in the environment by facilitating extracellular survival in low-temperature water samples (18, 22–25). Furthermore, L. pneumophila T2S helps to mediate the secretion of a surfactant that confers both sliding motility and antibacterial activity (26, 27). In the mammalian host, L. pneumophila T2S also has a multifactorial role by fostering intracellular multiplication in both macrophages and lung epithelial cells, dampening the chemokine and cytokine output from infected host cells, and elaborating tissue-destructive enzymes (7, 16, 28). Based primarily upon studies examining strain 130b, L. pneumophila T2S promotes the export of at least 25 proteins and 18 enzymatic activities (9, 28–32). From the analysis of a secreted acid phosphatase, we observed, early on, that some T2S-dependent exoproteins have striking similarity to eukaryotic proteins (33). A metalloprotease (ProA) and RNase (SrnA) have been shown to be required for infection of H. vermiformis, and a secreted chitinase (ChiA) has been linked to bacterial persistence in the lungs (19, 28, 34). However, T2S-dependent exoproteins that are required for infection of A. castellanii or macrophages or epithelial cells have yet to be identified. In our previous proteomic analysis of L. pneumophila T2S, a number of the proteins that were in wild-type strain 130b but not in lspF T2S mutant supernatants appeared to have very little if any similarity to known proteins (28). Therefore, to begin the present study, we sought to determine the importance of these potentially novel T2S substrates in infection. We demonstrate, among other things, that one of these Legionella-specific exoproteins significantly promotes infection of A. castellanii. Interestingly, the importance of this protein as well as several other T2S substrates proved to be dependent upon the amoebal host that is being infected.
MATERIALS AND METHODS
Bacterial strains and media.L. pneumophila strain 130b (American Type Culture Collection [ATCC] strain BAA-74) served as our wild-type strain (see Table S1 in the supplemental material). Mutants of strain 130b that were used in this study are listed in Table S1. Strains representing other Legionella species that were also examined are listed in Table S2 in the supplemental material. Legionellae were routinely grown at 37°C on buffered charcoal-yeast extract (BCYE) agar, which, when appropriate, was supplemented with chloramphenicol at 3 μg/ml, kanamycin at 25 μg/ml, or gentamicin at 2.5 μg/ml (27). Escherichia coli strain DH5α was the host for recombinant plasmids (Life Technologies, Carlsbad, CA). E. coli cells were grown in Luria-Bertani medium with kanamycin (50 μg/ml), chloramphenicol (30 μg/ml), or ampicillin (100 μg/ml). Unless otherwise noted, chemicals were obtained from Sigma-Aldrich (St. Louis, MO).
Assessments of bacterial extracellular growth and secreted factors.In order to monitor the extracellular growth of L. pneumophila strains as well as to isolate RNA, legionellae grown on BCYE agar were inoculated into buffered yeast extract (BYE) broth and then incubated at 37°C with shaking (35). The optical densities of the cultures were determined at 660 nm using a DU520 or DU720 spectrophotometer (Beckman Coulter, Indianapolis, IN). The secretion of pyomelanin was ascertained by the presence or absence of browning following bacterial growth in BYE broth to late stationary phase (36). Cell-free supernatants collected from late-log-phase BYE cultures were assayed for protease activity, as measured by azocasein hydrolysis, and for phosphatase activity, as measured by the release of p-nitrophenol from p-nitrophenol phosphate (33, 37). Starch-degrading activity was monitored as previously described (31). To look for the presence of astacin-like activity, culture supernatants were assayed for their ability to trigger the release of Congo red from elastin-Congo red, as previously described (38, 39). Swimming motility was determined by wet mount, and sliding motility was assessed by examining bacteria spotted onto BCYE medium containing only 0.5% agar (27).
DNA and protein sequence analysis.DNA was isolated from L. pneumophila as described previously (27). Primers used for sequencing or PCR were obtained from Integrated DNA Technologies (Coralville, IA). Primer names and their sequences are listed in Table S3 in the supplemental material. DNA sequences were analyzed using Lasergene (DNASTAR, Madison, WI), and protein alignments were done using the Clustal method. Phyre (www.sbg.bio.ic.ac.uk/∼phyre/) and BLAST homology searches were done using the GenBank at the NCBI and the other L. pneumophila databases at http://genolist.pasteur.fr/LegioList/.
Southern hybridizations.Southern blotting was carried out using EcoRI-restricted DNA. A digoxigenin nonradioactive labeling, detection, and stripping system was used (Roche Molecular Biochemicals, Indianapolis, IN). Probes were produced by PCR incorporation using 130b DNA as a template and primers LPW13951-F3 South and LPW13951-R3 South for detection of lpw13951 (nttA), LPW28721-F2 South and LPW28721-R2 South for detection of lpw28721 (nttB), LEGP-F2 South and LEGP-R2 South for detection of lpw32851 (legP), and LSPD-F1 South and LSPD-R1 South for detection of lspD. Low-stringency hybridization conditions (i.e., approximately 30% base pair mismatch allowed) were used as described previously (40).
RT-PCR analysis.Reverse transcription-PCR (RT-PCR) was done essentially as described previously (35). Bacterial RNA was isolated from BYE cultures by using RNA STAT-60 reagent (Tel-Test, Friendswood, TX) and following the manufacturer's instructions, with the exception that 80 μg/μl glycogen (Roche, Mannheim, Germany) and 300 mM sodium acetate were added during RNA precipitation (41). In order to isolate bacterial RNA from infected host cells, A. castellanii amoebae were infected with L. pneumophila (below). After removal of the culture medium, the monolayer was lysed with 50% RNA Protect (Qiagen, Valencia, CA)–1% saponin, and RNA was extracted using RNA STAT-60. Samples were treated with DNase I (Life Technologies, Carlsbad, CA) for 45 min at 37°C, extracted using acid-phenol-chloroform (Life Technologies), and precipitated with sodium acetate-ethanol (42). The purity and concentration of the RNA were confirmed by spectrophotometry (Synergy H1 Hybrid Reader; BioTek, Winooski, VT). cDNA was synthesized in a 20-μl reaction mixture containing 1 μg of RNA and the following items obtained from Life Technologies: 0.12 μg of random primers, 1× First Strand buffer, 2 mM each deoxynucleoside triphosphate (dNTP), 10 mM dithiothreitol (DTT), 40 U of RNaseOut, and 200 U of SuperScript III reverse transcriptase. Primers LPW13951-F4 and LPW13951-R4 were used to examine transcription of lpw13951, LPW28721-F3 and LPW28721-R3 were used for lpw28721, and LEGP-F3 and LEGP-R3 were used for legP. As a control, 16S rRNA gene transcription was assessed using primers 16S-F1 and 16S-F2. Control experiments in which the reverse transcriptase was omitted from the reaction mixture were done to rule out contributions from contaminating DNA. Relative, endpoint PCRs were separated by agarose-gel electrophoresis and detected with ethidium bromide staining.
Mutant constructions.To make mutants of L. pneumophila 130b lacking individual T2S-dependent exoproteins, we performed variations on allelic exchange as previously employed (19, 27, 30, 35). To obtain a mutant lacking lpw13951 (nttA), a 1,783-bp fragment containing lpw13951 was amplified using LPW13951-F1 and LPW13951-R1 and then ligated into pGEM-T Easy (Promega, Madison, WI) to yield pG13951. Next, pG13951 was digested with BglII, and blunt ends were generated with T4 DNA polymerase (Life Technologies). A kanamycin resistance gene (Kmr) from pMB2190 (35) was inserted approximately one-third of the way through lpw13951, yielding pG13951::Km. Following transformation of strain 130b (27) with pG13951::Km, mutant colonies were obtained by plating on BCYE medium containing kanamycin. Verification of the mutant genotype was carried out by PCR, using primers LPW13951-F1 and LPW13951-R1. The two independently derived lpw13951 mutants were designated NU415 and NU416. To obtain a mutant that lacked lpw28721 (nttB), a 1,862-bp fragment containing the gene was amplified using primers LPW28721-F1 and LPW28721-R1 and then ligated into pGEM-T Easy to yield pG28721. Next, pG28721 was digested with BsmI, and blunt ends were generated. A gentamicin resistance gene (Gmr) from pX1918-GT (35) was then inserted to yield pG28721::Gt, which has its Gmr cassette approximately one-fourth of the way through lpw28721. pG28721::Gt was introduced into strain 130b by transformation, and mutant colonies were selected for on BCYE agar containing gentamicin. Verification of the mutant (NU417) genotype was determined by PCR, using primers LPW28721-F1 and LPW28721-R1. To obtain a mutant lacking lpw32851 (legP), a 1,268-bp fragment containing legP as the only intact open reading frame (ORF) was amplified using primers LEGP-F1 and LEGP-R1 and ligated into pGEM-T Easy to yield pGLegP. A 665-bp fragment was removed from legP in pGLegP by digestion with Bst98I and BamHI. Blunt ends were generated, and a Kmr cassette was inserted to yield pGLegP::Km. Next, the SacI-SphI fragment of pGLegP::Km containing the disrupted gene was cloned into pRE112 (40), yielding pRELegP::Km. After pRELegP::Km was electroporated into strain 130b, colonies that were Kmr, chloramphenicol (Cm) sensitive, and sucrose resistant were obtained. The mutant (NU418) genotype was confirmed by PCR using primers LEGP-F1 and LEGP-R1. To obtain an lpw05041 (gamA) mutant, a 1,866-bp fragment containing the gene was amplified using primers GAMA-F1 and GAMA-R1 and ligated into pGEM-T Easy. The resulting pGgamA was digested with BamHI, and after blunt ends were generated, a Kmr cassette was inserted to yield pGgamA::Km. Following transformation of strain 130b with pGgamA::Km, mutants were obtained by plating on BCYE agar containing kanamycin. Verification of the mutant (NU419) genotype was done by PCR, using primers GAMA-F1 and GAMA-R1. The loss of GamA activity was confirmed by the absence of starch hydrolysis on indicator plates (data not shown) (31). In order to isolate an lpw30971 (plaC) mutant, plasmid pGplaC (7) was digested with BamHI, and blunt ends were generated. A chloramphenicol resistance gene obtained from pRE112 was then inserted at bp 165 in the gene to yield pGplaC::Cm. Next, pGplaC::Cm was introduced into strain 130b by transformation, and mutant colonies were obtained on chloramphenicol-containing BCYE agar. Verification of the mutant (NU420) genotype was performed by PCR, using primers PLAC-F1 and PLAC-R1. A double mutant lacking both plaC and proA (NU421) was obtained by introducing pGplaC::Cm into the proA mutant AA200 (19) and then selecting for acquisition of chloramphenicol resistance.
Genetic complementation.Complementation analysis of the plaC mutants was done by making a derivative of NU420 that had the plaC gene (alone) inserted into a neutral site in the chromosome, analogously to what we have previously done for other mutants (35). To that end, a 1,512-bp fragment that included 172 bp of the plaC promoter region was amplified using primers PLAC-F2 and PLAC-R2 and then digested using KpnI and XbaI. The resulting fragment was then ligated into a KpnI/XbaI-digested pMMB2002 (16), yielding pMplaC. Next, the XbaI/SacI fragment of pMplaC was cloned into XbaI/SacI-digested pZL1153 (35), resulting in pZLplaC. Finally, pZLplaC was transformed into NU420 and plated on kanamycin-containing BCYE agar. A Kmr, sucrose-resistant clone (NU423) was obtained and verified as having the plaC gene inserted between lpw27541 and lpw27551. Insertions at this site do not affect the ability of L. pneumophila to grow in standard medium and infection assays (35). Complementation of the lpw13951 mutants was achieved by a similar approach, inserting lpw13951 into the site between lpw27541 and lpw27551 in NU415. To that end, a 922-bp fragment containing the lpw13951 ORF plus 377 bp of upstream sequences was amplified using primers LPW13951-F2 and LPW13951-R2 and then ligated into pGEM-T Easy to yield pG13951c. The lpw13951-containing KpnI/XbaI fragment of pG13951c was ligated into pMMB2002, yielding pM13951c. Then, the SalI/SacI fragment of pM13951c containing lpw13951 was cloned into SalI/SacI-digested pZL1153Gt, resulting in pZL13951c. The pZL1153Gt vector had been generated by removing the Kmr cassette from pZL1153 and replacing it with the Gmr cassette obtained from pX1918-GT. After transformation of NU415 with pZL13951c, the Gmr, sucrose-resistant clone NU422 was obtained. Introduction of the two integrating plasmids, pZLplaC and pZL13951c, into the chromosome of wild-type strain 130b did not alter intracellular multiplication (data not shown).
Intracellular infection assays.To assess L. pneumophila growth within mammalian cells, we infected human U937 macrophages (ATCC CRL-1593.2) as previously done (7). To examine the ability of L. pneumophila strains to grow in protozoa, A. castellanii (ATCC 30234), H. vermiformis (ATCC 50237), and Naegleria lovaniensis (ATCC 30569) were infected as described previously (19, 27, 30, 43, 44). H. vermiformis and N. lovaniensis were infected in ATCC medium number 1034 that lacked its serum component, whereas infection of A. castellanii utilized protease peptone-yeast extract (PY) medium.
RESULTS
Identification of novel proteins that are secreted via T2S.This study began with a further examination of three T2S-dependent substrates of L. pneumophila strain 130b that had appeared to be novel based upon preliminary sequence analysis (28). The first substrate was a 53-kDa protein that, based upon the now-complete database of strain 130b (45), is encoded by ORF lpw13951 (Fig. 1A). Based upon current searches, the Lpw13951 protein had no similarity to anything outside the Legionella database. Thus, lpw13951 was designated nttA, for novel type two secreted protein A. A gene that is 98 to 99% identical to nttA was detected in the other currently sequenced strains of L. pneumophila. An NttA-like protein was identified in databases that are available for strains representing three non-pneumophila species of Legionella (see Table S2 in the supplemental material). Southern blot analysis further revealed the presence of an nttA-like gene in two of seven other non-pneumophila species examined (see Table S2). Together, these data indicated that NttA is conserved among L. pneumophila strains, present in a subset of the other Legionella species, and absent outside the Legionella genus.
L. pneumophila loci encoding novel secreted proteins. Depiction of the region of the strain 130b chromosome containing nttA (A), nttB (B), and legP (C). The horizontal arrows denote the locations, orientations, and relative sizes of the genes. The “lpw” numbers indicated within the arrows are ORF designations used in the database. Gene names and annotations are below the arrows. Sizes of the genes (bp) are given above the arrows, and the sizes of intergenic regions are noted between the arrows.
The second T2S-dependent substrate that we sought to further investigate was a 39-kDa protein that was encoded by lpw28721 (Fig. 1B). Lpw28721 displayed only very slight similarity (i.e., E value of ≥2 × 10−5) to hypothetical proteins in non-Legionella bacteria. In some cases, the proteins were annotated as putative cysteine proteases. Thus, we named lpw28721 nttB, for novel type two secreted protein B. A gene that is 97 to 99% identical to nttB was detected in the other sequenced L. pneumophila strains. L. pneumophila also had an NttB paralog; i.e., NttB shared 44% identity and 61% similarity across approximately 300 amino acids (aa) (E value of 2 × 10−75) with Lpw28341. No homolog or paralog of NttB was found in databases of Legionella longbeachae, Legionella drancourtii, and Legionella dumoffii. However, when we did Southern hybridizations, an nttB-like gene was detected in three of the seven species tested (see Table S2 in the supplemental material). These data indicated that NttB is another T2S-dependent substrate that appears to be unique to members of the Legionella genus.
The final T2S substrate that was reexamined was a 29-kDa protein encoded by ORF lpw32851. Previously, we had noted that Lpw32851 contained a short sequence that is similar to a metalloprotease domain that was first characterized in the crayfish enzyme astacin (28, 46). Other bioinformatic analyses noted the gene in L. pneumophila strains Philadelphia-1, Paris, and Lens as being “eukaryotic-like,” and one of the studies dubbed the gene legP for Legionella eukaryotic-like gene protease (47–49). Current results confirmed that LegP has similarity to hypothetical proteins in invertebrates, such as Drosophila sp. (XP_001999174, with an E value of 9 × 10−41), and vertebrates, such as Mus musculus (NP_766127, with an E value of 9 × 10−37). However, an examination of 130b culture supernatants failed to detect astacin-like activity (data not shown). Furthermore, the new BLASTP analysis revealed that LegP, despite its appellation, has its greatest similarity to hypothetical proteins of other bacteria, such as one from Roseobacter sp. (YP_004692382; E value of 1 × 10−68), and Cellulomonas fimi (YP_004453501; E value of 2 × 10−69). Although there was no homolog or paralog of LegP in databases of L. longbeachae, L. drancourtii, or L. dumoffii, Southern blotting detected a legP-like gene in Legionella cardiaca, Legionella moravica, and Legionella worsleiensis (see Table S2 in the supplemental material). In sum, even though LegP did not prove to have the level of uniqueness that the other two T2S substrates had, it is representative of an uncharacterized family of proteins that exists in both prokaryotes and eukaryotes.
Compatible with our previous proteomic analysis of strain 130b culture supernatants, RT-PCR analysis determined that nttA, nttB, and legP are all expressed when strain 130b is grown to stationary phase in BYE broth at 37°C (see Fig. S1A in the supplemental material).
L. pneumophila mutants lacking novel T2S substrates.In order to discern if the newly defined T2S substrates are needed for L. pneumophila growth, we generated 130b mutants specifically lacking nttA, nttB, or legP. All of the mutants grew normally in BYE broth (see Fig. S2 in the supplemental material), indicating that the mutants do not have a generalized growth defect and that nttA, nttB, and legP are not required for extracellular growth. These data were compatible with the fact that T2S (lsp) mutants of L. pneumophila grow normally in BYE broth (16). The new mutants displayed typical L. pneumophila colony morphology when grown on BCYE agar as well as normal cell shape and swimming motility (data not shown). They also behaved as parental 130b cells did in terms of surface translocation (i.e., sliding motility) and surfactant production (see Fig. S3 in the supplemental material). Finally, the mutants had normal levels of pyomelanin pigment, acid phosphatase, and protease activity within their BYE culture supernatants (data not shown), indicating that they do not have generalized defects in T2S and other forms of secretion.
Novel T2S substrate NttA is required for intracellular infection of A. castellanii but not infection of H. vermiformis.To begin to determine the importance of novel T2S substrates in infection, we assessed the ability of the new mutants to grow in U937 cell macrophages. The nttB and legP mutants grew as well as the wild type did (Fig. 2A), indicating that NttB and LegP are not needed for optimal infection of macrophages. Turning to protozoan models of intracellular infection, we observed that the same two mutants grew normally within H. vermiformis and A. castellanii, indicating that NttB and LegP are also not required for infection of multiple types of amoebae (Fig. 2B and C).
Intracellular infection of macrophages, H. vermiformis, and A. castellanii by L. pneumophila wild-type, nttB mutant, and legP mutant strains. U937 cells, H. vermiformis, and A. castellanii, as indicated on the figure, were infected with either the wild-type (WT) 130b, nttB mutant NU417, or legP mutant NU418 strain, and at the indicated times, CFU counts from the infected monolayers were determined. Data are the means and standard deviations for three to four infected wells. Data in panel A are representative of two independent experiments, whereas data in panels B and C represent three trials.
NttA mutant NU415, although growing normally in macrophages and in hartmannellae (Fig. 3A and B), exhibited reduced recovery upon infection of acanthamoebae (Fig. 3C). Indeed, at 48 h and 72 h postinoculation, the nttA mutant-infected A. castellanii cultures contained about 30-fold fewer bacteria. The nttA mutant did not exhibit reduced survivability when incubated in the PY assay medium (see Fig. S4A in the supplemental material) or in conditioned PY medium obtained from infected monolayers (see Fig. S4B), affirming that the strain's reduced recovery from infected monolayers was due to impaired intracellular infection. That NU415 was not defective when infecting H. vermiformis in PY medium (see Fig. S4C) further indicated that the reduced recoverability of the nttA mutant from A. castellanii was not simply due to the particular medium used in infection of those amoebae. Because a second, independently derived nttA mutant (i.e., NU416) had the same defect (Fig. 3D) and because the mutant phenotype was complemented by reintroduction of nttA into the chromosome (Fig. 3E), we concluded that NttA is required for optimal infection of A. castellanii. Compatible with this conclusion, RT-PCR analysis determined that nttA is expressed during L. pneumophila infection of A. castellanii (see Fig. S1B). Since the nttA mutants were not as defective in A. castellanii as an lspF mutant was (Fig. 3C), we posit that more T2S effectors facilitate intracellular infection of these amoebae.
Intracellular infection of macrophages, H. vermiformis, and A. castellanii by L. pneumophila wild type, nttA mutants, and a complemented nttA mutant. U937 cells (A), H. vermiformis (B) and A. castellanii (C) were infected with WT and nttA mutant NU415 strains. In addition, A. castellanii was infected with the lspF mutant NU275 (C) nttA mutant NU416 (D), and complemented nttA mutant NU422 (E), and at the indicated times, CFU counts from the infected monolayers were determined. Data are the means and standard deviations for three to four infected wells. Data in panels A to D are representative of three independent experiments, and data in panel E are representative of two trials. As shown in panel C, the recovery of the nttA mutant NU415 was significantly less than that of the WT at both 48 h (P = 0.024; Student's t test) and 72 h (P = 0.010). The recovery of the nttA mutant NU416 was significantly less than that of the WT at 48 h (P = 0.017) and 72 h (P = 0.009) (D). The recovery of the nttA mutant was again different from that of WT and also significantly less than that of the complemented nttA mutant at 48 h (P = 0.006) and 72 h (P = 0.005) (E). For the data shown in panel E, P = 0.012 and P = 0.024 at 48 and 72 h, respectively, for recovery of the WT compared to that of the complement.
T2S substrate SrnA is required for intracellular infection of H. vermiformis but not infection of A. castellanii.The fact that the nttA mutant was impaired for infection of A. castellanii but not for infection of H. vermiformis indicated that the importance of a T2S substrate can be dependent upon the amoebal host that is being infected. We had reached a similar conclusion before when testing a 130b mutant lacking secreted ProA although, in that instance, the mutant was impaired in H. vermiformis not in A. castellanii (19). However, the only other T2S substrate mutants that were tested in two amoebae were strains lacking either the LapA and LapB aminopeptidases or the endoglucanase CelA, and in these cases, no infection defects were seen (19, 30). In light of the results involving nttA and past data concerning proA, we considered the possibility that another T2S substrate(s) that had been previously assessed in only one amoebal model might prove to have a different importance if a second amoebal model were tried.
In one past study, an L. pneumophila mutant defective for gamA was tested only in A. castellanii and found to grow normally (31). Therefore, we made a 130b mutant lacking gamA and tested it in H. vermiformis. The gamA mutant was not impaired for infection of H. vermiformis (see Fig. S5A in the supplemental material), nor was it defective in A. castellanii as was expected (data not shown). Since a gamA mutant had not been previously tested in macrophages, we took the opportunity to test the new mutant in U937 cells, but once again it grew as well as the wild type did (see Fig. S5B). These data indicated that GamA is another T2S-dependent exoprotein that is not required for infection of H. vermiformis or A. castellanii.
In our own past studies, mutants of 130b lacking either chiA, lipA, lipB, map, plaA, plcA, or srnA were tested only for their ability infect H. vermiformis, and with the exception of the srnA mutant, all grew normally. Thus, we tested mutants lacking these genes in A. castellanii. In testing the importance of plcA, we utilized a recently made double mutant that lacks both PlcA and PlcB, a putative T2S-dependent phospholipase C that is closely related to PlcA (7). All of the mutants tested infected A. castellanii as well as the wild type did (Fig. 4). We also found that the plcA plcB mutant grew normally in H. vermiformis and U937 cells (data not shown). These data indicated that ChiA, LipA, LipB, Map, PlaA, PlcA, and PlcB, like CelA, LapA, and LapB, are not required for infection of H. vermiformis or A. castellanii. But the data did document that SrnA, like ProA, is required for infection of H. vermiformis but not infection of A. castellanii.
Intracellular infection of A. castellanii by L. pneumophila wild type and various other T2S substrate mutants. Acanthamoebae were infected, as indicated on the figure, with either the WT 130b, chiA mutant NU318, lipA lipB mutant NU373, map mutant NU254, plaA mutant NU270, plcA plcB mutant NU371, or srnA mutant NU328 strain, and then at the indicated times, CFU counts from the infected monolayers were determined. Data are the means and standard deviations for four infected wells, and each panel is representative of three experiments.
T2S substrate PlaC is required for intracellular infection of H. vermiformis but not infection of A. castellanii.In another previous study (50), an L. pneumophila mutant lacking plaC was tested in A. castellanii (and macrophages) and found to grow as well as the wild type did. Thus, we generated a new 130b mutant (NU420) lacking plaC and tested it in H. vermiformis. The plaC mutant, although still not impaired in A. castellanii, was defective in H. vermiformis (Fig. 5A and B). At 48 and 72 h postinoculation, the plaC mutant-infected cultures contained 4- to 10-fold fewer bacteria. The plaC mutant did not show reduced survivability when incubated in the 1034 assay medium alone with and without conditioning (data not shown). We also observed that NU420 was not defective when infecting A. castellanii in 1034 assay medium (data not shown), indicating that the reduced recovery of the mutant was not due to the medium used. Because a second, independent plaC mutant (NU367) had the same defect (Fig. 5C) and because complementation of the mutant phenotype was obtained (Fig. 5D), we concluded that PlaC is required for optimal intracellular infection of H. vermiformis. These data showed that PlaC is required for infection of H. vermiformis but not infection of A. castellanii, reinforcing our conclusion that T2S substrates can be important in a host cell-specific manner.
Intracellular infection of H. vermiformis and A. castellanii by L. pneumophila wild type, plaC mutants, a complemented plaC mutant, and a plaC proA mutant. A. castellanii and H. vermiformis were infected, as indicated on the figure, with either the WT 130b, plaC mutant NU420, plaC mutant NU367, complemented plaC mutant NU423, proA mutant AA200, or plaC proA mutant NU421 strain, and then at the indicated times, CFU counts from the infected monolayers were ascertained by plating. Data are the means and standard deviations for four infected wells. Data in all panels represent three independent experiments, except for the data in panel D, which represent two trials. As shown in panel B, the recovery of the plaC mutant NU420 was significantly less than that of the WT at both 48 h (P = 0.046; Student's t test) and 72 h (P = 0.029). The recovery of the plaC mutant NU367 was significantly less than that of the WT at 48 h (P = 0.005) and 72 h (P = 0.0005) (C). Recovery of the plaC mutant was again different from that of WT and also significantly less than that of complemented plaC mutant at 48 h (P = 0.027) and 72 h (P = 0.015) (D). For panel D, P = 0.501 and P = 0.668 at 48 and 72 h, respectively, for a comparison of the recovery of the WT to that of the complement. As shown in panel E, the recovery of the plaC mutant was again less than that of WT at 48 h (P = 0.025) and 72 h (P = 0.007), as were the recoveries of the proA mutant and proA plaC double mutant at 72 h (P = 0.008).
Role of ProA in H. vermiformis infection is related to its substrate PlaC.One of our earlier reports on T2S indicated that some T2S substrates might be cleaved by ProA (51), and subsequent work found that the lipolytic activity of PlaC is, in fact, dependent upon cleavage of PlaC by ProA (50, 52). Therefore, we made a proA plaC double mutant (NU421) and assessed its infectivity relative to that of the proA mutant and plaC mutant. In the H. vermiformis infection model, the double mutant was no more defective than either of the single mutants (Fig. 5E), suggesting that the importance of ProA in H. vermiformis infection is related to its role in activating PlaC.
T2S and substrates PlaC, ProA, and SrnA promote infection of N. lovaniensis.By fully utilizing two amoebal models, we were able to learn more about the importance of individual T2S-dependent substrates. Thus, we next sought to further improve our appreciation of T2S by incorporating a third amoebal host into our analysis. Although hartmannellae and acanthamoebae are the amoebae most frequently implicated as natural hosts for L. pneumophila, studies have implicated Naegleria species as being another natural reservoir (2, 53, 54). Therefore, we developed a model using N. lovaniensis. To begin, we determined that strain 130b can infect the naegleriae quite well, achieving a level of growth that was comparable to that observed with acanthamoebae and hartmannellae (Fig. 6A). An lspF mutant of 130b, but not its complement, was severely impaired in N. lovaniensis (Fig. 6A), demonstrating that L. pneumophila T2S is critical in at least three different amoebal hosts. Turning to an analysis of T2S substrates, we found that the nttA, nttB, and legP mutants were not impaired in N. lovaniensis (Fig. 6B to D), nor were most of our other mutants (see Fig. S6 in the supplemental material). However, mutants lacking plaC, proA, or srnA were impaired in N. lovaniensis (Fig. 6E). Unlike the result obtained in infection of H. vermiformis, the proA mutant and the plaC proA double mutant were more defective than the plaC mutant (Fig. 6F), suggesting that in some host cells ProA can have key activities beyond that of activating PlaC.
Intracellular infection of N. lovaniensis by L. pneumophila wild type and various T2S substrate mutants. The naegleriae were infected with either the WT 130b, lspF mutant NU275 and its complemented derivative ([dagger]), legP mutant NU418, nttA mutant NU415, nttB mutant NU417, srnA mutant NU328, plaC mutant NU420 and proA mutant AA200, or plaC proA double mutant strain, as indicated on the figure, and then at the indicated times, the numbers of CFU from the infected monolayers were determined by plating. Data are the means and standard deviations for four infected wells. Data in panels A to D are representative of two independent trials, and data in panels E and F represent three independent experiments. As shown in panel A, the recovery of the lspF mutants was significantly less than that of the WT and the complemented lspF mutant at 24, 48, and 72 h postinoculation (P < 0.05; Student's t test). The recoveries of the srnA and plaC mutants (E and F) were less than the recovery of the WT at 48 and 72 h, whereas the recoveries of the proA and plaC proA mutants were less than the WT recovery at 24, 48, and 72 h (P < 0.05; Student's t test).
DISCUSSION
The results presented here advance our appreciation of T2S by L. pneumophila in several ways. With the demonstration of a role for NttA in A. castellanii infection, we have our first example of a substrate promoting growth in an Acanthamoeba host. The documentation of a role for PlaC in growth within H. vermiformis brings to three (PlaC, ProA, and SrnA) the number of substrates that are required for infection of a Hartmannella host. Together, these data have confirmed the hypothesis that the inability of T2S (lsp) mutants to infect A. castellanii and H. vermiformis is due, at least in part, to the loss of secreted substrates as opposed to being due to potential cell-associated defects. The relatively modest effect of each of the individual substrate mutations is compatible with a scenario in which the role of T2S is an additive effect of multiple, secreted proteins. That the importance of NttA was revealed after testing only three of the novel substrates (NttA, NttB, and LegP)—whereas a greater effort that had focused on proteins with similarity to known enzymes yielded only three required promoters of intracellular infection (PlaC, ProA, and SrnA)—suggests that substrates that are more unique in sequence and specific to Legionella may be critical for L. pneumophila persistence. With the demonstration of a role for lspF, plaC, proA, and srnA in N. lovaniensis infection, we conclude that T2S is required for L. pneumophila infection of at least three genera of amoebae. The fact that an nttA mutant was impaired in A. castellanii but not H. vermiformis or N. lovaniensis and that a plaC mutant, proA mutant, and srnA mutant were impaired in H. vermiformis and N. lovaniensis but not in A. castellanii documents that the importance of a particular T2S substrate can be dependent on the amoeba being infected and implies that the exoprotein repertoire has a role in shaping host range. Presumably, different amoebae present different intracellular environments for infecting legionellae (to surmount); e.g., NttA may have evolved to interact with a factor that is present or highly expressed in A. castellanii but absent from or poorly expressed in H. vermiformis. Alternatively, the expression levels of T2S substrates are different in different amoebae. Based on the behavior of our mutants, infection of H. vermiformis and infection of N. lovaniensis appear to be more similar to each other than they are to infection of A. castellanii.
Given the novelty of NttA at this moment, it is difficult to predict what the protein might be doing within the infected amoebae. One possible direction to pursue is determining the protein's three-dimensional (3D) structure as it may reveal a conserved fold that can then give a clue to function. Unlike NttA, PlaC is an known enzyme, a glycerophospholipid:cholesterol acyltransferase that has phospholipase A and lysophospholipase activities (50). Thus, we posit that PlaC is altering ergosterol-containing membranes (52) of H. vermiformis and N. lovaniensis (but not A. castellanii) and thereby might influence processes such as the trafficking of the phagosome or the flux of factors into or out of it. Another consideration is that the fatty acid profile and position of fatty acids in glycerophospholipids may be different in the different amoebae. Based on our analysis of the plaC proA mutant, the key role of ProA in H. vermiformis infection appears to be to activate PlaC. However, this does not appear to be the case during infection of N. lovaniensis. Hence, ProA may, in certain hosts, be helping L. pneumophila obtain amino acids or cleave other secreted proteins (besides PlaC) that promote intracellular growth. As the last of the known T2S-dependent potentiators of intracellular infection, SrnA might be degrading H. vermiformis and N. lovaniensis RNA in order to obtain nucleotides and phosphate.
Although NttB and LegP proved not to be required for infection of macrophages, A. castellanii, H. vermiformis, or N. lovaniensis, this does not necessarily mean that they are irrelevant for infection. Indeed, another factor may compensate for their loss, or they may be more important in another host or other niches such as biofilms (2, 55). LegP joins the phosphatase Map (33) as being a T2S-dependent substrate that is eukaryotic-like, reflecting the relationship that L. pneumophila has with eukaryotic hosts. Although we along with others have shown that LegP is secreted into culture supernatants in a T2S-dependent manner (28, 32), other work indicates that during infection of J774 mouse macrophages, the protein is translocated out of the bacterial phagosome and into the host cell cytoplasm in a type IV secretion (Dot/Icm)-dependent manner (56). This raises the intriguing scenario whereby LegP and perhaps other proteins are directly secreted or indirectly influenced by multiple secretion pathways, with environmental conditions (e.g., extracellular versus intracellular) dictating which secretion pathway is most critical.
Table 1 is a summary of the T2S-dependent exoproteins of L. pneumophila that have now been examined for their broad role in infection. Based upon these data, it may be instructive to incorporate even more hosts into the analysis; e.g., L. pneumophila can grow in at least 18 more amoebae, including seven other species of Acanthamoeba, five other species of Naegleria, another species of Hartmannella, and species of Balamuthia, Dictyostelium, Echinamoeba, Vahlkampfia, and Willaertia (58–77), as well as three species of Tetrahymena (64, 69, 78–80). It has often been argued that when an effector mutant lacks an infection defect, it is the result of functional redundancy; i.e., the loss of a single, secreted protein is compensated for by the (up-) expression of another protein. In light of current data, which follow from our earlier assessment of ProA (19), mutants are best tested in multiple protozoa before the argument of functional redundancy is invoked. This strategy may be even more important for investigation of L. pneumophila type IV secretion which mediates the secretion of >270 effectors (6, 81).
Role of T2S and individual T2S substrates in intracellular infection
The Legionella genus consists of 57 species (82). Previously, we found that all seven non-pneumophila species examined contained lsp genes encoding T2S (16). Based on the data in Table S2 in the supplemental material, we can now say that nine more non-pneumophila species have lsp genes. Thus, we suspect that lsp genes exist throughout the Legionella genus, as one would predict, given the prevalence of T2S among Proteobacteria (12). However, we have been gaining evidence that the T2S output varies between Legionella species (22, 26, 82). Hybridization analysis of nttA, nttB, and legP, coupled with BLAST analysis of recently sequenced non-pneumophila genomes, indicates that the presence of genes encoding substrates varies among species. Thus, the variations in T2S output are most likely due to differences in substrate gene content versus the presence or absence of the T2S apparatus. It will be interesting to more systematically assess substrate genotypes among different species as this may give clues to evolutionary relationships. A molecular evolution analysis of (just) L. pneumophila strains has already concluded that ProA and SrnA have been selected due to their role in virulence mechanisms (83).
ACKNOWLEDGMENTS
We thank members of the Cianciotto lab for their helpful comments. We also thank Felizza Gunderson for providing RNA, Kessler McCoy-Simandle for assisting with the plcA plcB and chiA mutants, and Catherine Stewart for help with analyzing plaC.
J.Y.T. was partly supported by NIH training grant T32 AI0007476. This study was funded by NIH grant AI043987 awarded to N.P.C.
FOOTNOTES
- Received 11 January 2013.
- Returned for modification 7 February 2013.
- Accepted 11 February 2013.
- Accepted manuscript posted online 19 February 2013.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.00045-13.
- Copyright © 2013, American Society for Microbiology. All Rights Reserved.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
