Logo of jbacterPermissionsJournals.ASM.orgJournalJB ArticleJournal InfoAuthorsReviewers
J Bacteriol. 2005 Nov; 187(22): 7716–7726.
PMCID: PMC1280299

Evidence for Acquisition of Legionella Type IV Secretion Substrates via Interdomain Horizontal Gene Transfer


Intracellular pathogens exploit host cell functions to create a replication niche inside eukaryotic cells. The causative agent of Legionnaires' disease, the γ-proteobacterium Legionella pneumophila, resides and replicates within a modified vacuole of protozoan and mammalian cells. L. pneumophila translocates effector proteins into host cells through the Icm-Dot complex, a specialized type IVB secretion system that is required for intracellular growth. To find out if some effector proteins may have been acquired through interdomain horizontal gene transfer (HGT), we performed a bioinformatic screen that searched for eukaryotic motifs in all open reading frames of the L. pneumophila Philadelphia-1 genome. We found 44 uncharacterized genes with many distinct eukaryotic motifs. Most of these genes contain G+C biases compared to other L. pneumophila genes, supporting the theory that they were acquired through HGT. Furthermore, we found that several of them are expressed and up-regulated in stationary phase in an RpoS-dependent manner. In addition, at least seven of these gene products are translocated into host cells via the Icm-Dot complex, confirming their role in the intracellular environment. Reminiscent of the case with most Icm-Dot substrates, most of the strains containing mutations in these genes grew comparably to the parent strain intracellularly. Our findings suggest that in L. pneumophila, interdomain HGT may have been a major mechanism for the acquisition of determinants of infection.

The γ-proteobacterium Legionella pneumophila is an opportunistic human pathogen that multiplies within alveolar macrophages and causes the nosocomial and community-acquired pneumonia known as Legionnaires' disease (18, 25, 48). Human disease occurs when aerosolized L. pneumophila is inhaled from man-made or natural freshwater reservoirs harboring the bacteria. L. pneumophila poses a significant worldwide public health problem, particularly for individuals with compromised immune systems (19, 38, 40). Eradication of the pathogen from freshwater, industrial settings has proven difficult, since L. pneumophila thrives in environments that exclude antibacterial agents, such as biofilms and the intracellular compartments of protozoa (8, 46, 47).

In order to create a replicative niche inside eukaryotic cells suitable for replication, L. pneumophila is believed to modulate host cell functions by the delivery of effector proteins through a type IVB secretion system known as the Icm-Dot complex (52, 59). Effector proteins presumably regulate several pathways in the host, including up-regulation of phagocytosis (23), delay in phagosome-lysosome fusion (24), recruitment of ARF1 to the phagosome (43), acquisition of endoplasmic reticulum-derived vesicles (29), and nonlytic egress from the host cell (11).

Several laboratories found Icm-Dot substrates through genetic screens and bioinformatic approaches (3, 7, 11, 13, 36, 43, 44, 57). Most of the known effector proteins are not individually required for intracellular multiplication, since knocking out the corresponding genes does not result in growth defects within protozoa or macrophages under laboratory conditions (except if the effector is required for the integrity of the translocation apparatus itself, which is the case for LidA).

Two hypotheses have been proposed to explain the origins of virulence determinants in bacterial pathogens: convergent evolution and horizontal gene transfer (HGT) (58). In the former, the pathogen's own genetic material is modified to evolve a protein that has a function similar to that of a host protein. In the latter, the pathogen acquires foreign genetic material from eukaryotic cells that is incorporated into its genome and retains at least part of its original activity.

Of the known L. pneumophila effector proteins, it has been suggested that at least one (RalF) has been acquired through interdomain HGT (43). Structural studies have confirmed that the three-dimensional structure of this protein resembles the well-known eukaryotic Sec7 domain fold (1). However, it is currently unknown if interdomain HGT is a major mechanism through which L. pneumophila and other pathogens acquire effector proteins.

We decided to take advantage of the recently completed genome sequence of L. pneumophila Philadelphia-1 (12) to identify genes that may have arisen via HGT from eukaryotes to bacteria. By looking for domains of eukaryotic origin in all open reading frames of the genome, we identified, in addition to several previously identified Icm-Dot substrates (see Table Table4),4), 46 uncharacterized genes (see Table Table3).3). Two of these 46 genes were subsequently studied and found to be effector proteins by Roy and colleagues while the manuscript was in preparation (7). In the course of analyzing the novel genes similar to those in eukaryotes (referred to as “eukaryotic-like genes”), two other isolates of L. pneumophila (Paris and Lens) were sequenced, annotated, and found to have some of the same eukaryotic-like genes and/or different genes with similar domains (9) (see Table Table5).5). However, Buchrieser and colleagues highlighted in their genome study (9) only 26 of the 46 eukaryotic-like genes described here, showing that genome annotation alone is not sufficient to find all genes with eukaryotic domains in bacterial genomes. Additionally, no experimental evidence was provided to show that the eukaryotic-like genes from the Paris and Lens isolates were expressed or involved in host-pathogen interactions.

Summary of Legionella eukaryotic-like genes identified in this study
Summary of previously characterized genes that were also identified in this study's bioinformatic screen
Domain and gene conservation in different Legionella pneumophila isolates

Here, we provide evidence supporting the theory that the Legionella eukaryotic-like genes (leg) were acquired through interdomain HGT. Furthermore, we show that several of these genes are in fact expressed in L. pneumophila and that their expression is up-regulated in stationary phase in an RpoS-dependent manner. We also found that seven of the Leg proteins, none of which were described previously or were highlighted in genome studies, are translocated into host cells via the Icm-Dot complex. Finally, we analyzed the abilities of 20 strains containing transposon insertions in the leg genes to grow intracellularly. Reminiscent of strains containing knockouts in other Legionella effector proteins, we found that none of the leg genes tested here are required for intracellular growth.


Plasmid constructions.

Plasmids and strains used in the study are summarized in Table Table1.1. Plasmids were constructed essentially as described elsewhere (11) or using standard molecular biology techniques. To create CyaA hybrid proteins, candidate leg genes were amplified by PCR using Proofstart polymerase (QIAGEN, California) with primers that contained unique restriction sites at the ends. The PCR products were digested with either KpnI/XbaI or BamHI. These fragments were then ligated into the KpnI/XbaI or BamHI sites of the previously described pJC141 vector (11). For complementation studies, the same fragments were ligated into the KpnI/XbaI or BamHI sites of the pMMB207C (11) vector. For allelic exchange, we amplified 2,000 bp upstream of legLC8 using Proofstart polymerase (QIAGEN). The PCR product was digested with EcoRV/BamHI. This fragment was then ligated into the EcoRV/BamHI sites of pLAW344 to generate pKS80 (61). We then PCR amplified 2,398 bp downstream of the legLC8 gene and digested the product with BamHI/NotI. This fragment was ligated into the BamHI/NotI sites of pKS80 to generate pKS81. We then amplified the kanamycin resistance cassette of Tn903dIIlacZ and digested the product with BamHI. The kanamycin resistance cassette was ligated into the BamHI site of pKS81 to generate pKS82. Primer sequences are available upon request.

Plasmids used in this study

Strain constructions.

Transposon-mutagenized strains are listed in Table Table22 along with other relevant strains used in this study. The ordered transposon library and pools are described elsewhere (49). By performing PCR in the larger pools (ca. 600 mutants) with a primer annealing to the transposon and one annealing to the upstream region of the gene of interest, we were able to determine, based on the sizes of the PCR products, if there was a transposon insertion in the gene of interest in one of the large pools. If the result was positive, we then performed PCR in the smaller pools representing single microtiter plates (∼60 mutants) to identify in what plate our clone of interest was located. Finally, we screened the individual plate by the same PCR technique to identify the mutant of interest. PCRs were performed using Taq polymerase (New England Biolabs, Massachusetts) with the following protocol: denaturation at 95°C for 5 min, followed by 35 cycles of denaturation at 95°C for 10 s, annealing at 68°C for 30 s, and amplification at 72°C for 3 min. The PCR products were then sequenced to confirm that the transposon insertion was in fact located in the gene of interest. Primers are available upon request.

Strains used in this study

We performed allelic exchange as described previously (61) to generate the isogenic legLC8 strain KS83.

Sequence analysis.

PSI-BLAST searches were performed using default parameters in the nonredundant database as of April 2005. Values for overall G+C content and G+C content in codon positions 1, 2, and 3 of coding regions were calculated by SeqStat, a C program (28). The program was modified to process a large number of coding regions in FASTA format for analysis of the Legionella pneumophila Philadelphia-1 genome.

RNA extraction, reverse transcription, and real-time PCR.

The L. pneumophila strain JR32 (a streptomycin-resistant, restriction-defective strain of the Philadelphia-1 isolate) and an isogenic rpoS null strain, LM1376, were grown in AYE medium (16) on a shaker at 300 rpm at 37°C. Samples of bacterial culture were removed during logarithmic and stationary phases. Total RNA was isolated using the RNeasy kit (QIAGEN). Reverse transcription was performed using 10 μg of RNA with Superscript II Plus RNase H- reverse transcriptase (RT) (Invitrogen, California) and random primers. After alkaline hydrolysis of source RNA with 1 N NaOH, target cDNA was purified using the QiaQuick PCR purification kit (QIAGEN). Real-time PCR was performed using the SYBR Green PCR master mix (Applied Biosystems, California) on an ABI PRISM 7000 (Applied Biosystems) using 25 pmol of each of the gene-specific primers (available upon request). The following protocol was used: denaturation at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 15 s and annealing-extension at 67°C for 30 s. For the amplification reaction we used 100 pg of cDNA from JR32 and LM1376 at log and stationary phases. The amplification levels were determined with the ABI PRISM 7000 SDS software (Applied Biosystems). The following equation was used to calculate fold induction in the RT-PCR experiments: Ratio = equation M1, where ΔCT = CT sample1CT sample2.

The CT value represents the cycle at time (T) in which amplification occurs.

Translocation assays.

Translocation assays were performed essentially as described elsewhere (11), except that Dictyostelium discoideum was used as a host. Briefly, Legionella pneumophila strains carrying the leg-cyaA hybrids and controls were grown overnight to stationary phase with the appropriate antibiotics. The bacteria were then subcultured into fresh AYE medium (16) to a final optical density (OD) of 0.08 and grown with shaking at 37°C until they reached an OD of 0.2. Isopropyl-beta-d-thiogalactopyranoside (IPTG) was then added to a final concentration of 1 mM to induce expression of the hybrid proteins, which are under the control of the Ptac promoter. The cultures were then incubated at 37°C for an additional 3 h. Dictyostelium discoideum DH1 cells (4 × 105) in MB medium {20 mM MES [2-(N-morpholino)ethanesulfonic acid] (pH 6.9), 0.7% yeast extract, 1.4% BBL thistone E peptone} containing IPTG to a final concentration of 1 mM were seeded into the wells of 96-well plates and incubated at 25°C for several hours. After IPTG induction, the bacterial cultures were spun and resuspended in MB medium containing 1 mM IPTG to a final OD of 0.2. The Dictyostelium discoideum cells were then infected with 2 × 107 L. pneumophila cells (multiplicity of infection = 50) containing the different hybrid proteins or appropriate controls. The microtiter plate was then spun at 2,000 rpm for 15 min at 25°C to synchronize the infection. The plate was then incubated for 15 h at 25°C. The supernatant was removed, and the whole contents of the well were lysed with 100 μl of ice-cold 2.5% perchloric acid. A 0.4 M concentration of Trizma base was then added in appropriate amounts to neutralize the contents of the well. cAMP was then measured by using the cAMP ELISA Biotrak kit (Amersham, New Jersey) following the instructions of the manufacturer.

Intracellular Growth Assays in Eukaryotic Hosts.

Intracellular growth curves in Acanthamoeba castellanii and terminally differentiated U937 cells were performed essentially as described elsewhere (11) with an multiplicity of infection of 0.001 (4 × 106 host cells and 4 × 103 bacterial cells) in 24-well tissue culture plates. Briefly, infections were synchronized by spinning the bacteria at 2,000 rpm for 10 min onto the adhered host cells. After 1 h of incubation, the wells were washed once to remove host cells that were not adhered and bacterial cells that were not internalized. Fifty-microliter aliquots were removed every 24 h, serially diluted in sterile, distilled water, and plated in charcoal-yeast extract plates (16).


Genomewide screen for L. pneumophila eukaryotic-like genes.

To identify L. pneumophila genes containing “eukaryotic domains,” we performed computational screens using the web-based programs PSI-BLAST, Pfam, SMART, and COILS (Fig. (Fig.1).1). We scanned every open reading frame (ORF) of the L. pneumophila Philadelphia-1 genome (12) for those that contain specific sequences at the amino acid level that can be classified as “eukaryotic domains.” We consider “eukaryotic domains” to be protein motifs that are widespread in eukaryotic species and significantly unrepresented in archaeal and prokaryotic species (45) and that have known cellular functions associated with eukaryotes. We also used PSI-BLAST to confirm that the full-length ORFs are not widespread in other bacterial species. The Legionella eukaryotic-like genes identified by this screen are described in Table Table3.3. Confirming the validity of the approach as a tool to find putative effector proteins, we identified several previously characterized Icm-Dot substrates that also have eukaryotic-like domains (Table (Table4).4). The genes identified in this bioinformatics screen include approximately 60% of the previously confirmed Icm-Dot substrates. The expect values listed in Table Table44 pertain to the alignment of the domain regions only against the current consensus sequences available in the conserved domains database. These expect values for the domain regions ranged from 5 × 10−4 to 6 × 10−50. When selecting hits containing coiled-coil domains, we took into consideration only ORFs that had regions of predicted coiled coils for stretches larger than 50 residues and that did not resemble known bacterial coiled-coil proteins based on sequence alignments. Since the current number of complete eukaryotic genomic sequences available is relatively small, it is difficult to predict the flow of horizontal gene transfer by analyzing the gene of the organism to which a given Legionella eukaryotic-like gene aligns best. Also, we found that while the regions pertaining to the predicted eukaryotic domains in the leg genes align significantly with the consensus sequences of the domains and with the sequences of several eukaryotic genes containing the same types of domains, the remaining parts of most of the genes do not have any clear orthologs in the database (data not shown). This may suggest that a certain degree of amelioration may have occurred since the gene was first acquired or that closer orthologs will be identified only as many more eukaryotic genomes are sequenced. During this work, two other L. pneumophila isolates were fully sequenced and annotated (9, 12). This genome study highlighted only 26 out of the 46 genes described here (Table (Table5).5). This argues strongly that gene annotation alone greatly underestimates the number of eukaryotic-like genes in bacterial genomes.

FIG. 1.
Flow chart of the bioinformatic screen performed to identify Legionella eukaryotic-like genes.

leg gene domains.

A variety of domains were found in the leg genes. Most of these correspond to protein binding motifs commonly found in eukaryotes. Several are predicted to encode products with enzymatic activities related to eukaryotic cell functions. A few, however, are not fully characterized even in eukaryotes. Following is a brief description of the domains and their prevalence in bacterial pathogens.

Ankyrin repeats.

The gene products LegA1 to -15 are predicted to contain one or more ankyrin repeats, which are interaction domains usually found in eukaryotic proteins that play a role in the attachment of cytoskeletal components (41). Other bacterial pathogens are also known to carry genes with this type of domain (10, 39), including the closely related intracellular bacterial pathogen Coxiella burnetti (the causative agent of Q fever), which contains 13 genes with this type of domain in its genome (55), one of which is conserved in L. pneumophila (legA3 and CBU1292; E = 1 × 10−93). Also, the parasitic endosymbiont Wolbachia pipientis contains several proteins with ankyrin repeats speculated to play a role in its intracellular lifestyle (26, 62).

Leucine-rich repeats.

Leucine-rich repeat domains, which usually form protein-binding interfaces (32), were identified in eight of the leg genes (legL1 to LC8). Eukaryotic proteins with these domains participate in a wide range of functions. Leucine-rich repeat proteins have also been identified in other bacterial pathogens (30), such as Salmonella enterica serovar Typhimurium, Yersinia pestis, Ralstonia solanacearum, Shigella flexneri, and the obligate amoebal symbiont Parachlamydia. However, leucine-rich repeat-containing proteins are rarely found in nonpathogenic bacteria but are present in nearly all sequenced eukaryotic species (45).

Coiled coils.

Extended regions of coiled coils usually form protein interaction domains that can participate in multiple cell functions, including the formation of fusogenic complexes through SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) proteins (6). Several of the currently known Icm-Dot substrates possess eukaryotic-like coiled-coil regions (11, 36). legC1 to -C8, legLC4, and legLC8 encode products with coiled-coil regions.

Ser/Thr kinases.

The L. pneumophila genome contains three ORFs predicted to encode Ser/Thr kinases (legK1 to -K3). The presence of kinase effector proteins in bacterial pathogens whose substrates are host cell proteins is not unprecedented (58). LegK1 and -K3 contain the conserved glycine loop motif and a catalytic aspartate preceded by an arginine—typical features of Ser/Thr kinases. Besides its enzymatic active site, LegK1 has another interesting property: its C terminus is similar to a region in the Salmonella effector protein SseG previously shown to contain a golgi localization domain (50).

Signaling lipid-related domains.

The legS2 gene contains an active site predicted to modify the eukaryotic signaling lipid sphingosine 1-phosphate. Sphingosine 1-phosphate can cause the mobilization of Ca2+ ions from the endoplasmic reticulum, which in turn activates pathways important for phagocytosis and proper phagosome maturation (37). LegS1 has a predicted lipid phosphoesterase active site that aligns weakly with sphingomyelinases. To our knowledge, similar active sites have not been identified in other prokaryotes.

GDP-GTP exchange.

Two of the leg genes (legG1 and legG2) are believed to be involved in GDP-to-GTP exchange in GTP-binding proteins. The gene legG1 has an RCC1 domain which can act as a guanine-nucleotide dissociation stimulator for the Ran GTP-binding protein (54), while legG2 has a Ras GEF domain, which exchanges GDP for GTP in Ras. Other pathogens also contain proteins that modify eukaryotic nucleotide-binding proteins (58); however, such proteins are rare in other bacterial species (45).

Ubiquitin-related domains.

Two gene products, LegAU13 and LegU1, contain an FBOX domain, which can link target proteins to a ubiquitin ligase in eukaryotes (31). LegAU13 contains both ankyrin repeats and an FBOX domain. LegU2 contains a UBOX domain, which, in eukaryotes, may possess E3 ubiquitin ligase activity (22). To our knowledge, such domains are exclusive to eukaryotes.


The other five predicted enzymes among the leg gene products include one SET lysine methyltransferase (legAS4), one amylase (legY), two putative phytanoyl-coenzyme A (CoA) dioxygenases (legD1 and -D2), and one astacin zinc protease (legP). The gene product of legN encodes a unique motif within this collection of genes. It contains two predicted cNMP binding domains in the middle portion of the ORF. The LegT protein contains a thaumatin domain, which makes part of the more generalized class of domains found in plant pathogenesis-related proteins, generally thought to be membrane-permeabilizing factors (4).

Evidence for interdomain horizontal gene transfer.

One theory to explain the presence of eukaryotic domains in the leg genes is that they may have been acquired by interdomain HGT. Koonin and colleagues have done extensive sequence analysis to identify the origin of genes with eukaryotic domains in bacteria (45). The presence of such genes in both the bacterial and eukaryotic lineages and not the archaeal lineage contradicts the notion of common ancestry, since if genes with these domains had arisen first in bacteria, they should have been transmitted vertically to both eukaryotes and archaea after the division of the domains. This suggests either that such genes were lost very early on in the archaeal lineage or that they were horizontally transferred between eukaryotes and bacteria. Even though interdomain horizontal gene transfer is not a well-characterized phenomenon, there are several reports suggesting that genetic material can be exchanged between prokaryotes and eukaryotes (5, 14, 21, 33, 51).

Two lines of evidence support the theory that at least some of the leg genes may have been acquired through interdomain HGT. First, the domains found in the leg genes are present in eukaryotes and, in a few cases, in a small number of other bacterial species (mostly pathogenic or endosymbiotic ones), but not in archaea (45). The presence of the eukaryotic domains found in the leg genes in few other bacterial species raises the possibility that some of the leg genes were acquired prior to the modern-day speciation of L. pneumophila and/or that interbacterial HGT occurred after the gene was originally acquired from a eukaryote. The second line of evidence is that the leg genes differ from the average L. pneumophila coding region values in both overall G+C content and G+C content by codon position. Differences in G+C content and distribution can often be seen in genes acquired from another species by HGT (35). For comparison, we selected 34 L. pneumophila “informational” genes, such as those involved in transcription and translation. These genes are less likely to be subject to HGT (27). We then calculated the overall G+C content and the G+C content in codon positions 1, 2, and 3 for all 46 leg genes (Table (Table3),3), the 34 “informational” genes, and all 3,005 predicted open reading frames in the L. pneumophila Philadelphia-1 genome. The leg genes differ in their overall G+C content and G+C content in codon positions 1 and 2 from L. pneumophila genomic averages and from the “informational” genes (two-tailed P values obtained from t tests are provided in Fig. Fig.2).2). We also found that the previously characterized genes for Icm-Dot substrates identified in our bioinformatics screen (Table (Table4)4) have G+C content biases similar to those of the leg genes (data not shown). These differences are consistent with the hypothesis that the leg genes and other effectors may have been acquired through HGT from a species with sequence characteristics distinct from those of L. pneumophila. The fact that the G+C content in codon position 3, the “wobble” codon, did not differ significantly suggests that there has been sufficient time for amelioration to the genomic average in codon position 3. The “wobble” codon allows for the greatest proportion of synonymous mutations and can ameliorate more quickly than codon positions 1 or 2 (34).

FIG. 2.
Dot plot analysis showing overall and codon position G+C content differences in the leg genes. The plot shows overall G+C content (A) and content of codon position 1 (B), codon position 2 (C), and codon position 3 (D) in the coding regions. ...

RpoS-dependent expression.

Several studies have suggested that in L. pneumophila the sigma factor RpoS regulates genes involved in intracellular replication (2, 20, 36, 63). We decided to find out if the leg genes are actually expressed and if they are regulated by RpoS. We chose the genes legA3, -AS4, -K1, -C1, -C3, -C4, -C5, -L5, -L7, -LC8, -G2, and -T because they represent distinct types of domains. We used 16S rRNA and Mip (15) as negative controls and the gene lpg0981 as a positive control based on our preliminary studies showing that Mip and 16S rRNA are not up-regulated in stationary phase, while lpg0981 is highly up-regulated in stationary phase in an RpoS-dependent manner (S. Pampou, S. Kalachikov, and H. A. Shuman, unpublished data). We isolated RNA from cells in logarithmic phase and in stationary phase both in an rpoS+ strain (JR32) and in an isogenic rpoS mutant (LM1376) and performed RT-PCR for the 12 genes listed above and the controls. We calculated the induction ratio for each gene based on the equation described in Methods. All the leg genes selected are expressed and up-regulated in stationary phase in an RpoS-dependent manner, albeit to various degrees (Fig. (Fig.33).

FIG. 3.
Stationary-phase up-regulation of leg genes in an RpoS-dependent manner. We performed RT-PCR for 12 leg genes, 2 negative controls (16S rRNA and Mip), and 1 positive control (lpg0981). We isolated bacterial RNA from cells growing in stationary and logarithmic ...

Translocation of Leg proteins into a eukaryotic host.

L. pneumophila encodes a type IVB secretion system called the Icm-Dot complex that translocates proteins from the pathogen into host cells (11, 36, 42, 43). The use of gene fusions to the catalytic domain of the calmodulin-dependent adenylate cyclase protein, CyaA of Bordetella pertussis, provides a convenient and reproducible method for measuring hybrid protein translocation from bacteria to eukaryotic cells (11, 42). In order to find out if the Leg proteins are translocated into host cells, we constructed hybrid proteins with CyaA, expressed the proteins in L. pneumophila, and then measured the capacity of these fusion proteins to be translocated into Dictyostelium discoideum. Increases in the level of cAMP in the host cell are indicative of translocation.

We amplified 20 candidate leg genes and created C-terminal translational gene fusions to the catalytic domain coding region of cyaA. We measured cAMP production following infection by both JR32 and dotA (LELA 3118) L. pneumophila expressing these genes. The dotA strain of L. pneumophila is unable to translocate effector proteins into the host because of a mutation in one of the essential components of the type IVB secretion system (11, 42). We considered the Leg-CyaA proteins to be translocated in an Icm-Dot-dependent manner if there were increases in the levels of host cAMP when using dot+ L. pneumophila and no increases when using the dotA mutant of L. pneumophila expressing the leg-cyaA hybrids. We were able to detect translocation for 8 out of 20 leg gene products tested (Fig. (Fig.4).4). Recent work by Roy and colleagues confirms the translocation of YlfB/LegC2 and shows that YlfA/LegC7, which was not tested in this study, is also a translocated eukaryotic-like protein (7). A lack of signal does not mean that the gene product is not translocated, since there are several reasons that could cause the assay to be unsuccessful for any particular hybrid protein. For instance, it has been shown that for the RalF effector, C-terminal CyaA fusions prevent translocation of the hybrid into host cells (42), while for the LepA and LepB effectors, C-terminal CyaA fusions do not block translocation (11).

FIG. 4.
Measurement of Icm-Dot-dependent translocation of several Leg-CyaA hybrid proteins into D. discoideum. JR32 and the isogenic strain dotA harboring constructs with Leg-CyaA fusion proteins were used to infect D. discoideum in microtiter plates. We obtained ...

Role of the leg genes in intracellular multiplication.

To evaluate the contribution of the leg genes to intracellular growth, we took advantage of an ordered library of transposon insertions in the L. pneumophila genome. We were able to find 20 leg knockout strains (legA3, -AS4, -A5, -A9, -A11, -A12, -L1, -LC4, -LC8, -C1, -C2, -C3, -C4, -C5, -K1, -S2, -G2, -U2, -D1, and -T) of the 44 searched for in the transposon library.

To find out if the leg gene products are required for intracellular growth, we tested the abilities of the 20 knockout strains to grow within Acanthamoeba castellanii. Most of the tested strains grow comparably to the parent strain in this host. However, the strain containing a transposon insertion in gene legLC8 is unable to grow to the same levels of JR32 (data not shown). Despite the fact that we could complement this growth defect by providing the gene in trans, when we constructed a full deletion of the gene by allelic exchange we did not further observe an intracellular growth defect (data not shown). This raises the possibility that an N-terminal fragment of the protein that is expressed has a negative effect in the context of intracellular growth. We conclude that consistent with previous findings from several laboratories (7, 36, 43, 44), most strains lacking the genes for putative effector proteins do not have intracellular growth defects.


Currently, there are two nonexclusive theories to explain how bacterial pathogens acquire genes that increase their ability to exploit host cells. One theory suggests that mutations in the bacterial chromosome and selective pressure can generate novel factors that lead to increased infectivity (58). The second theory proposes that these factors are acquired horizontally from host cells or other eukaryotic cells with conserved biological pathways. Additionally, the “food hypothesis” suggests that certain unicellular organisms that feed on bacteria are exposed to prokaryotic DNA that may be stably incorporated into their chromosomes (14, 17, 33). The “food hypothesis” can be extended to explain the flow of genes in the opposite direction: intracellular bacteria that parasitize on eukaryotic cells are exposed to eukaryotic DNA and may take up the foreign genetic material to enhance their infectivity. In addition, it has been reported that several pathogens, including L. pneumophila, are naturally competent (56), offering a plausible mechanism for the acquisition of eukaryotic DNA. It has been hypothesized that bacterial effector proteins have been acquired by both horizontal gene transfer and convergent evolution (58), and yet it is not clear which mechanism is predominant or if the lifestyle of a particular pathogen favors one mechanism over the other. Our findings suggest that in L. pneumophila, HGT may have been a major means through which effector proteins were acquired.

Eukaryotic-like genes in bacterial genomes are not always identified by genome annotation alone. Specific bioinformatics screens that search for genes that could have been acquired by eukaryotes can be considerably more sensitive. Here, for instance, we identified 20 genes that were not highlighted before in genome studies, despite the fact that three L. pneumophila strains had been fully sequenced and annotated. Importantly, among these 20 genes at least 8 are translocated into eukaryotic hosts. It is interesting that we could detect translocation for only three classes of leg genes: those containing coiled coils, leucine-rich repeats, or the one putative effector containing a Ras GEF domain. Since a CyaA translocation assay can fail for many reasons, it will be interesting to determine if the other leg genes are indeed translocated and if they play a role in virulence. Since many paralogs of certain eukaryotic-like genes exist (presumably a product of gene duplication), it is possible that only some are translocated and that the remaining ones may eventually be lost from the genome.

Several theories have been proposed to explain the lack of intracellular multiplication defects of L. pneumophila strains containing mutations in the genes that encode effector proteins. These theories include functional redundancy, nonnatural laboratory conditions, and host specificity. We studied mutations in 20 of the leg genes (including 8 whose gene products are translocated) and found that none are required for intracellular multiplication. Given that many of the leg genes have the same type of eukaryotic domains, it seems that there may be significant functional redundancy among this collection of genes. It seems counterintuitive that L. pneumophila would retain so many genes that can fully substitute for one another. It has been argued that retaining functionally redundant genes may prevent single genetic lesions from making the bacteria unable to multiply intracellularly. However, a single mutation in several genes of the Icm-Dot complex can render L. pneumophila completely avirulent. The leg genes identified here contain domains that seem to have been duplicated and shuffled throughout the chromosome. It is therefore possible that a negative pressure exists on the leg genes and that the high level of redundancy may eventually be reduced. It is interesting that a few of the leg genes are unique to either the Paris, Lens, or Philadelphia-1 isolates, indicating that deletions and/or duplications of the eukaryotic-like genes are still occurring in isolates of the same species. Alternatively, it is possible that genes that seem functionally redundant may actually play an essential role under particular environmental conditions or in other hosts that were not tested. This seems like a more plausible hypothesis to explain the lack of intracellular growth defect in strains containing mutations in effector proteins that do not have paralogs within the genome, such as RalF or LegG2.

Many of the eukaryotic-like genes (36 out of the 46 leg genes) encode putative protein binding domains. It should be possible to identify host cell interaction partners for these Leg proteins, particularly the ones for which we could detect translocation. Most of the other Leg proteins contain predicted enzymatic active sites. It will be interesting to determine if the three predicted Ser/Thr kinases have host substrates, since there is such a precedent for phosphorylation of host proteins by kinases in other pathogenic bacteria (58, 60). Also, L. pneumophila has two proteins with domains predicted to modify eukaryotic signaling lipids that participate in intracellular trafficking. An attractive hypothesis is that Legionella modifies both host proteins and signaling lipids in order to subvert organelle trafficking.

The large number of domains encoded by the leg genes reinforces the theory that L. pneumophila uses a varied arsenal of proteins to modulate intracellular trafficking to its own advantage. This suggests that this pathogen subverts multiple host cell functions during different stages of infection and that the biogenesis of a replicative vacuole probably involves many cellular factors. The large number of effector proteins in L. pneumophila may also help explain why this pathogen is capable of multiplying within a variety of phylogenetically unrelated eukaryotic hosts. Understanding the origin and mechanistic function of the leg proteins will certainly help clarify how L. pneumophila acquired effector proteins and how they subvert host cell functions. It will also be informative to find out if other pathogens encode such a large number of eukaryotic-like genes and if interdomain HGT is a general mechanism for the acquisition of factors necessary for bacterial pathogenesis.


We are grateful to the reviewers for many insightful suggestions. We thank David Figurski, Paul Planet, Richard Sucgang, Jonathan Dworkin, Sean Murray, and all members of the Shuman and Kessin laboratories for technical advice and helpful discussions.

This work was supported by NIH grants AI23549 (H.A.S.) and AI64481 (H.A.S.).


1. Amor, J. C., J. Swails, X. Zhu, C. R. Roy, H. Nagai, A. Ingmundson, X. Cheng, and R. A. Kahn. 2005. The structure of RalF, an ADP-ribosylation factor guanine nucleotide exchange factor from Legionella pneumophila, reveals the presence of a cap over the active site. J. Biol. Chem. 280:1392-1400. [PubMed]
2. Bachman, M. A., and M. S. Swanson. 2001. RpoS co-operates with other factors to induce Legionella pneumophila virulence in the stationary phase. Mol. Microbiol. 40:1201-1214. [PubMed]
3. Bardill, J. P., J. L. Miller, and J. P. Vogel. 2005. IcmS-dependent translocation of SdeA into macrophages by the Legionella pneumophila type IV secretion system. Mol. Microbiol. 56:90-103. [PubMed]
4. Batalia, M. A., A. F. Monzingo, S. Ernst, W. Roberts, and J. D. Robertus. 1996. The crystal structure of the antifungal protein zeamatin, a member of the thaumatin-like, PR-5 protein family. Nat. Struct. Biol. 3:19-23. [PubMed]
5. Budd, A., S. Blandin, E. A. Levashina, and T. J. Gibson. 2004. Bacterial alpha2-macroglobulins: colonization factors acquired by horizontal gene transfer from the metazoan genome? Genome Biol. 5:R38. [PMC free article] [PubMed]
6. Burkhard, P., J. Stetefeld, and S. V. Strelkov. 2001. Coiled coils: a highly versatile protein folding motif. Trends Cell Biol. 11:82-88. [PubMed]
7. Campodonico, E. M., L. Chesnel, and C. R. Roy. 2005. A yeast genetic system for the identification and characterization of substrate proteins transferred into host cells by the Legionella pneumophila Dot/Icm system. Mol. Microbiol. 56:918-933. [PubMed]
8. Cargill, K. L., B. H. Pyle, R. L. Sauer, and G. A. McFeters. 1992. Effects of culture conditions and biofilm formation on the iodine susceptibility of Legionella pneumophila. Can. J. Microbiol. 38:423-429. [PubMed]
9. Cazalet, C., C. Rusniok, H. Bruggemann, N. Zidane, A. Magnier, L. Ma, M. Tichit, S. Jarraud, C. Bouchier, F. Vandenesch, F. Kunst, J. Etienne, P. Glaser, and C. Buchrieser. 2004. Evidence in the Legionella pneumophila genome for exploitation of host cell functions and high genome plasticity. Nat. Genet. 36:1165-1173. [PubMed]
10. Chang, J. H., J. M. Urbach, T. F. Law, L. W. Arnold, A. Hu, S. Gombar, S. R. Grant, F. M. Ausubel, and J. L. Dangl. 2005. A high-throughput, near-saturating screen for type III effector genes from Pseudomonas syringae. Proc. Natl. Acad. Sci. USA 102:2549-2554. [PMC free article] [PubMed]
11. Chen, J., K. S. de Felipe, M. Clarke, H. Lu, O. R. Anderson, G. Segal, and H. A. Shuman. 2004. Legionella effectors that promote nonlytic release from protozoa. Science 303:1358-1361. [PubMed]
12. Chien, M., I. Morozova, S. Shi, H. Sheng, J. Chen, S. M. Gomez, G. Asamani, K. Hill, J. Nuara, M. Feder, J. Rineer, J. J. Greenberg, V. Steshenko, S. H. Park, B. Zhao, E. Teplitskaya, J. R. Edwards, S. Pampou, A. Georghiou, I. C. Chou, W. Iannuccilli, M. E. Ulz, D. H. Kim, A. Geringer-Sameth, C. Goldsberry, P. Morozov, S. G. Fischer, G. Segal, X. Qu, A. Rzhetsky, P. Zhang, E. Cayanis, P. J. De Jong, J. Ju, S. Kalachikov, H. A. Shuman, and J. J. Russo. 2004. The genomic sequence of the accidental pathogen Legionella pneumophila. Science 305:1966-1968. [PubMed]
13. Conover, G. M., I. Derre, J. P. Vogel, and R. R. Isberg. 2003. The Legionella pneumophila LidA protein: a translocated substrate of the Dot/Icm system associated with maintenance of bacterial integrity. Mol. Microbiol. 48:305-321. [PubMed]
14. de Koning, A. P., F. S. Brinkman, S. J. Jones, and P. J. Keeling. 2000. Lateral gene transfer and metabolic adaptation in the human parasite Trichomonas vaginalis. Mol. Biol. Evol. 17:1769-1773. [PubMed]
15. Engleberg, N. C., C. Carter, D. R. Weber, N. P. Cianciotto, and B. I. Eisenstein. 1989. DNA sequence of mip, a Legionella pneumophila gene associated with macrophage infectivity. Infect. Immun. 57:1263-1270. [PMC free article] [PubMed]
16. Feeley, J. C., R. J. Gibson, G. W. Gorman, N. C. Langford, J. K. Rasheed, D. C. Mackel, and W. B. Baine. 1979. Charcoal-yeast extract agar: primary isolation medium for Legionella pneumophila. J. Clin. Microbiol. 10:437-441. [PMC free article] [PubMed]
17. Field, J., B. Rosenthal, and J. Samuelson. 2000. Early lateral transfer of genes encoding malic enzyme, acetyl-CoA synthetase and alcohol dehydrogenases from anaerobic prokaryotes to Entamoeba histolytica. Mol. Microbiol. 38:446-455. [PubMed]
18. Fields, B. S., J. M. Barbaree, E. B. Shotts, Jr., J. C. Feeley, W. E. Morrill, G. N. Sanden, and M. J. Dykstra. 1986. Comparison of guinea pig and protozoan models for determining virulence of Legionella species. Infect. Immun. 53:553-559. [PMC free article] [PubMed]
19. Gump, D. W., and M. Keegan. 1986. Pulmonary infections due to Legionella in immunocompromised patients. Semin. Respir. Infect. 1:151-159. [PubMed]
20. Hales, L. M., and H. A. Shuman. 1999. The Legionella pneumophila rpoS gene is required for growth within Acanthamoeba castellanii. J. Bacteriol. 181:4879-4889. [PMC free article] [PubMed]
21. Hall, C., S. Brachat, and F. S. Dietrich. 2005. Contribution of horizontal gene transfer to the evolution of Saccharomyces cerevisiae. Eukaryot. Cell 4:1102-1115. [PMC free article] [PubMed]
22. Hatakeyama, S., and K. I. Nakayama. 2003. U-box proteins as a new family of ubiquitin ligases. Biochem. Biophys. Res. Commun. 302:635-645. [PubMed]
23. Hilbi, H., G. Segal, and H. A. Shuman. 2001. Icm/dot-dependent upregulation of phagocytosis by Legionella pneumophila. Mol. Microbiol. 42:603-617. [PubMed]
24. Horwitz, M. A., and F. R. Maxfield. 1984. Legionella pneumophila inhibits acidification of its phagosome in human monocytes. J. Cell Biol. 99:1936-1943. [PMC free article] [PubMed]
25. Horwitz, M. A., and S. C. Silverstein. 1980. Legionnaires' disease bacterium (Legionella pneumophila) multiples intracellularly in human monocytes. J. Clin. Investig. 66:441-450. [PMC free article] [PubMed]
26. Iturbe-Ormaetxe, I., G. R. Burke, M. Riegler, and S. L. O'Neill. 2005. Distribution, expression, and motif variability of ankyrin domain genes in Wolbachia pipientis. J. Bacteriol. 187:5136-5145. [PMC free article] [PubMed]
27. Jain, R., M. C. Rivera, and J. A. Lake. 1999. Horizontal gene transfer among genomes: the complexity hypothesis. Proc. Natl. Acad. Sci. USA 96:3801-3806. [PMC free article] [PubMed]
28. Jovanovic, O. S. 2002. Computational, genetic and biochemical analysis of broad host-range plasmid RK2. Doctoral thesis. Columbia University, New York, N.Y.
29. Kagan, J. C., and C. R. Roy. 2002. Legionella phagosomes intercept vesicular traffic from endoplasmic reticulum exit sites. Nat. Cell Biol. 4:945-954. [PubMed]
30. Kedzierski, L., J. Montgomery, J. Curtis, and E. Handman. 2004. Leucine-rich repeats in host-pathogen interactions. Arch. Immunol. Ther. Exp. (Warsz) 52:104-112. [PubMed]
31. Kipreos, E. T., and M. Pagano. 2000. The F-box protein family. Genome Biol 1:REVIEWS3002. [PMC free article] [PubMed]
32. Kobe, B., and J. Deisenhofer. 1995. Proteins with leucine-rich repeats. Curr. Opin. Struct. Biol. 5:409-416. [PubMed]
33. Kondo, N., N. Nikoh, N. Ijichi, M. Shimada, and T. Fukatsu. 2002. Genome fragment of Wolbachia endosymbiont transferred to X chromosome of host insect. Proc. Natl. Acad. Sci. USA 99:14280-14285. [PMC free article] [PubMed]
34. Lawrence, J. G., and H. Ochman. 1997. Amelioration of bacterial genomes: rates of change and exchange. J. Mol. Evol. 44:383-397. [PubMed]
35. Lawrence, J. G., and H. Ochman. 1998. Molecular archaeology of the Escherichia coli genome. Proc. Natl. Acad. Sci. USA 95:9413-9417. [PMC free article] [PubMed]
36. Luo, Z. Q., and R. R. Isberg. 2004. Multiple substrates of the Legionella pneumophila Dot/Icm system identified by interbacterial protein transfer. Proc. Natl. Acad. Sci. USA 101:841-846. [PMC free article] [PubMed]
37. Malik, Z. A., C. R. Thompson, S. Hashimi, B. Porter, S. S. Iyer, and D. J. Kusner. 2003. Cutting edge: Mycobacterium tuberculosis blocks Ca2+ signaling and phagosome maturation in human macrophages via specific inhibition of sphingosine kinase. J. Immunol. 170:2811-2815. [PubMed]
38. Marrie, T. F., and R. S. Martin. 1988. Persistent Legionella pneumophila bacteraemia in an immunocompromised host. J. Infect. 16:203-204. [PubMed]
39. McGraw, E. A., and S. L. O'Neill. 2004. Wolbachia pipientis: intracellular infection and pathogenesis in Drosophila. Curr. Opin. Microbiol. 7:67-70. [PubMed]
40. Miyara, T., K. Tokashiki, T. Shimoji, K. Tamaki, M. Koide, and A. Saito. 2002. Rapidly expanding lung abscess caused by Legionella pneumophila in immunocompromised patients: a report of two cases. Intern. Med. 41:133-137. [PubMed]
41. Mosavi, L. K., T. J. Cammett, D. C. Desrosiers, and Z. Y. Peng. 2004. The ankyrin repeat as molecular architecture for protein recognition. Protein Sci. 13:1435-1448. [PMC free article] [PubMed]
42. Nagai, H., E. D. Cambronne, J. C. Kagan, J. C. Amor, R. A. Kahn, and C. R. Roy. 2004. A C-terminal translocation signal required for Dot/Icm-dependent delivery of the Legionella RalF protein to host cells. Proc. Natl. Acad. Sci. USA 102:826-831. [PMC free article] [PubMed]
43. Nagai, H., J. C. Kagan, X. Zhu, R. A. Kahn, and C. R. Roy. 2002. A bacterial guanine nucleotide exchange factor activates ARF on Legionella phagosomes. Science 295:679-682. [PubMed]
44. Ninio, S., D. M. Zuckman-Cholon, E. D. Cambronne, and C. R. Roy. 2005. The Legionella IcmS-IcmW protein complex is important for Dot/Icm-mediated protein translocation. Mol. Microbiol. 55:912-926. [PubMed]
45. Ponting, C. P., L. Aravind, J. Schultz, P. Bork, and E. V. Koonin. 1999. Eukaryotic signalling domain homologues in archaea and bacteria. Ancient ancestry and horizontal gene transfer. J. Mol. Biol. 289:729-745. [PubMed]
46. Rogers, J., A. B. Dowsett, P. J. Dennis, J. V. Lee, and C. W. Keevil. 1994. Influence of temperature and plumbing material selection on biofilm formation and growth of Legionella pneumophila in a model potable water system containing complex microbial flora. Appl. Environ. Microbiol. 60:1585-1592. [PMC free article] [PubMed]
47. Rogers, J., and C. W. Keevil. 1992. Immunogold and fluorescein immunolabelling of Legionella pneumophila within an aquatic biofilm visualized by using episcopic differential interference contrast microscopy. Appl. Environ. Microbiol. 58:2326-2330. [PMC free article] [PubMed]
48. Rowbotham, T. J. 1980. Preliminary report on the pathogenicity of Legionella pneumophila for freshwater and soil amoebae. J. Clin. Pathol. 33:1179-1183. [PMC free article] [PubMed]
49. Sadosky, A. B., L. A. Wiater, and H. A. Shuman. 1993. Identification of Legionella pneumophila genes required for growth within and killing of human macrophages. Infect. Immun. 61:5361-5373. [PMC free article] [PubMed]
50. Salcedo, S. P., and D. W. Holden. 2003. SseG, a virulence protein that targets Salmonella to the Golgi network. EMBO J. 22:5003-5014. [PMC free article] [PubMed]
51. Schlieper, D., M. A. Oliva, J. M. Andreu, and J. Lowe. 2005. Structure of bacterial tubulin BtubA/B: evidence for horizontal gene transfer. Proc. Natl. Acad. Sci. USA 102:9170-9175. [PMC free article] [PubMed]
52. Segal, G., M. Purcell, and H. A. Shuman. 1998. Host cell killing and bacterial conjugation require overlapping sets of genes within a 22-kb region of the Legionella pneumophila genome. Proc. Natl. Acad. Sci. USA 95:1669-1674. [PMC free article] [PubMed]
53. Segal, G., J. J. Russo, and H. A. Shuman. 1999. Relationships between a new type IV secretion system and the icm/dot virulence system of Legionella pneumophila. Mol. Microbiol. 34:799-809. [PubMed]
54. Seki, T., N. Hayashi, and T. Nishimoto. 1996. RCC1 in the Ran pathway. J. Biochem. (Tokyo) 120:207-214. [PubMed]
55. Seshadri, R., I. T. Paulsen, J. A. Eisen, T. D. Read, K. E. Nelson, W. C. Nelson, N. L. Ward, H. Tettelin, T. M. Davidsen, M. J. Beanan, R. T. Deboy, S. C. Daugherty, L. M. Brinkac, R. Madupu, R. J. Dodson, H. M. Khouri, K. H. Lee, H. A. Carty, D. Scanlan, R. A. Heinzen, H. A. Thompson, J. E. Samuel, C. M. Fraser, and J. F. Heidelberg. 2003. Complete genome sequence of the Q-fever pathogen Coxiella burnetii. Proc. Natl. Acad. Sci. USA 100:5455-5460. [PMC free article] [PubMed]
56. Sexton, J. A., and J. P. Vogel. 2004. Regulation of hypercompetence in Legionella pneumophila. J. Bacteriol. 186:3814-3825. [PMC free article] [PubMed]
57. Shohdy, N., J. A. Efe, S. D. Emr, and H. A. Shuman. 2005. Pathogen effector protein screening in yeast identifies Legionella factors that interfere with membrane trafficking. Proc. Natl. Acad. Sci. USA 102:4866-4871. [PMC free article] [PubMed]
58. Stebbins, C. E., and J. E. Galan. 2001. Structural mimicry in bacterial virulence. Nature 412:701-705. [PubMed]
59. Vogel, J. P., H. L. Andrews, S. K. Wong, and R. R. Isberg. 1998. Conjugative transfer by the virulence system of Legionella pneumophila. Science 279:873-876. [PubMed]
60. Walburger, A., A. Koul, G. Ferrari, L. Nguyen, C. Prescianotto-Baschong, K. Huygen, B. Klebl, C. Thompson, G. Bacher, and J. Pieters. 2004. Protein kinase G from pathogenic mycobacteria promotes survival within macrophages. Science 304:1800-1804. [PubMed]
61. Wiater, L. A., A. B. Sadosky, and H. A. Shuman. 1994. Mutagenesis of Legionella pneumophila using Tn903 dlllacZ: identification of a growth-phase-regulated pigmentation gene. Mol. Microbiol. 11:641-653. [PubMed]
62. Wu, M., L. V. Sun, J. Vamathevan, M. Riegler, R. Deboy, J. C. Brownlie, E. A. McGraw, W. Martin, C. Esser, N. Ahmadinejad, C. Wiegand, R. Madupu, M. J. Beanan, L. M. Brinkac, S. C. Daugherty, A. S. Durkin, J. F. Kolonay, W. C. Nelson, Y. Mohamoud, P. Lee, K. Berry, M. B. Young, T. Utterback, J. Weidman, W. C. Nierman, I. T. Paulsen, K. E. Nelson, H. Tettelin, S. L. O'Neill, and J. A. Eisen. 2004. Phylogenomics of the reproductive parasite Wolbachia pipientis wMel: a streamlined genome overrun by mobile genetic elements. PLoS Biol. 2:E69. [PMC free article] [PubMed]
63. Zusman, T., O. Gal-Mor, and G. Segal. 2002. Characterization of a Legionella pneumophila relA insertion mutant and roles of RelA and RpoS in virulence gene expression. J. Bacteriol. 184:67-75. [PMC free article] [PubMed]

Articles from Journal of Bacteriology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...