![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
dotL Lethality Phenotype
Department of Molecular Microbiology,1 Department of Cell Biology and Physiology, Washington University School of Medicine, St. Louis, Missouri 63110,3 Apath, LLC, 893 North Warson Road, Saint Louis, Missouri 631412
Received 28 June 2006/ Accepted 13 September 2006
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
dotL lethality could be suppressed by inactivation of the Dot/Icm complex via mutations in other dot/icm genes. Here we report the isolation of non-dot/icm suppressors of this phenotype. These dotL suppressors include insertions that disrupt the function of the L. pneumophila homologs of cpxR, djlA, lysS, and two novel open reading frames, lpg0742 and lpg1594, that we have named ldsA and ldsB for lethality of dotL suppressor. In addition to suppressing dotL lethality, inactivation of these genes in a wild-type strain background causes a range of defects in L. pneumophila virulence traits, including intracellular growth, implicating these factors in the proper function of the Dot/Icm complex. Consistent with previous data showing a role for the cpx system in regulating expression of several dot/icm genes, the cpxR insertion mutant produced decreased levels of three Dot/Icm proteins, DotA, IcmV, and IcmW. The remaining four suppressors did not affect the steady-state levels of any Dot/Icm protein and are likely to represent the first identified factors necessary for assembly and/or activation of the Dot/Icm secretion complex. |
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Numerous cell biological processes have been implicated in the intracellular survival and replication of L. pneumophila, including modulation of both apoptosis and autophagy (1, 2). Although the role of these processes in L. pneumophila virulence is not fully understood, it is clear that alteration of the host cell endocytic pathway is critical to bacterial multiplication, as most avirulent L. pneumophila mutants are unable to replicate inside host cells because they mistarget and fail to form a replicative phagosome (5, 38). Complementation experiments utilizing these mutants revealed that L. pneumophila employs a specialized secretion system encoded by 26 "dot" or "icm" genes to prevent nascent L. pneumophila-containing phagosomes from entering into the endocytic pathway (39, 49).
The dot/icm genes encode an adapted conjugation apparatus that has been classified as a type IVB secretion system (T4SS) (10, 45). The L. pneumophila Dot/Icm T4SS is responsible for injecting a dozen or more bacterial proteins into the cytoplasm of host cells (4, 13, 14, 26, 30, 32). Although the specific biological functions of most Dot/Icm substrates have yet to be identified, many of these secreted proteins are expressed only in the early stationary phase of growth (4, 26). This may explain why the pathogen is not infectious during exponential growth and must differentiate into a transmissive form prior to infecting a new host cell (7).
Although critical to intracellular replication, most of the L. pneumophila dot/icm genes are not required for in vitro growth on bacteriological media (6, 40). However, it was recently reported that three dot genes, dotL, dotM, and dotN, are essential for the viability of L. pneumophila strain Lp02 (6). The dotL gene is also essential for the viability of an unrelated L. pneumophila strain, AA100, suggesting that this phenotype may be a conserved trait (C. D. Vincent and J. P. Vogel, unpublished data). Interestingly, it was discovered that the dotL lethality phenotype could be suppressed by mutations in the majority of the dot/icm genes (6). These observations led to the model that loss of dotL is lethal to L. pneumophila due to the accumulation of a toxic complex, perhaps an unregulated secretion pore, in the envelope of the bacterium (6). According to this model, suppression of dotL lethality could be achieved either directly or indirectly. For example, mutation of a component of the T4SS would directly disrupt the toxic secretion channel and would thus restore viability. Alternatively, the mechanism of suppression could be indirect, for example, by inactivation of genes needed for the proper assembly and/or activity of the secretion channel.
To test this, we performed a screen for transposon mutants that were able to survive in the absence of dotL (6). As predicted, the screen yielded a number of insertions in known dot/icm genes, including four in dotA, two in dotG, one in dotI, five in dotO, three in icmF, and one in icmX (6) and three in a new dot gene, dotV (J. A. Sexton et al., unpublished data). We report here the identification of non-dot/icm suppressor mutations, including disruptions of the L. pneumophila homologs of cpxR, lysS, djlA, and two novel genes, ldsA and ldsB. Although inactivation of each of these genes suppresses dotL lethality, their inactivation in the presence of wild-type DotL results in varying effects on L. pneumophila virulence. These findings indicate that multiple genes are likely to play key roles in regulating the assembly and/or activation of the Dot/Icm secretion complex.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
dotL lethality.
The assay used to screen for dotL lethality suppression has been described previously (6). Plasmid pJB1005 was transferred to L. pneumophila by using the RP4 conjugation system encoded on plasmid pRK600 (22). L. pneumophila strains that had integrated the plasmid onto the chromosome were selected by plating on medium containing streptomycin and kanamycin. The resulting dotL/dotL::Cmr merodiploid strains were plated on medium containing sucrose to select for resolution of the merodiploid to either dotL or dotL::Cmr. Sucrose-resistant colonies were streaked on medium containing kanamycin to confirm that the integrated plasmid had been lost and on medium containing chloramphenicol to screen for dotL or dotL::Cmr. Plasmid construction. All complementing clones were made in the RSF1010 cloning vector pJB908 (43). The cpxR cpxA clone (pJB3243) was generated by amplifying the cpxR-cpxA operon, using primers JVP1147 and JVP1148. The yitW lysS clone (pJB3197) was generated by amplifying the yitW-lysS operon, using primers JVP1159 and JVP1218. The lysS clone (pJB3198) was generated by amplifying lysS, using primers JVP1217 and JVP1218. The ldsA clone (pJB3411) was generated by amplifying ldsA, using primers JVP1229 and JVP1230. The djlA clone (pJB3230) was generated by amplifying djlA, using primers JVP1149 and JVP1150. To generate a djlA-complementing clone in a plasmid that could be transferred by the Dot/Icm complex, the djlA gene from pJB3230 was cloned into pJB1627, generating plasmid pJB3192. All constructs were sequenced to confirm that no errors were introduced during PCR amplification.
Growth of L. pneumophila strains in mouse bone marrow-derived macrophages. Mouse bone marrow-derived macrophages were isolated from the femurs of female A/J mice as described previously (47). Intracellular growth of L. pneumophila strains was assayed as described previously (4, 47).
Western analysis. Western blotting was performed using standard techniques (4). Whole-cell extracts were generated from L. pneumophila cells grown to early stationary phase in AYE. Western blotting was performed using antibodies specific to DotL (diluted 1:10,000), DotB (diluted 1:500), DotG (diluted 1:1,000), DotO (diluted 1:3,000), IcmR (diluted 1:3,000), IcmX (diluted 1:10,000), RalF (diluted 1:3,000), SdeC (diluted 1:3,000), or isocitrate dehydrogenase (ICDH) (diluted 1:10,000). Generation of DotL-, DotB-, and SdeC-specific antibodies has been described previously (4, 6, 43). Antibodies to DotG, IcmX, IcmR, and RalF were generated against purified amino-terminal His6 fusions injected into rabbits (Cocalico, Inc.). Antibody specific to ICDH was kindly provided by L. Sonenshein. Antibody specific to DotO was generously provided by R. Isberg. All antibodies were detected using goat anti-rabbit secondary antibody conjugated to horseradish peroxidase (Sigma) diluted 1:10,000 and an ECL detection kit (Amersham Biosciences).
Conjugation assay. Conjugation assays were performed as described previously (42). L. pneumophila strains were grown to early stationary phase, and 1 x 109 L. pneumophila donor cells were mixed with 1 x 1010 E. coli (strain ER1821) recipient cells. Cell mixtures were incubated for 2 h at 37°C on 0.45 µm-pore-size filters (Millipore) placed on CYE plus thymidine agar plates. Cells were resuspended in sterile water and plated on LB plus ampicillin (100 µg/ml) to determine the number of E. coli transconjugants. The conjugation frequency was determined by dividing the number of E. coli plasmid recipients by the number of L. pneumophila donors in each reaction. Shown are the averages and standard deviations for triplicate conjugation reactions.
Cytotoxicity assay. Contact-dependent cytotoxicity assays were performed as described previously (23, 46). L. pneumophila strains were added at the indicated multiplicities of infection (MOI) to mouse bone marrow macrophages plated on glass coverslips in 24-well tissue culture plates at a density of 1.5 x 105 bone marrow macrophages per coverslip. Bacteria were pelleted onto the macrophages at 168 x g for 5 min at room temperature and then incubated for 1 hour at 37°C in 5% CO2. The coverslips were then inverted onto a drop of phosphate-buffered saline (PBS) containing 25 µg/ml ethidium bromide and 5 µg/ml acridine orange (23) and were immediately observed. To calculate cytotoxicity, the number of ethidium bromide-positive macrophages was divided by the total number of macrophages observed in each field of view. Shown are the averages and standard deviations for four randomly selected fields of view scored for each strain at each MOI.
Intracellular targeting assay. Targeting assays were performed essentially as described previously (48). L. pneumophila cells grown to stationary phase were added to the macrophages at an MOI of approximately 5, and infections were allowed to proceed for 1 h at 37°C in 5% CO2. Infected macrophages were fixed by the addition of paraformaldehyde-lysine-periodate containing 5% sucrose, followed by incubation at room temperature for 20 min. Extracellular bacteria were stained with polyclonal serum to L. pneumophila (3, 49) diluted 1:10,000 in PBS containing 2% goat serum (PBSG), followed by goat anti-rabbit secondary antibody conjugated to Cascade Blue (Molecular Probes) diluted 1:10,000 in PBSG. Macrophages were then permeabilized with methanol, and intracellular bacteria were decorated with polyclonal serum to L. pneumophila diluted 1:10,000 in PBSG and goat anti-rabbit secondary antibody conjugated to Alexa Fluor 594 (Molecular Probes) diluted 1:10,000 in PBSG. LAMP-1 was labeled with the antibody ID4B, which was developed by J. Thomas August and was obtained from the Developmental Studies Hybridoma Bank, University of Iowa, Department of Biological Sciences, Iowa City. It was diluted 1:2.5 in PBSG and goat anti-rat secondary antibody conjugated to Alexa Fluor 488 (Molecular Probes) diluted 1:100 in PBSG. Only intracellular bacteria were scored for colocalization with the endocytic marker LAMP-1. The averages and standard deviations of numbers obtained from three sets of 100 intracellular bacteria are reported for each strain.
Immunofluorescence assay of SdeC secretion. Immunofluorescence staining of secreted SdeC was performed as described previously (4). Cells were stained with affinity-purified polyclonal serum raised to SdeC diluted 1:10, followed by incubation with goat anti-rabbit antibody conjugated to Oregon green (Molecular Probes) diluted 1:100. Bacterial and macrophage DNAs were stained with 4',6-diamidino-2-phenylindole (DAPI). For each strain, three counts of 100 bacteria were averaged. Results were normalized to wild-type secretion of SdeC set at 100%.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
dotL lethality can be suppressed by dot/icm and non-dot/icm mutations.
Although the dot/icm genes are required for the intracellular growth of L. pneumophila, most of these genes are dispensable for growth on bacteriological media (39, 40, 49). In contrast, dotL is essential for the viability of the L. pneumophila strain Lp02 under all growth conditions (6). We previously took advantage of dotL lethality to perform a screen for mutations that allow growth in the absence of dotL (6). The assay was based on resolution of a dotL/dotL::Cmr merodiploid strain, as observed in Fig. 1. Total resolution events can be detected by plating on CYE-sucrose plates, as the merodiploid cassette contains a counterselectable marker that confers sensitivity to sucrose. Specific resolution to the dotL::Cmr locus can be identified by growth of the strain on CYE plates containing sucrose and chloramphenicol. Therefore, by comparing growth of the merodiploid on the two types of media, it can be determined if a gene is essential or dispensable. The essentiality of dotL can be observed by the lowered plating efficiency of the wild-type dotL/dotL::Cmr merodiploid on CYE-sucrose plates containing chloramphenicol (Fig. 1D). Of the total resolution events (Fig. 1A), fewer than 1 in 1,000 resolved to the dotL::Cmr locus.
|
dotL::Cmr merodiploid strain on sucrose-chloramphenicol plates occasionally did result in growth of a few colonies. These colonies were confirmed to contain dotL deletions and were deduced to be alive because they contained an additional mutation(s) capable of suppressing dotL lethality (6). Based on the frequency at which these suppressors were observed, it seemed likely that there must be a large number of genes whose inactivation could alleviate the loss of dotL. This prediction proved true, as inactivation of almost any dot/icm gene, including dotA, is sufficient to suppress loss of dotL (Fig. 1B and E).
To identify additional suppressors, we previously reported a screen for mini-Tn10 insertions that allowed L. pneumophila to survive without the dotL gene (6). We identified 33 independent strains containing insertions that suppressed the loss of dotL. Nineteen of these strains possessed insertions in known dot/icm genes (6) or in dotV (Sexton et al., unpublished data). The remaining 14 suppressors contained insertions in open reading frames (ORFs) that did not resemble known dot/icm genes. These insertion mutations were able to suppress dotL lethality at a frequency similar to that of a dotA mutation (Fig. 1C and 1F). In summary, dotL is an essential gene, but loss of dotL can be suppressed by insertions in dot/icm genes and non-dot/icm genes.
Identification of non-dot/icm suppressors of dotL lethality.
We report the sites of transposon insertion for 7 of the 15 non-dot/icm suppressors (Fig. 2 and Table 2). These seven insertions are in five genes: the L. pneumophila homologs of cpxR, djlA, and yitW, as well as two previously uncharacterized L. pneumophila open reading frames, lpg0742 and lpg1594 (9). Based on their ability to suppress loss of dotL, we have named these genes ldsA and ldsB for lethality of dotL suppressor. We obtained a single insertion in ldsB, cpxR, and yitW and two independent insertions in ldsA and djlA (Fig. 2). However, it is worth noting that the original screen described by Buscher et al. was not performed to saturation (6).
|
|
Of the novel suppressors, ldsA codes for a 417-amino-acid protein that is predicted to be localized to the inner membrane via eight transmembrane domains (9). Although homologs to LdsA are conserved in other L. pneumophila strains, including Paris and Lens (Lpp0807 and Lpl0778, respectively) (8), there do not appear to be highly homologous ldsA genes in any other bacterial species, including the closely related pathogen Coxiella burnetii (41). Four ORFs with very limited homology to ldsA can be detected in GenBank, including SO0975, GSU1615, MCA2787, and AdehDRAFT_0196 from Shewanella, Geobacter, Methylococcus, and Anaeromyxobacter species, respectively. Each of these ORFs is predicted to encode an inner membrane protein with a size similar to that of LdsA, although none has been assigned a function.
The ldsB gene is also predicted to encode a small inner membrane protein with eight transmembrane domains. Similar to the case for ldsA, the L. pneumophila strains Lens and Paris contain highly conserved homologs to ldsB (lpl1431 and lpp1552, respectively). Most gram-negative bacteria, including C. burnetii, do not possess an LdsB homolog. The exception may be Burkholderia pseudomallei, which is predicted to encode a protein with limited homology to LdsB (BLAST score of 
1e–08 to LdsB). This B. pseudomallei hypothetical protein has been annotated as encoding a membrane-associated phospholipid phosphatase, although we could not detect this similarity and LdsB does not appear to resemble a phosphatase. In summary, LdsA and LdsB can be viewed as novel proteins that likely do not have true homologs outside of the L. pneumophila family.
Confirmation of suppressor phenotypes.
To confirm that the dotL lethality suppressor phenotype selected for in our screen was specifically linked to the transposon insertions we identified, the transposon and surrounding DNA were transformed into an unresolved dotL/dotL::Cmr merodiploid by using natural competence (44). The resulting strains were then assayed for their ability to suppress loss of dotL by resolving the strains on selective media as described above. In each case, resolution of merodiploids to the dotL::Cmr locus was compared to that of a dotA mutant merodiploid. As shown in Fig. 2, each of the seven insertions was able to suppress dotL lethality, with recovery rates ranging from 18% to 101% (normalized to a dotA mutant set at 100%). Although the percent recovery of dotL mutants varied, we observed that the insertions with lower recovery rates of dotL mutants also exhibited a corresponding diminished effect on Dot/Icm-related traits (see below). Significantly, no resolution events to the dotL::Cmr locus were recovered from the wild-type merodiploid (0/300 colonies scored). Thus, each newly created insertion mutant appears able to suppress loss of dotL, demonstrating linkage of this phenotype to the identified insertion.
The dotL suppressor mutants have varied effects on intracellular replication.
To determine the mechanism of dotL lethality suppression, we assayed the effects of the insertion mutations on the virulence properties of L. pneumophila. Since the suppressor mutations were isolated in a strain that lacked dotL, they were first transformed into a wild-type strain of L. pneumophila by natural competence (44). These strains were then grown in vitro to confirm that each resembled the wild-type strain in growth rate, colony morphology, and motility (data not shown). The mutants were then assayed for their ability to replicate within macrophages derived from the bone marrow of A/J mice (47). In these cells, the wild-type strain Lp02 was able to multiply approximately 125-fold over the course of the assay, whereas a dotA mutant was unable to replicate (Fig. 3).
|
dotL lethality suppressors vary significantly in their intracellular replication defects, and we observed a positive correlation between the severity of their attenuation and the level of recovery of dotL mutants. The growth defects of the ldsA and djlA insertions could be fully restored by introduction of the corresponding gene on a plasmid (pJB3411 for ldsA+ or pJB3230 for djlA+) (Fig. 3). Since cpxR is in an operon with cpxA, and the genes each encode an element of a two-component regulator whose components function together, we assayed only complementation of the cpxR insertion by introduction of a plasmid encoding the cpxR-cpxA operon. The cpxR-cpxA operon plasmid pJB3243 was able to fully restore growth to the cpxR insertion strain JV1807 (Fig. 3).
Likewise, yitW is found in an operon with a second gene, lysS. Therefore, the growth defect of the yitW insertion could be due to inactivation of yitW, of the downstream gene lysS, or of both genes. Similar to the case for the cpxR insertion, addition of a plasmid (pJB3197) containing both genes to the yitW insertion strain (JV1485) was able to fully restore growth (data not shown), suggesting that both genes might be involved. However, JV1485 could also be fully complemented by introduction of just the lysS gene (pJB3198) (Fig. 3), leading to the conclusion that the phenotype of the yitW transposon insertion must be due to polar effects on expression of lysS. Complementation of the ldsB mutant (JV1484) was not performed, as this strain exhibited no intracellular growth defect. As a result, we can conclude that ldsB is dispensable for intracellular replication, the cpxR-cpxA operon and lysS are partially required, and ldsA and djlA are necessary. The requirement for djlA closely resembled that of most dot/icm genes.
The cpxR mutation, but not the other suppressor mutations, affects expression of the Dot/Icm apparatus proteins.
Based on the observation that inactivation of various dot/icm genes is able to suppress the lethality associated with loss of dotL (6), we hypothesized that these non-dot/icm suppressors might affect the expression and/or the stability of components of the Dot/Icm type IV secretion machinery. For example, it is well established for other bacterial species that cpxR/cpxA mutations can affect the expression of many proteins (34, 37). To determine if any of the dotL suppressor mutations affected Dot/Icm protein levels, we performed Western blot analysis on whole-cell lysates by using antibodies to 20 Dot/Icm proteins and 2 secreted substrates. Six representative Dot/Icm proteins and the two secreted substrates, RalF and SdeC, are shown in Fig. 4A. An antibody that recognized the constitutively expressed housekeeping protein ICDH was included as a control. None of the suppressor mutations significantly affected expression of the majority of the Dot/Icm apparatus proteins or of the secreted substrates in either the wild-type or dotL strain background in either the exponential or stationary phase of growth (Fig. 4A and data not shown).
|
dotL lethality (6), we hypothesized that decreased expression of IcmV and DotA was the mechanism of dotL lethality suppression of the cpxR mutant. To test this, we transformed a plasmid containing the icmV dotA operon under a constitutive promoter or the corresponding vector into the cpxR::Kanr dotL mutant. Although transformants were obtained with the vector, no transformants were recovered when the icmV dotA plasmid was transformed into this strain (data not shown). This result was dependent on inactivation of both dotL and cpxR, as transformants could be obtained in the cpxR::Kanr mutant background. Thus, the cpxR::Kanr mutant appears to suppress dotL lethality through decreased levels of IcmV and DotA. However, the virulence defects of the remaining suppressor mutants do not appear to be due to altered expression or stability of the Dot/Icm T4SS but may instead be due to the improper assembly or activation of the secretion apparatus.
The dotL suppressor mutants have varied effects on plasmid mobilization.
Having eliminated the most likely explanation for the dotL suppressor phenotype, we employed a number of Dot/Icm-dependent assays to test the dotL suppressor mutants for Dot/Icm T4SS functionality. The first assay we used is based on the ability of the L. pneumophila Dot/Icm T4SS to transfer a plasmid from one bacterium to another (39, 49). A wild-type L. pneumophila strain, JV4157, mobilized pKB5 with an efficiency of 8.1 x 10–6 transconjugants per donor, whereas a dotA mutant (JV4158) was completely deficient for plasmid transfer (limit of detection = 1 x 10–9) (Fig. 5A).
|
The dotL suppressor mutants have varied effects on contact-dependent cytotoxicity.
As an additional measure of Dot/Icm T4SS functionality, we assayed the ability of the dotL suppressor mutants to cause rapid cytotoxicity of host macrophages. This form of cytotoxicity, also termed "contact-dependent cytotoxicity," is observed at high MOI and results in host cell death consistent with the insertion of pores in the host cell membrane (23). Rapid contact-dependent cytotoxicity requires the majority of the Dot/Icm proteins, excluding the type IV adaptors IcmS and IcmW, and therefore accurately reflects the presence of an assembled and functional T4SS (12, 23).
To assay cytotoxicity, L. pneumophila strains were pelleted onto mouse bone marrow-derived macrophages at a range of MOI. After incubation for 1 hour, the infected macrophages were stained with acridine orange and ethidium bromide in order to determine their intactness. Live, intact cells stain green, whereas permeabilized cells stain red due to the uptake of ethidium bromide (23). Representative images of the cytotoxic effects of a wild-type strain, a dotA mutant, and the dotL suppressor mutant strains are shown in Fig. 6A to H. Uninfected cells are green, as they are impermeable to ethidium bromide (Fig. 6A). Infection using a wild-type strain at an MOI of 50 bacteria/macrophage resulted in the majority of the macrophages being permeabilized (Fig. 6B). Infection using a dotA mutant at an identical MOI caused no cytotoxicity (Fig. 6C). Macrophages infected with the ldsB mutant (Fig. 6D), the cpxR cpxA mutant (Fig. 6E), and the lysS mutant (Fig. 6F) closely resembled those infected with wild-type cells (Fig. 6B). In contrast, the ldsA mutant (Fig. 6G) and the djlA mutant (Fig. 6H) were significantly less cytotoxic but did not appear to be as defective as the dotA mutant (compare Fig. 6G and H with Fig. 6C).
|
The ldsA and djlA mutants are defective in altering the endocytic pathway of host cells. Since the ldsA and djlA mutants were the only suppressors that exhibited severe defects in dot/icm-related phenotypes, we examined the intracellular growth defect of these mutants further. It is possible that these mutants are impaired for growth within macrophages simply because they replicate at lower rates inside host cells. Alternatively, they could lack a function that is necessary for intracellular replication. For example, they could be unable to alter intracellular trafficking of the host cell, a process that absolutely requires a functional Dot/Icm type IV secretion system. To assess L. pneumophila avoidance of the endocytic pathway, we quantitated colocalization of Legionella-containing phagosomes with the endocytic marker LAMP-1 (48). As shown in Fig. 7, the majority of wild-type L. pneumophila bacteria did not colocalize with LAMP-1, whereas only 10% of dotA mutants were LAMP-1 negative. Both the ldsA mutant and the djlA mutant resembled the dotA mutant in that they were severely defective in altering the endocytic pathway. The ldsA mutant was slightly less attenuated than the dotA mutant for avoiding the endocytic marker LAMP-1 (17% LAMP-1 negative), while the djlA mutant was indistinguishable from the dotA mutant (8% LAMP-1 negative). As observed with intracellular growth, the targeting defect of each of the mutants was fully restored by introduction of complementing clones (86% and 73% LAMP-1 avoidance, respectively). Because the severity of the intracellular growth defect correlated with the defect in targeting for both mutants, it seems likely that the observed defects in intracellular growth are caused solely by the inability to alter the host endocytic pathway.
|
To assay SdeC export by the ldsA and djlA mutants, the strains were added to mouse bone marrow-derived macrophages, and infections were allowed to proceed for 30 min prior to fixation and staining. DAPI (blue) was used to stain macrophage and bacterial DNAs, whereas the SdeC antibody was decorated with a goat anti-rabbit antibody conjugated to Oregon green. Representative images of each strain are shown in Fig. 8A, with quantitation provided in Fig. 8B. As described previously (4), wild-type cells (JV1139) frequently exhibit foci of SdeC staining at one or both poles of the phagosome, whereas dotA mutants (JV1141) do not. Consistent with the defect in altering the endocytic pathway as observed by LAMP-1 colocalization, the djlA mutant (JV4077) was almost completely defective for SdeC secretion (Fig. 8A and B). The ldsA mutant was also defective for SdeC export, although not to the same extent (data not shown). Secretion of SdeC was fully restored in each mutant by the introduction of the corresponding complementing clone (Fig. 8 and data not shown). Thus, the intracellular growth defect of the dotL suppressor mutants ldsA and djlA appears to be due to a failure to properly export T4SS substrates and therefore a failure to alter the endocytic pathway of the host cell.
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
dotL suppressor). Each of these suppressor mutations conferred the ability of L. pneumophila to survive in the absence of dotL. When assayed in a dotL+ background, these mutants exhibited differential effects on Dot/Icm-dependent assays. These results support the hypothesis that there are multiple means to suppress dotL lethality, including by decreasing expression of dot/icm genes and by diminishing the assembly and/or activity of the L. pneumophila Dot/Icm secretion system. The CpxRA (conjugative plasmid expression) stress response system is a two-component regulatory system induced in response to extracytoplasmic signals generated in the cell envelope, principally misfolded proteins in the periplasm (34, 37). Once these signals are detected, the system is responsible for activating the expression of a number of factors required for responding to cell envelope stress (35). In addition to a role in F plasmid conjugation (28), the Cpx system has been implicated in several virulence mechanisms, including invasion (Salmonella enterica), transcriptional regulation (Shigella sonnei), P-pilus biogenesis (uropathogenic E. coli), and assembly of type IV bundle-forming pili (enteropathogenic E. coli) (reviewed in reference 34).
Based on our model of dotL toxicity and the important role of the Cpx system in responding to periplasmic stress in other organisms, it is not surprising that this system is involved in the response to loss of dotL. We hypothesized that the cpx mutation could affect expression of dot/icm genes, thereby suppressing dotL lethality due to lowered levels of Dot/Icm proteins. This hypothesis was supported by a previous study, using ß-galactosidase reporters, that showed decreased transcription of icmR, icmV, and icmW in a cpxR mutant (16). Although we did not observe that IcmR protein levels were significantly affected by a cpxR insertion (Fig. 4A), we did observe an effect on DotA, IcmV, and IcmW. Interestingly, the cpxR::Kanr mutant had only a minor effect on the levels of IcmX (Fig. 4A), a protein encoded downstream of icmW. By assaying the effects of overexpression of IcmV and DotA, we were able to demonstrate that decreased levels of IcmV and DotA were solely responsible for suppression of dotL lethality in the cpxR::Kanr mutant. This is consistent with our previous observation that disruption of icmV or dotA, but not icmW, suppresses dotL lethality (6). Surprisingly, although decreased IcmV and DotA levels were sufficient to suppress loss of dotL in the cpxR::Kanr mutant, it was still able to conjugate and remained cytotoxic. Thus, the dotL lethality phenotype is more sensitive to decreased levels of Dot/Icm proteins than other Dot/Icm-associated traits.
The remaining non-dot/icm suppressors described here likely represent mutations that indirectly affect the activity of the L. pneumophila T4SS or mutations that do not affect the Dot/Icm complex itself but allow the strain to survive in the absence of dotL. These mutations can be subdivided into three phenotypic classes: (i) mutations that affect virulence by disrupting the assembly and/or the activation of the Dot/Icm system (e.g., ldsA and djlA), (ii) mutations that negatively affect the ability of the strain to grow in macrophages but do not appear to grossly affect the assembly and function of the Dot/Icm complex (e.g., lysS), and (iii) mutations that do not affect any of the virulence-related assays (e.g., ldsB). Before examining the overall implications of these suppressors, we will discuss each class separately.
The first class of mutants reported here includes the ldsA and djlA mutants. Each of these mutants was much more attenuated in intracellular growth than the other suppressor mutants, and the djlA mutant was completely defective for growth in macrophages. Further analysis revealed that the reason for the attenuation of these mutants is that they likely affect the assembly and/or activity of the Dot/Icm apparatus. Although these mutants still make wild-type levels of Dot/Icm proteins and substrates, they are less efficient than the wild type at avoiding phagosome-lysosome fusion and exhibit decreased Dot/Icm-mediated plasmid transfer and contact-dependent cytotoxicity. Finally, both mutants are severely attenuated for the secretion of a T4SS substrate, SdeC, into the cytoplasm of macrophages. Thus, as predicted, our dotL lethality suppressor screen did identify factors involved in the assembly or activity of the Dot/Icm apparatus.
The ldsA gene encodes a polytopic inner membrane protein that does not contain any distinguishing motifs, making prediction of its molecular function difficult. Based on the protein's putative membrane localization and the Dot/Icm-related defects associated with loss of LdsA, it is formally possible that LdsA is an additional Dot/Icm protein. However, we do not favor this possibility since LdsA lacks homology to a component of any known type IV secretion system and is encoded distantly from the two Dot/Icm pathogenicity islands (49).
In contrast to the ldsA mutant, it is easier to propose an explanation for the djlA mutant. DjlA (for DnaJ-like protein) is an inner membrane-anchored homolog of DnaJ with its J domain located in the cytoplasm (11). Similarly to DnaJ, E. coli DjlA is known to interact with the DnaK (Hsp70) chaperone and stimulate its ATPase activity (50). DjlA has been proposed to function as a chaperone for the assembly and/or activity of membrane proteins and may play a role as a sensor of envelope stress (21). The involvement of a stress response system in the suppression of dotL lethality is consistent with our model that the absence of DotL leads to the accumulation of a toxic subcomplex in the envelope of cells. Based on the proposed activities of E. coli DjlA, the highly conserved L. pneumophila homolog is likely to be involved in sensing misfolded proteins in the membrane and assisting in their proper folding. Therefore, L. pneumophila may require DjlA to properly fold and assemble components of the Dot/Icm complex. Loss of djlA would therefore phenocopy dot/icm mutations and suppress dotL lethality by a similar mechanism. Additionally, djlA has also been implicated as being required for the intracellular growth of another Legionella species, Legionella dumoffii (33). Because our djlA mutant still appears to synthesize components of the Dot/Icm apparatus and yet is severely defective for all Dot/Icm-dependent traits, it likely serves as a key factor in the assembly/quality control of the L. pneumophila T4SS. To our knowledge, DjlA is the first such factor identified.
The second class of mutations reported here includes an insertion in yitW that affects the expression of the lysyl-tRNA synthetase gene lysS. This mutation caused subtle defects in intracellular growth in mouse bone marrow-derived macrophages. However, the mutant did not display a significant defect in plasmid transfer or contact-dependent cytotoxicity, traits requiring a functional Dot/Icm secretion apparatus. LysS is an aminoacyl-tRNA synthetase that catalyzes the formation of Lys-tRNALys, which is then used to insert lysine into proteins (31). It was surprising to obtain an insertion that appeared to down-regulate the expression of lysS, since most tRNA synthetases are essential. However, it is possible that the yitW::Kanr allele exhibited only an incomplete polar effect on lysS, thereby partially down-regulating its expression. Alternatively, lysS may not be essential because Legionella possesses a homolog of lysU, which encodes a secondary, inducible form of lysyl-tRNA synthetase in other bacterial species (31). Although the molecular mechanism of suppression by inactivation of lysS remains cryptic, the Lys-tRNALys molecule has been implicated in additional roles, including functioning as a signaling molecule (31). It is thus possible that the mechanism of lysS suppression of dotL lethality is indirect and may be unlinked to its role in protein synthesis.
The insertion in ldsB was the only mutant in the third class identified in this screen whose inactivation suppressed dotL lethality but did not affect the virulence of a dotL+ strain of L. pneumophila. LdsB is predicted to be a novel polytopic, inner membrane protein that does not possess any apparent protein motifs, thus precluding an obvious prediction of function. Although we do not understand the molecular mechanism of ldsB suppression of dotL lethality, the ability of this mutant to replicate normally in macrophages suggests that the suppression may be mediated by an indirect mechanism of action (see below). Nevertheless, the identification of ldsB is important for two reasons. First, since a dotL strain is not viable in the Lp02 background, it was previously only possible to examine this strain lacking dotL in the presence of another dot/icm mutation, e.g., dotL dotA, thus severely complicating the analysis of the dotL phenotype. As a result of this discovery, it is now possible to examine the dot/icm phenotypes of Lp02 lacking dotL in the dotL ldsB strain. Second, the existence of a dotL lethality suppressor that does not affect the intracellular replication of L. pneumophila provides credence to our proposal that the JR32 dotL strain is viable because the JR32 strain background contains a suppressor mutation (6).
Based on our initial observation that dot/icm mutations could suppress loss of dotL, we proposed that an L. pneumophila dotL strain is not viable because it accumulates a toxic substrate in the cell envelope. This poisonous structure could be a misfolded Dot/Icm subcomplex, similar to pilin subunits that accumulate in the absence of the P-pilus chaperone PilD (25). Alternatively, the toxic substrate might be a normally assembled Dot/Icm subcomplex that functions improperly in the absence of DotL, perhaps as an unregulated secretion pore that leads to a loss of homeostasis. The former model implies that accumulation of a subcomplex, likely to consist of a few Dot/Icm proteins, is responsible for the toxicity. If this were the case, inactivation of only a few dot/icm genes should suppress dotL lethality. Instead, inactivation of almost any dot/icm gene suppresses loss of dotL, which is more consistent with the toxicity being due to an improperly functioning Dot/Icm complex.
Taken in this context, inactivation of factors that are required for the proper assembly or activity of the Dot/Icm complex, such as DjlA, should also suppress dotL lethality. However, the existence of suppressors that replicate in macrophages and have only subtle or no effects on the Dot/Icm complex in a dotL+ background is perplexing. We propose two possible explanations for this conundrum. First, the number of functional Dot/Icm complexes required for intracellular growth may be significantly less than the number of toxic dotL complexes required for cell death. In this case, a mutation that decreases the number of functional complexes may be able to suppress dotL lethality but have no effect on intracellular growth in a wild-type background. Alternatively, one could imagine a type of mutation that induces a response capable of allowing a dotL strain to live that is not needed for the normal assembly/activity of the complex. For example, inactivation of a repressor of a periplasmic protease might allow the elimination of a toxic Dot/Icm subcomplex but yet not be required in a wild-type cell. In either case, further examination of the dotL lethality suppressors should reveal insights into both the normal assembly/activity of the T4SS and envelope stress response systems in L. pneumophila.
|
ACKNOWLEDGMENTS |
|---|
This work was funded by NIH grant AI48052 to J.P.V.
|
FOOTNOTES |
|---|
Published ahead of print on 22 September 2006.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
