Characterization of the icmH and icmF Genes Required for Legionella pneumophila Intracellular Growth, Genes That Are Present in Many Bacteria Associated with Eukaryotic Cells

Abstract

Legionella pneumophila, the causative agent of Legionnaires' disease, replicates intracellularly within a specialized phagosome of mammalian and protozoan host cells, and the Icm/Dot type IV secretion system has been shown to be essential for this process. Unlike all the other known Icm/Dot proteins, the IcmF protein, which was described before, and the IcmH protein, which is characterized here, have homologous proteins in many bacteria (such as Yersinia pestis, Salmonella enterica, Rhizobium leguminosarum, and Vibrio cholerae), all of which associate with eukaryotic cells. Here, we have characterized the L. pneumophila icmH and icmF genes and found that both genes are present in 16 different Legionella species examined. The icmH and icmF genes were found to be absolutely required for intracellular multiplication in Acanthamoeba castellanii and partially required for intracellular growth in HL-60-derived human macrophages, for immediate cytotoxicity, and for salt sensitivity. Mutagenesis of the predicted ATP/GTP binding site of IcmF revealed that the site is partially required for intracellular growth in A. castellanii. Analysis of the regulatory region of the icmH and icmF genes, which were found to be cotranscribed, revealed that it contains at least two regulatory elements. In addition, an icmH::lacZ fusion was shown to be activated during stationary phase in a LetA- and RelA-dependent manner. Our results indicate that although the icmH and icmF genes probably have a different evolutionary origin than the rest of the icm/dot genes, they are part of the icm/dot system and are required for L. pneumophila pathogenesis.

Bacterial pathogens, as well as bacteria that live in close contact with eukaryotic cells, have developed many mechanisms to subvert their host cells and grow in intimate association with them. Many bacterial pathogens, such as Yersinia spp., Salmonella enterica, Pseudomonas aeroginosa, and Escherichia coli O157, use the type III secretion system as part of their pathogenesis determinants (14). Other bacteria such as Agrobacterium tumefaciens, Bordetella pertussis, and Legionella pneumophila use type IV secretion systems, which are functionally homologous to type III secretion systems but are evolutionarily related to bacterial conjugation systems, as opposed to the type III secretion systems, which are evolutionarily related to the bacterial flagellar basal body (9, 13).

L. pneumophila, the causative agent of Legionnaires' disease, is a facultatively intracellular pathogen that is able to infect, multiply within, and kill human macrophages, as well as free-living amoebae (32, 48). Two regions of icm/dot genes that constitute the L. pneumophila icm/dot type IV secretion system have been discovered (reviewed in references 53 and 64). Region I contains 7 genes (icmV, -W, and -X and dotA, -B, -C, and -D) (3, 6, 39, 63), and region II has been shown to contain 17 genes (icmT, -S, -R, -Q, -P, -O, -N, -M, -L, -K, -E, -G, -C, -D, -J, -B, and -F) (1, 46, 50, 52, 63). The icm/dot genes were shown to participate in many aspects of L. pneumophila pathogenesis, such as phagocytosis (29, 66), immediate cytotoxicity (36, 71), inhibition of phagosome-lysosome fusion at early times during infection (11, 60, 61, 67), association of the phagosome with the rough endoplasmic reticulum (35, 44), apoptosis (70), and exit from the phagosome (42). As a consequence of all these features, the icm/dot genes were found to be essential for intracellular multiplication in all of the hosts examined: HL-60- and U937-derived human macrophages (50, 67), murine bone marrow-derived macrophages (63), and the protozoan hosts Acanthamoeba castellanii (55) and Dictyostelium discoideum (57). Thus far, the icm/dot type IV system has been shown to translocate two effector proteins (RalF and LidA) into the host cell during infection (12, 44), and additional effectors for the system are expected to be found.

Eighteen proteins encoded by the icm/dot genes (IcmT, -P, -O, -M, -L, -K, -E, -G, -C, -D, -J, -B, -V, and -X and DotA, -B, -C, and -D) contain significant sequence homology to conjugation-related proteins from the IncI plasmid R64 (37, 56). The origins of the six icm/dot genes (icmS, icmW, icmR, icmQ, icmN, and icmF) that do not have homologues on the IncI plasmids are unknown. Among the products of these six genes, the IcmS protein was found to interact with IcmW and the IcmR protein was shown to interact with IcmQ (10, 73). The icmN gene is predicted to encode a lipoprotein that belongs to the OmpA protein family, which contains homologous proteins found in many gram-negative bacteria (50).

In this study, the icmF transcriptional unit was analyzed, and it was found to contain a new icm gene, which was named icmH. Homologues of the L. pneumophila IcmH and IcmF proteins were found in several bacteria that live in intimate contact with eukaryotic cells (such as Yersinia pestis, E. coli O157, Vibrio cholerae, Rhizobium leguminosarum, and A. tumefaciens), which is a unique property in comparison to the other Icm/Dot proteins. Our results indicate that the IcmF and IcmH proteins are required for L. pneumophila intracellular multiplication, but a partially functional Icm/Dot complex is probably present in the bacteria even in their absence. Our results indicate a possible role for these proteins in the interaction of L. pneumophila and other bacteria with eukaryotic cells.

MATERIALS AND METHODS

Bacterial strains, plasmids, primers, and media.

The L. pneumophila and E. coli strains used in this work are listed in Table 1. The plasmids and primers used in this work are described in Table 2 and Table 3, respectively. Bacterial media, plates, and antibiotic concentrations were used as described before (52).

TABLE 1.

Bacterial strains

Strain	Genotype and features	Reference or Source
L. pneumophila
25D	Icm− avirulent mutant	31
GS3011	JR32 icmT3011::Km	54
GS3015	JR32 icmF3015::Km	This study
GS3016	JR32 icmH3016::Km	This study
GS-RelA	JR32 relA::Km	72
JR32	Homogeneous salt-sensitive isolate of AM511	49
LELA3118	JR32 dotA3118::Tn903dIIlacZ	49
LM1376	JR32 rpoS::Km	25
MW627	JR32 tphA::Km	46
OG2001	JR32 letA::Km	23
OG2002	JR32 cpxR::Km	22
OG2003	JR32 rpoE::Km	22
OG2004	JR32 cpxA::Km	22
E. coli
MC1022	araD139 Δ(ara leu)7697 Δ(lacZ)M15 galU galK strA	8
MC1061	araD139 Δ(ara leu)7697 ΔlacX74 galU galK strA	8

TABLE 2.

Plasmids used in this study

Plasmid	Feature(s)	Reference or Source
pGS-lac-02	pAB-1 with a promoterless lacZ gene	24
pGS-Lc-55-14	The icmH-icmF-tphA operon in pMMB207αb-Km-14	55
pGS-Lc-70-14	The icmH-icmF-tphA operon in pMMB207αb-Km-14	This study
pGS-Lc-70-G2S-14	pGS-Lc-70-14 with a mutation in the icmF ATP-binding site	This study
pGS-Lc-70-K2A-14	pGS-Lc-70-14 with a mutation in the icmF ATP-binding site	This study
pGS-Lc-76-14	icmH in pMMB207αb-Km-14	This study
pGS-Lc-77-14	pGS-Lc-70-14 with an in-frame deletion in icmH	This study
pGS-Lp-71	The icmH gene in pUC-18	This study
pGS-Lp-72	Part of the icmH-icmF-tphA operon in pUC-18	This study
pGS-Lp-74	pGS-Lp-71 with an in-frame deletion in icmH	This study
pGS-Lp-75	icmH and part of icmF in pHG-165	This study
pGS-Lp-72-Km	pGS-Lp-72 with kanamycin cassette in icmF	This study
pGS-Lp-72-Km-GR	Insert of pGS-Lp-72-Km cloned in pLAW344	This study
pGS-Lp-78	pGS-Lp-75 with an in-frame deletion in icmH	This study
pGS-Lp-78-Km	pGS-Lp-78 with kanamycin cassette in icmH	This study
pGS-Lp-78-Km-GR	Insert of pGS-Lp-78-Km cloned in pLAW344	This study
pGS-reg-F4	151 bp of the regulatory region of icmH in pGS-lac-02	This study
pGS-reg-F4-M2	pGS-reg-F4 with a mutation in the regulatory region	This study
pGS-reg-F5	278 bp of the regulatory region of icmH in pGS-lac-02	This study
pGS-reg-F5-M2	pGS-reg-F5 with a mutation in the regulatory region	This study
pGS-reg-F6	251 bp of the regulatory region of icmH in pGS-lac-02	This study
pGS-reg-F7	228 bp of the regulatory region of icmH in pGS-lac-02	This study
pGS-reg-F8	217 bp of the regulatory region of icmH in pGS-lac-02	This study
pGS-reg-F9	178 bp of the regulatory region of icmH in pGS-lac-02	This study
pGS-reg-F10	116 bp of the regulatory region of icmH in pGS-lac-02	This study
pGS-reg-F11	102 bp of the regulatory region of icmH in pGS-lac-02	This study
pGS-reg-F12	240 bp of the regulatory region of icmH in pGS-lac-02	This study
pGS-reg-F13	137 bp of the regulatory region of icmH in pGS-lac-02	This study
pGS-reg-F14	126 bp of the regulatory region of icmH in pGS-lac-02	This study
pHG165	oriR(colEI) MCS Apr	59
pLAW344	oriR(colEI) sacB MCS oriT(RK2) Cmr Apr	68
pMMB207αb-Km-14	oriV(RSF1010) MCS lacIq CmrmobA::Km	54
pUC18	oriR(colEI) MCS Apr	69
pZT-reg-F15	91 bp of the regulatory region of icmH in pGS-lac-02	This study
pZT-reg-F16	81 bp of the regulatory region of icmH in pGS-lac-02	This study

TABLE 3.

Primers used in this study

Primer name	Sequence (5′-3′)
icmH-F	ATCGCTGTAGGTATTGTAATTCTAGC
icmH-R	ATCAGGCTCTGTGATTGCAAGACG
F-G2S-F	ACGCCCAATCTAAATCGGCACTGTTAAAGCA
F-G2S-R	TGCCGATTTAGATTGGGCGTTTTTTCCGGTGA
F-K2A-F	CCCAAGGCGCATCGGCACTGTTAAAGCAAAG
F-K2A-R	CAGTGCCGATGCGCCTTGGGCGTTTTTTCCGG
F-lac	CGGGGGATCCCCGTATTGCTCAGTTGTCATTATAT
F-M2-F	TTTACCTAGAGAATAAATAAGAATTATAGGAATAC
F-M2-R	CTTATTTATTCTCTAGGTAAAATCTGCTTCTAG
RT-F-2	TCCTTGAAATTGGCGGCACT
RT-F-GSP	GTGCTGGTTCGGCGATCAGT
RT-H-2	GAATTACGCGCCTTTCACAG
RT-Tp-1	AAGCCCACCAATTGGACGCA
RT-Tp-GSP	GCCGTCCTACATGATCAGCA
T-F4	GCCGGAATTCGAACAAGGAGCAAGTATTTC
T-F5	GCCGGAATTCGCATGTATTAATTCTAAGCTTGAC
T-F6	GCCGGAATTCCTTGAACAAAGAGTCGTATATAAC
T-F7	GCCGGAATTCCGTATCATCTTCTTTTTAATTGAG
T-F8	GCCGGAATTCCTTTTTGATTAGCATAAATGGACC
T-F9	GCCGGAATTCTTTCTTGTAGGACTGATGGACTTG
T-F10	GCCGGAATTCTTACCTATAGAATAAATAAGAATTATAGG
T-F11	GCCGGAATTCAATAAGAATTATAGGAATACTTACTAATC
T-F12	GCCGGAATTCAGTCGTATATAACGTATCATCTTC
T-F13	GCCGGAATTCAAGTATTTCTAGAAGCAGATTTTAC
T-F14	GCCGGAATTCGAAGCAGATTTTACCTATAGAATA
T-F15	GCCGGAATTCTAGGAATACTTACTAATCAACTAGC
T-F16	GCCGGAATTCTACTAATCAACTAGCAATCACG
RACE-GSP	TTTACTGTGAAAGGCGCGTAA
RACE-F2	CGGGGGATCCGTGAGGACGGGTATTGCTCAGTTGTC

Plasmid construction for complementation.

The plasmid pGS-Lc-55-14 was used to construct new complementing plasmids for the icmH-icmF-tphA operon. The plasmid pGS-Lc-55-14 was digested with XhoI, SacII, and a 7,109-bp fragment that contains the icmH-icmF-tphA operon, as well as part of an open reading frame located upstream from icmH, was filled in and cloned into the SmaI site of the vector pMMB207αb-km-14 to generate pGS-Lc-70-14. The plasmid pGS-Lc-70-14 was digested with KpnI, and a 3,182-bp fragment was deleted from the plasmid after self-ligation to generate pGS-Lc-76-14. The resulting plasmid contained the icmH gene by itself (the icmF and tphA genes were deleted due to the KpnI digestion). To generate a complementing plasmid that contained a nonpolar in-frame deletion in icmH, the plasmid pGS-Lc-70-14 was digested with SphI and XbaI, and the resulting 1,081-bp fragment was cloned into the corresponding sites in pUC-18 to generate pGS-Lp-71. An in-frame deletion in icmH was constructed by PCR, as described before (33), using the icmH-F and icmH-R primers (Table 3). The first 3 nucleotides of both primers were changed from the original sequence to form an EcoRV site after self-ligation (each primer contains half of an EcoRV site at its 5′ end). The PCR conditions were 30 cycles at 95°C for 1 min, 60°C for 0.5 min, and 75°C for 5 min, performed in 100-μl reaction mixtures using the buffer supplied with the enzyme, 200 mM each nucleotide, ∼0.1 μg of pGS-Lp-71 plasmid DNA, 50 pmol of each primer, and 2 U of Vent DNA polymerase (New England BioLabs). The PCR product was gel purified and self-ligated. After transformation, the plasmids prepared were examined for the presence of the EcoRV site expected to be generated by the ends of the two primers. The insert of the plasmid harboring the in-frame deletion in icmH was sequenced and named pGS-Lp-74. The plasmid pGS-Lp-74 was digested with XbaI and SphI, and the insert containing the in-frame deletion in icmH was cloned into pGS-Lc-70-14 to generate pGS-Lc-77-14. The plasmid pGS-Lc-77-14 contains an in-frame deletion in the icmH gene with the icmF-tphA part of the operon unchanged.

Allelic exchange.

To construct icmF and icmH knockout strains, plasmids containing parts of the icmH and icmF genes were constructed using the sequence information available from the L. pneumophila genome sequence database (http://genome3.cpmc.columbia.edu/∼legion/index.html). For the construction of the icmF knockout strain, the plasmid pGS-Lc-55-14 was digested with EcoRI and XbaI, and a 3,797-bp fragment was cloned into the corresponding sites in pUC-18 to generate pGS-Lp-72. Then, the kanamycin resistance cassette (Pharmacia) was cloned into it instead of an internal 1,872-bp EcoRV-Eco47III fragment to generate pGS-Lp-72-Km. The plasmid pGS-Lp-72-Km was digested with PvuII, and the insert was cloned into the EcoRV site of the allelic-exchange vector pLAW344 to form pGS-Lp-72-Km-GR. This plasmid was used for allelic exchange, as described previously (52). The resulting strain (GS3015) contained the first 139 amino acids of the IcmF protein, as well as its last 175 amino acids. The internal 625 amino acids were deleted during the construction, and the kanamycin resistance cassette was cloned instead. Several isolates were analyzed by PCR to confirm that the right change occurred on the chromosome (data not shown). For the construction of the icmH knockout strain, the plasmid pGS-Lc-55-14 was digested with PstI, and a 4,417-bp fragment generated was cloned into the corresponding site in pHG-165 to generate pGS-Lp-75. Then, an in-frame deletion was constructed in the icmH gene with the same primers and by using the same method described for the construction of pGS-Lp-74 to generate pGS-Lp-78. The kanamycin resistance cassette was cloned into the EcoRV site generated by PCR to generate pGS-Lp-78-Km. The plasmid pGS-Lp-78-Km was digested with PvuII, and the insert was cloned into the EcoRV site of the allelic-exchange vector pLAW344 to form pGS-Lp-78-Km-GR. This plasmid was used for allelic exchange, as described previously (52). The resulting strain (GS3016) contained the first 19 amino acids of the IcmH protein, as well as its last 36 amino acids. The internal 207 amino acids were deleted during the construction, and the kanamycin resistance cassette was cloned instead. Several isolates were analyzed by PCR to confirm that the right change occurred on the chromosome (data not shown).

Construction of mutations in the predicted ATP/GTP binding site of IcmF.

The plasmid pGS-Lp-71 (described above) was used as a template in site-directed mutagenesis using the overlap extension PCR method (30). For each mutation, two primers that contained the mutation and overlapped one another by 20 bp were designed (F-G2S-F, F-G2S-R, F-K2A-F, and F-K2A-R [Table 3]). The PCR mutagenesis includes two steps. In the first step, two PCR fragments were generated using the following primers: (i) a primer located on the vector upstream from the regulatory region (one of the primers containing the mutation) and (ii) a primer located on the vector (the second primer containing the mutation, on the complementary strand). The resulting two fragments were gel purified and used as templates in the second step, which included a third PCR using the two primers located on the vector. The resulting PCR product was digested with XbaI and SphI and cloned into the same sites in pGS-Lc-70-14 to generate pGS-Lc-70-G2S-14 and pGS-Lc-70-14-K2A-14. The mutations were confirmed by sequencing the whole region amplified during the PCR.

Construction of lacZ fusions.

The promoterless lacZ vector pGS-lac-02 was used for cloning different fragments originating from the regulatory region of the icmH-icmF-tphA operon. The fragments cloned were generated by PCR using the F-lac primer (Table 3) and a second primer located at different distances (Table 2) from the IcmH first methionine (primers T-F4 to T-F16 [Table 3]). The fragments generated were digested with BamHI and EcoRI and subsequently cloned into the same sites in pGS-lac-02. The plasmids generated were sequenced to confirm that no changes occurred in the regulatory region and were named according to the primer (e.g., the plasmid resulting from the PCR performed with F-lac and T-F4 was named pGS-reg-F4).

The site-directed mutagenesis was performed by the overlap extension PCR method (30), using the primers F-M2-F and F-M2-R (Table 3). The same mutation was constructed in the plasmids pGS-reg-F4 and pGS-reg-F5 to generate pGS-reg-F4-M2 and pGS-reg-F5-M2, respectively.

Intracellular growth in A. castellanii.

Intracellular-growth assays were performed similarly to that described previously (55). A. castellanii (ATCC 30234) (1.5 × 10⁵ organisms) in PYG (Proteose Peptone-yeast extract-glucose) was added to wells of a 24-well microtiter plate, and the amoebae were incubated for 1 h at 37°C to let the amoebae adhere. Then, the PYG was aspirated, and the wells were washed once with 0.5 ml of warm (37°C) Acanthamoeba buffer (Ac buffer), and 0.5 ml of warm Ac buffer was added to the wells. Then, L. pneumophila in Ac buffer was added to the wells at a multiplicity of infection (MOI) of ∼0.1. The plate was incubated for 30 min at 37°C, the Ac buffer was aspirated, the wells were washed three times with 0.5 ml of warm Ac buffer, and 0.6 ml of warm Ac buffer was added to the wells. The supernatant of each well was sampled at intervals of ∼12 or 24 h, and the numbers of CFU were determined by plating samples on ABCYE {ACES[N-(2-acetamido)-2-aminoethanesulfonic acid]-buffered charcoal-yeast extract} plates.

Intracellular growth in HL-60-derived human macrophages.

Intracellular-growth assays were performed similarly to those previously described (54). Wells of a 24-well plate containing 2 × 10⁶ differentiated HL-60-derived human macrophages were used for infection. L. pneumophila was added to the wells at an MOI of ∼0.1, and the infected HL-60-derived macrophages were incubated for 1 h at 37°C under CO₂ (5%). Then, the wells were washed three times, and 0.6 ml of RPMI containing 2 mM Gln and 10% normal human serum was added to the wells. The supernatant of each well was sampled at intervals of ∼24 h, and the numbers of CFU were determined by plating samples on ABCYE plates.

In the experiments in which the numbers of intracellular bacteria were determined, at each time point a well containing infected cells was lysed by adding sterile double-distilled water to the well, and the numbers of bacteria inside the cells were determined by plating samples on ABCYE plates.

When entry of the bacteria into HL-60-derived human macrophages was determined, the infection was performed at an MOI of 1, followed by centrifugation and incubation with gentamicin for 1 h. Then, the cells were washed three times and the number of intracellular bacteria was determined as described above.

Immediate cytotoxicity to HL-60-derived human macrophages.

The immediate-cytotoxicity assay was preordered as described before (7) with several modifications. Wells of a 96-well plate containing 4 × 10⁵ differentiated HL-60-derived human macrophages were infected with twofold serial dilutions of L. pneumophila in RPMI, starting from ∼10⁸ bacteria/well. Then, the plate was centrifuged for 2 min at 880 × g and incubated for 1 h at 37°C under CO₂ (5%). Later, the plate was washed three times, and 0.1 ml of RPMI containing 2 mM Gln, 10% normal human serum, and 10% Alamar-Blue (Biosource) was added to the wells. After incubation at 37°C under CO₂ (5%) for 4 h, the fluorescence intensity was measured at 590 nm (excitation at 530 nm) to determine the extent of macrophage killing.

Sodium sensitivity.

The sodium sensitivity assay was performed essentially as described before (7). A wild-type L. pneumophila strain (JR32) and several mutants were grown for 72 h on an ABCYE plate, scraped off the plate, and calibrated to an optical density at 600 nm (OD₆₀₀) of 4. Then, eight 10-fold serial dilutions were plated on ABCYE plates containing or lacking 100 mM NaCl. The sodium sensitivity was determined by comparing the numbers of bacteria growing on the plates, and it is presented as the percentage of bacteria that grew on the ABCYE plates containing NaCl in relation to the standard ABCYE plates.

β-Galactosidase assays.

β-Galactosidase assays were performed as described elsewhere (41). L. pneumophila strains were grown for 48 h on ABCYE plates containing chloramphenicol. The bacteria were scraped off the plate and suspended in ACES-yeast extract broth, and the bacterial OD₆₀₀ was calibrated to 0.1 in ACES-yeast extract broth. The resulting cultures were grown on a roller drum for ∼18 h until they reached an OD₆₀₀ of ∼3.8 (stationary phase). To test the levels of expression at exponential phase, the cultures described were diluted to an OD₆₀₀ of 0.1 and grown for an additional 6 to 7 h until they reached an OD₆₀₀ of ∼0.7 (exponential phase). The assays were done for 20, 50, or 100 μl of culture, and the substrate for lacZ hydrolysis was o-nitrophenyl-β-d-galactopyranoside.

RNA manipulations.

RNA was prepared as described before (22). To determine the transcription start site of the icmH-icmF-tphA operon, 5′ rapid amplification of cDNA ends (RACE) (Invitrogen) was performed as described by the manufacturer. The RACE-GSP primer (Table 3) was used for generating the cDNA; this primer, together with the AA (abridged-anchor) primer supplied with the kit, was used for the first PCR, and the AUA (abridged universal amplification) primer supplied with the kit and RACE-F2 (Table 3) were used for the nested PCR. The resulting fragment from the second PCR was subsequently cloned, and seven different clones were sequenced to determine the transcription start site of the mRNA.

To determine if the icmH, icmF, and tphA genes are located on one transcriptional unit, a reverse-transcription (RT) reaction was performed using the primers RT-F-GSP and RT-Tp-GFP and avian myeloblastosis virus reverse transcriptase (Invitrogen). The cDNA product was analyzed by PCR using the primers RT-Tp-1 and RT-F-2 (Table 3) to discover whether icmF and tphA are located on one transcriptional unit, and the primers RT-F-GSP, RT-H-2, and RT-F-2 (Table 3) were used to discover if icmF and icmH are located on the same transcriptional unit.

Southern hybridization.

Legionella chromosomal DNA was prepared as previously described (51). The icmF probe was purified from agarose gel and radiolabeled with [α-³²P]dCTP by the random prime labeling kit (Roche). Hybridization was performed, using a nitrocellulose membrane, at 42°C for ∼16 h in a solution containing 5× SSPE (0.18 M sodium chloride, 10 mM sodium phosphate [pH 7.7], 1 mM EDTA), 2.5× Denhardt's solution (0.02% Ficoll, 0.02% polyvinylpyrrolidone, 0.02% bovine serum albumin), 0.25% sodium dodecyl sulfate (SDS), 150 mg of denatured herring sperm DNA per ml, and 20% formamide. After the hybridization, the filter was washed three times (briefly) in 2× SSPE-0.1% SDS at room temperature and two times with 2× SSPE-0.1% SDS at 42°C for 20 min each time. Then, the membrane was air dried and exposed to X-ray film (Fuji).

RESULTS

Previously, it was reported that the L. pneumophila icm/dot region II contains 17 genes (50). This region was shown to contain 16 icm genes (icmT, -S, -R, -Q, -P, -O, -N, -M, -L, -K, -E, -G, -C, -D, -J, and -B) organized in the same direction (in the order written) and an additional icm gene (icmF) located at the 5′ end of this region and pointing in the opposite direction (50). It was proposed that the icmF gene is organized in the same transcriptional unit with another gene (tphA) located downstream of it that was shown to be dispensable for L. pneumophila intracellular growth (46, 55). Later, signature-tagged mutagenesis performed with L. pneumophila identified an intracellular-growth-defective mutant containing an insertion in a gene located upstream from icmF, but it was not determined if the phenotype was due to polarity on icmF or to the lack of the gene product of the mutated gene (the mutant identified was designated 47:8g) (20).

Proteins homologous to the L. pneumophila IcmF protein are present in several bacteria.

Of the 17 genes characterized in region II (icmT, -S, -R, -Q, -P, -O, -N, -M, -L, -K, -E, -G, -C, -D, -J, -B, and -F), 12 (icmT, -P, -O, -M, -L, -K, -E, -G, -C, -D, -J, and -B) have significant degrees of homology to genes involved in conjugation and are evolutionarily related to the IncI conjugative system present in plasmids such as R64 (37, 56). The five remaining genes (icmS, -R, -Q, -N, and -F) have no homologues on the R64 plasmid. Very recently, in silico analysis of the V. cholerae icmF homologue revealed that the gene is found in many proteobacteria, all of which associate with eukaryotic cells. The analysis indicated that icmF is usually found as one of a large number of genes (between 10 and 15) that have a conserved sequence and organization (17). We found that only one of these genes is present in L. pneumophila and named it icmH (see below). Examination of the gene region that surrounds the icmF- and icmH-homologous genes in the other bacteria revealed that there are several types of organizations for the two genes, and in several bacteria they are found in more than one copy and in more than one form of organization (Fig. 1). The functions of the proteins encoded by these genes in the different bacteria are not known, but one of them was identified previously during a transposon mutagenesis performed on R. leguminosarum. This gene (impJ) (Fig. 1) was shown to be located on the sym plasmid and to be required for nodulation (4, 47). In addition, this gene region was also found in S. enterica subspecies I in centisome 7. The genes in this region (sci genes) (Fig. 1) were found to affect the ability of the bacteria to enter eukaryotic cells (21). Moreover, the expression of an icmF-homologous gene in V. cholerae was found to be induced during infection (15, 16). It is important to mention that in several cases a gene region containing homologues to the L. pneumophila icmF and icmH genes was found in one or two species that belong to the same genus, and the genes were always found in the more pathogenic species. For example, in Y. pestis there are four gene regions that contain icmF- and icmH-homologous genes (21) (two of which are presented in Fig. 1), while in the less pathogenic Yersinia enterocolitica (http://www.sanger.ac.uk/Projects/Y_enterocolitica/) and Yersinia pseudotuberculosis (http://bbrp.llnl.gov/bbrp/bin/y.pseudotuberculosis_blast) there are no homologues of these genes. In E. coli K-12, there are no genes homologous to the icmF and icmH genes (5), while in the pathogenic E. coli O157 these genes are present in the two complete genomes available (28, 45) (Fig. 1). In addition, in S. enterica it was found, using DNA hybridization, that the sci genes are present only in subspecies I (out of seven subspecies), which is the only subspecies that was shown to be associated with warm-blooded organisms (21). All this information indicates that the functions of these genes are probably related to the intimate association of these bacteria with eukaryotic cells.

FIG. 1.

Schematic representations of several gene regions containing icmH and icmF homologues from different bacteria. Genes are represented as boxes with arrowheads indicating their orientations. Homologous genes are shown with the same pattern; open reading frames that have no homologues in the other regions presented are represented by open boxes. Where applicable, only the part of the gene encoding a protein homologous to IcmH is shaded. The gene regions from R. leguminosarum and S. enterica were described before (21, 47), and these genes are identified by their original names (imp and sci, respectively). In Y. pestis, two additional gene regions that contain homologues to icmH and icmF were found (21). In P. aeruginosa, one additional region that contains only the icmF gene was found (21). In E. coli O157, the two genes located upstream from the icmF homologue (wavy-line boxes) are homologous to one gene found in the same location in both V. cholerae and Y. pestis.

The icmF and icmH genes are present in many Legionella species.

As described above, the gene regions containing the icmF and icmH genes were found in one of several species that belong to the same genus or, in the case of S. enterica, in only one subspecies. To examine whether the icmF and icmH genes are also present in other Legionella species besides L. pneumophila (as was shown before for several icm/dot genes [icmD, icmE, icmG, icmX, dotA, and dotB] in Legionella micdadei and for icmX in several Legionella species [34, 40]), we used low-stringency Southern hybridization with the genes. The hybridizations performed with the icmF (Fig. 2) and icmH (data not shown) genes clearly indicate that homologues of the two genes are found in all the Legionella species examined. This result indicates that even though it seems that the genes were incorporated into the icm/dot system from another evolutionary source, in contrast to most of the other icm/dot genes (which probably originated from an IncI plasmid), they are found in all the Legionella species examined. This might indicate the importance of these genes for Legionella intracellular multiplication and fits in with the hypothesis that in nature all the Legionella species are intracellular parasites of amoebae and protozoa, and therefore all of them contain these genes.

FIG. 2.

icmF homologues are present in other Legionella species. Low-stringency Southern hybridization using the L. pneumophila icmF gene as a probe was performed as described in Materials and Methods. The probe was hybridized to EcoRI-digested chromosomal DNA that was prepared from each of the following Legionella species (indicated above the lanes): L. pneumophila serogroup 1 (JR32), L. pneumophila serogroup 3 (ATCC 33155), L. cincinnatiensis (ATCC 43753), L. tucsonensis (ATCC 49180), L. sainthelensi (ATCC 49322), L. gormanii (ATCC 43769), L. birminghamensis (ATCC 43702), L. feeleii (ATCC 35849), L. dumoffii (ATCC 35850), L. dumoffii (ATCC 33343), L. bozemanii (ATCC 33217), L. longbeachae (ATCC 33462), L. longbeachae (ATCC 33484), L. hackeliae (ATCC 35250), L. gratiana (ATCC 49413), L. oakridgensis (ATCC 700515), L. micdadei (ATCC 33204), and L. micdadei (ATCC 33218). The hybridizations performed with the icmH gene gave similar results, and both genes were always located on the same EcoRI fragment (data not shown).

icmH is required for L. pneumophila intracellular multiplication in A. castellanii.

Previously, it was shown that the icmF gene product is required for intracellular growth of L. pneumophila in the protozoan host A. castellanii and partially required for intracellular growth in HL-60-derived human macrophages (55). To determine whether icmH is required for intracellular multiplication as well, a deletion substitution was constructed in it (GS3016). An additional strain that was constructed contains a deletion substitution in the icmF gene (GS3015). This strain was constructed instead of the previously characterized transposon insertion in the icmF gene (LELA1718) because the latter strain was only partially complemented for intracellular growth (46). The deletion substitution mutants in the icmF and icmH genes (GS3015 and GS3016, respectively), were unable to multiply in the protozoan host A. castellanii (Fig. 3A and B). Both of these mutants were complemented to wild-type levels of intracellular growth with a plasmid containing the icmF and icmH genes (pGS-Lc-70-14). A plasmid containing the icmH gene by itself (pGS-Lc-76-14) did not complement the mutant strain in this gene (GS3016), indicating that the deletion substitution in icmH has a polar effect on icmF. However, a plasmid containing an in-frame nonpolar deletion in icmH (pGS-Lc-77-14) was able to complement the deletion substitution in the icmF gene but not the mutant containing the deletion substitution in the icmH gene. These results clearly indicate that icmH is required for intracellular multiplication and that the lack of growth observed with strain GS3016 did not occur due to polarity on icmF alone.

FIG. 3.

The L. pneumophila icmH and icmF genes are required for intracellular growth. Intracellular-growth experiments in the protozoan host A. castellanii (A and B) and in HL-60-derived human macrophages (C and D) were performed as described in Materials and Methods. Shown are icmF mutant GS3015 (A and C) and icmH mutant GS3016 (B and D) containing the following plasmids: pMMB207αb-Km-14 (solid box), icmH-icmF-tphA operon (pGS-Lc-70-14) (solid triangle), icmH gene (pGS-Lc-76-14) (solid circle), and icmF and tphA genes (pGS-Lc-77-14) (open triangle). Solid diamond, wild-type L. pneumophila (JR32); open circle, 25D intracellular defective mutant. The experiments were performed at least three times, and similar results were obtained.

Analysis of the icmH and icmF mutants in HL-60-derived human macrophages.

As was previously shown for the icmF gene (55), the icmH gene was found to be only partially required for intracellular multiplication in HL-60-derived human macrophages (Fig. 3C and D). The lack of both the icmH and icmF genes did not have an additive effect in comparison to the lack of one of them, as the phenotypes observed for the icmF insertion, the icmH insertion expressing the icmF gene, and the icmF insertion containing the vector were the same. These results might indicate that icmF and icmH perform their functions together, as was expected from the bioinformatics analysis.

The partial intracellular-growth phenotype observed with the icmH and icmF mutants in HL-60-derived human macrophages might occur due to several possible defects. As the growth rates observed in the wild-type and the icmH and icmF mutants were similar (Fig. 3C and D), it might be that the 48-h delay in the appearance of the bacteria outside the cells occurred due to a defect in entry, intracellular growth, or exit of the bacteria from the cells (since in the assays performed, bacteria were sampled from the media surrounding the cells). To distinguish these three possibilities, we examined the frequency of entry of the icmH mutant strain into HL-60-derived human macrophages and compared the intracellular and extracellular bacterial numbers during infection, and the results are presented in Fig. 4. As can be seen in Fig. 4A, the icmH mutant had an entry defect, and only 20% of the bacteria entered the cells in comparison to the wild-type strain and the complemented mutant strain. A similar entry phenotype was described before for all the icm/dot mutant strains examined (29). However, this fivefold reduction in entry cannot account for the 2-log-unit difference in the numbers of CFU observed 68 and 92 h postinfection (Fig. 3C and D). One other possibility was that the icmH and icmF mutants were defective in exit from the cells, and due to that, the bacterial numbers in the medium outside the cells were lower than those of the wild-type strain. If this was the situation, the numbers of an icmH mutant inside the cells should have been similar to those of the wild-type strain. However, as can be seen clearly in Fig. 4B, the numbers of the icmH mutant inside the cells 44, 68, and 92 h postinfection were significantly lower (∼2 log units) than those of the wild-type strain and the complemented mutant strain, indicating that the intracellular-growth phenotype observed in HL-60-derived human macrophages did not occur due to a defect in exit from the cells. Therefore, it is most likely that the defect observed with the icmH and icmF mutant strains occurs mainly due to a defect in intracellular growth. However, unlike most of the other icm/dot mutants, the icmH and icmF mutants were able to grow intracellularly to some extent.

FIG. 4.

icmF and icmH are partially required for intracellular growth in HL-60-derived human macrophages. (A) Entry experiments were performed as described in Materials and Methods. The bacteria examined were wild-type L. pneumophila JR32 (W.T.), the icmH mutant GS3016 containing the vector pMMB207αb-Km-14 (icmH) and containing the icmH-icmF-tphA operon (icmH+icmHF). (B) Analysis of intracellular bacteria in HL-60-derived macrophages 44 (open bars), 68 (shaded bars), and 92 (solid bars) h postinfection. The bacterial strains examined are the same as in panel A. The experiments were performed three times, and similar results were obtained. The intracellular growth of the icmH mutant strain was found to be significantly different (P > 0.005) from the intracellular growth of the wild-type strain or the complemented icmH mutant, as determined by the standard t test. The error bars indicate standard deviations.

IcmH and IcmF are partially required for immediate cytotoxicity and salt sensitivity.

Two additional phenotypes that were shown to be associated with mutations in the icm/dot genes are immediate cytotoxicity (36) and salt resistance (49). It was previously shown that L. pneumophila, at a high MOI, rapidly kills host cells in an icm/dot-dependent manner, a phenomenon called immediate cytotoxicity (36). On the other hand, it has been known for a long time that wild-type L. pneumophila is salt sensitive while icm/dot mutants become salt resistant (49). As the icmH and icmF genes seem to have a different evolutionary origin than most of the other icm/dot genes, the deletion substitutions in the icmH and icmF genes were examined for these two phenotypes, as shown in Fig. 5.

FIG. 5.

icmF and icmH are partially required for immediate cytotoxicity and salt sensitivity. (A) Immediate cytotoxicity to HL-60-derived human macrophages was determined as described in Materials and Methods. Solid diamonds, wild-type L. pneumophila (JR32); open triangles, icmT insertion mutant (GS3011); solid triangles, dotA transposon insertion (LELA3118); ×, tphA insertion mutant (MW627); open squares, icmF mutant GS3015 containing the vector pMMB207αb-Km-14; solid squares, icmH-icmF-tphA operon (pGS-Lc-70-14); open circles, icmH mutant GS3016 containing the vector pMMB207αb-Km-14; solid circles, icmH-icmF-tphA operon (pGS-Lc-70-14). (B) The strains indicated in part A were also analyzed for salt sensitivity as described in Materials and Methods. The error bars indicate standard deviations.

The icmH and icmF genes were found to be partially required for immediate cytotoxicity, in comparison to the wild-type strain (JR32) and insertion mutations in the icmT and dotA genes (GS3011 and LELA3118, respectively) (Fig. 5A). When the same strains were analyzed for salt sensitivity, similar results were obtained (Fig. 5B). The icmH and icmF insertion mutants (GS3015 and GS3016, respectively) were found to be partially resistant to sodium, a phenotype that was clearly distinguishable from the degree of sensitivity of the wild-type strain (JR32) and the degree of resistance of the icmT and dotA mutants (Fig. 5B). Both the immediate-cytotoxicity and salt resistance phenotypes were completely complemented when the icmH and icmF genes were supplied on a plasmid (Fig. 5). These results indicate that, similar to what was observed for intracellular growth in HL-60-derived human macrophages, the icmH and icmF genes are partially required for two additional phenotypes related to the icm/dot system.

Analysis of the predicted ATP/GTP binding site of the L. pneumophila IcmF protein.

The N-terminal parts of all the IcmF homologous proteins contain a putative ATP/GTP binding motif (Fig. 6A) that corresponds to the Walker box A nucleotide-binding site ([A/G]XXXXGK[T/S]) found in many ATP-binding proteins (65). In the icm/dot system, four proteins were found to contain this motif (Fig. 6B). To determine the importance of this site for the function of the L. pneumophila IcmF protein, two mutations were constructed in it. The changes were made in 2 amino acids (glycine to serine and lysine to alanine) (Fig. 6C) that are the most conserved amino acids identified in this motif. In addition, the mutations constructed were chosen based on previous reports indicating that such changes result in inactivation of the site (38, 58, 62). The two mutations described were constructed on the complementing plasmid that contains both icmH and icmF (pGS-Lc-70-14), and the plasmids containing the changes (pGS-Lc-70-G2S-14 and pGS-Lc-70-K2A-14) were examined for the ability to complement the icmF deletion substitution mutant (GS3015) for intracellular growth. As can be seen in Fig. 6D, the icmF mutant strain (GS3015) harboring the plasmids containing the mutations was partially attenuated in intracellular growth. The major defect was observed between 48 and 60 h postinfection, when up to 1-log-unit reduction in the number of bacterial CFU in comparison to the wild-type strain (JR32) or the icmF deletion substitution mutant (GS3015) containing the complementing plasmid without the mutations (pGS-Lc-70-14) was observed. These results were reproducible, and they were also observed at different MOIs and always in the same period postinfection (Fig. 6E and data not shown). When the same mutants were examined in HL-60-derived human macrophages, they fully complemented the icmF mutation (data not shown).

FIG. 6.

The IcmF ATP/GTP binding site is partially required for intracellular growth. (A) Predicted ATP/GTP binding sites in IcmF-homologous proteins. The bacteria, the sequences of the predicted ATP/GTP binding sites, and the amino acid positions are indicated. The conserved amino acids ([A/G]XXXXGK[T/S]) are in boldface. (B) Predicted ATP/GTP binding sites of L. pneumophila Icm/Dot proteins, presented as in panel A. (C) Mutations constructed in the L. pneumophila IcmF predicted ATP/GTP binding site; the changes constructed are underlined. (D) Intracellular-growth experiments in the protozoan host A. castellanii were performed as described in Materials and Methods. Solid diamonds, wild-type L. pneumophila (JR32); asterisks, icmF mutant GS3015 containing the vector pMMB207αb-Km-14; solid squares, wild-type icmF gene (pGS-Lc-70-14); open circles, icmF gene containing the K-to-A mutation (pGS-Lc-70K2A-14); open triangles, icmF gene containing the G-to-S mutation (pGS-Lc-70G2S-14). (E) Analysis of intracellular growth of the strains shown in panel D. Solid bars, wild-type L. pneumophila (JR32); stippled bars, icmF mutant GS3015 containing the wild-type icmH-icmF-tphA operon (pGS-Lc-70-14); shaded bars, same plasmid containing the K-to-A mutation in IcmF (pGS-Lc-70K2A-14); open bars, same plasmid containing the G-to-S mutation in IcmF (pGS-Lc-70G2S-14). The experiment was performed at different MOIs, and the results obtained 48 h postinfection are presented. Similar differences were obtained 60 h postinfection (data not shown). The error bars indicate standard deviations. The intracellular growth of the icmF mutant complemented with the mutated ATP/GTP binding site was found to be significantly different (P > 0.0001) from the intracellular growth of the wild-type strain or the icmF mutant complemented with the wild-type genes at all the MOIs examined, as determined by the standard t test.

mRNA analysis of the icmH-icmF-tphA transcriptional unit.

The finding that in V. cholerae the icmF gene was activated during infection (16) led us to examine the icmH regulatory region carefully. Using RT analysis, we found that icmH, icmF, and tphA are located on the same transcriptional unit (data not shown), as was expected from the complementation experiments and the gene organizations. In addition, using the RACE system, the transcription start site of the icmH-icmF-tphA transcriptional unit was determined and found to be located 52 bp upstream from the first ATG of the icmH gene (Fig. 7A). This result indicates that the −10 promoter element of this transcriptional unit is unique among the icm genes, as it does not contain the conserved TATACT consensus sequence that was identified in the other icm gene regulatory regions (24).

FIG. 7.

Truncation analysis of the icmH-icmF-tphA operon regulatory region. (A) Sequences of the icmH-icmF-tphA regulatory region and the truncations constructed. The first nucleotide of each truncation is marked with an asterisk, and the truncation number is indicated above it. The transcription start site (TSS) is in boldface and underlined, the first ATG codon of the IcmH protein and the predicted promoter sequence are in boldface, and the single nucleotide that was mutated (T to G) is underlined (Fig. 8A). (B and C) β-Galactosidase activities of different truncations (the length of the regulatory region of each truncation is given in Table 2) in L. pneumophila during exponential growth. β-Galactosidase activity was measured as described in Materials and Methods. The results are the averages ± standard deviations of at least three different experiments. The β-galactosidase activities of the F14 and F11 truncations were found to be significantly different (P > 0.0001), as determined by the standard t test.

Analysis of the regulatory region of the icmH-icmF-tphA transcriptional unit.

To obtain additional information about the regulation of the icmH-icmF-tphA transcriptional unit, truncation analysis was performed (Fig. 7A). As can be seen clearly from the results presented in Fig. 7B and C, there were three points where the level of expression of the icmH::lacZ fusions dropped dramatically. (i) Between truncations F6 and F12, the deletion of 11 nucleotides (from 251 to 240 bp upstream) (Fig. 7A) reduced the level of expression from ∼12,000 to ∼2,500 Miller units (MU). (ii) Between truncations F13 and F14, the deletion of 11 nucleotides (from 137 to 126 bp upstream) (Fig. 7A) reduced the level of expression from ∼2,500 to ∼70 MU. (iii) Between truncations F10 and F11, the deletion of 14 nucleotides (from 116 to 102 bp upstream) (Fig. 7A) reduced the level of expression to ∼30 MU (Fig. 7C), and it got very close to the promoter region (Fig. 7A) and to the level of expression of the vector control (data not shown). These results indicate that there are at least two additional regulatory elements located in the icmH regulatory region besides the promoter element.

In a previous study, a regulatory element (CTATAGAAT) located between the F10 and F11 truncations was identified, and one nucleotide substitution in this site (constructed on the F4 icmH::lacZ fusion) completely abolished the expression of the fusion (24). To determine the relationship between this site and the potential regulatory element located between the F6 and F12 truncations, the same mutation was introduced into the F5 icmH::lacZ fusion. As can be seen in Fig. 8A, the mutation constructed completely abolished the expression from the F4 icmH::lacZ fusion, but it reduced the level of expression from the F5 icmH::lacZ fusion by ∼2,500 MU. This result indicates that the upstream regulatory element (probably located between the F6 and F12 icmH::lacZ fusions) is probably independent of the site mutated (located between the F10 and F11 icmH::lacZ fusions), since the level of expression of the F5 icmH::lacZ fusion was reduced only by the portion contributed by the downstream regulatory element. Comparison of the level of expression of the F5 icmH::lacZ fusion to the level of expression of the other icm::lacZ fusion (24) clearly shows that the icmH-icmF-tphA transcriptional unit has the highest level of expression of all the icm genes and operons examined.

FIG. 8.

Regulatory elements and factors affecting expression of the icmH::lacZ fusion. (A) Effect of a single-base-pair substitution (Fig. 7A) on the levels of expression of the F4 and F5 truncations. Expression from the F5 and F4 truncations (open bars) was compared to the expression of the fusions containing the mutation indicated in Fig. 7A (shaded bars). The β-galactosidase activities of the F5 truncation and the mutated F5 truncation were found to be significantly different (P > 0.0001), as determined by the standard t test. (B) Effects of growth phase and regulators related to stationary phase on the level of expression of the F5 truncation. The level of expression of truncation F5 was examined at exponential (open bars) and stationary (shaded bars) phases in the wild-type (WT) strain, the relA mutant strain GS-RelA (RelA), the rpoS mutant strain LM1376 (RpoS), and the letA mutant strain OG2001 (LetA). β-Galactosidase activity was measured as described in Materials and Methods. The results are the averages ± standard deviations of at least three different experiments. The β-galactosidase activities of the F5 truncation at exponential and stationary phases were found to be significantly different (P > 0.0001), as determined by the standard t test.

The expression of the icmH-icmF-tphA operon is influenced by RelA and LetA.

Because the regulation of genes required for intracellular growth was postulated to be activated during stationary phase (2, 26, 27, 43), we examined the level of expression of the F5 icmH::lacZ fusion during stationary phase in comparison to exponential growth phase. As can be seen in Fig. 8B, the level of expression of this fusion increased at stationary phase (from ∼12,000 MU at exponential phase to ∼15,000 MU at stationary phase). This result led us to examine whether regulators related to gene expression during stationary phase are required for this phenomenon. When the icmH::lacZ fusion was examined in relA and letA insertion mutants (GS-RelA and OG2001, respectively), no increase was observed at stationary phase, while the same increase was observed when the fusion was examined in an rpoS insertion mutant (LM1376) (Fig. 8B). These results clearly indicate that the icmH-icmF-tphA transcriptional unit is regulated by RelA and LetA, which were shown previously to be part of one regulatory circuit in L. pneumophila (43). When additional regulators (CpxR, CpxA, and RpoE) were examined to see if they were involved in the regulation of the icmH-icmF-tphA transcriptional unit, no effect was observed (data not shown).

DISCUSSION

Thus far, 25 icm/dot genes required for intracellular multiplication have been described and characterized in L. pneumophila. The majority of the Icm/Dot proteins (18 proteins) encoded by the icm/dot genes are homologous to proteins involved in conjugation encoded by IncI plasmids (37, 56). The finding that most of the icm/dot genes probably constitute part of the secretion complex directed most of the research to the icm/dot genes that have no homologues in conjugative systems (icmN, icmS, icmW, icmQ, icmR, icmF, and icmH). Besides the icmN gene product, which has many homologous proteins in other bacteria, the six other proteins can be divided into pairs.

The IcmS and IcmW proteins were found to interact with one another, and this property was also found to be conserved in the IcmS and IcmW homologues from Coxiella burnetii (10, 73). These two proteins were found not to be required for pore formation (10), and null mutants in each of them, as well as in both of them together, can still replicate to some extent in HL-60-derived human macrophages (10, 52, 71, 73). The specific functions of these two genes and their relationship to the rest of the Icm/Dot system are not known.

The IcmQ and IcmR proteins were also found to interact with one another. The IcmQ protein was found to form homopolymers, and the IcmR protein was shown to possess chaperon activity for IcmQ and to prevent its polymerization (10, 18). In addition, an icmR insertion mutant was found to retain some small ability to multiply inside host cells and to kill HL-60-derived human macrophages (10, 52). Very recently, it was reported that the IcmQ protein is exposed on the surfaces of bacteria after contact with macrophages and that it forms pores in lipid membranes, and the pore formation was shown to be inhibited by IcmR (19). Surprisingly, the exposure of the IcmQ protein on the surfaces of bacterial cells was found to be independent of the other icm/dot components (19). However, mutants in the icmR and icmQ genes become salt resistant, indicating that they are connected to the icm/dot system (49, 52).

The IcmF and IcmH proteins are the subjects of this report, and our results indicate that icmH and icmF probably perform their function together, because the phenotype of a mutant lacking both genes was similar to that of a mutant that lacked either gene (possible interactions between IcmH and IcmF were examined using a bacterial two-hybrid system, but no interactions could be obtained, perhaps due to their expected membrane location [T. Zusman, G. Segal, unpublished results]). In addition, the information obtained from the analysis of immediate cytotoxicity indicates that bacteria lacking these genes are still able to translocate a pore into host cells, but less efficiently than the wild-type strain. Moreover, the partial phenotype for salt resistance gives further strength to the assumption that a functional icm/dot system is present in these bacteria, as salt resistance is a property thought to be associated with a defective icm/dot secretion system (64). Both these results and the partial phenotype for intracellular multiplication in HL-60-derived human macrophages indicate that IcmH and IcmF are probably not integral parts of the icm/dot system, because a partially functional icm/dot system is present in bacteria that lack both proteins. Additional support for this assumption comes from analysis of the icmH-icmF-tphA regulatory region, which was found to contain a different promoter sequence than the other icm/dot genes and operons. These genes were shown to contain a conserved −10 promoter element (24), which is missing in the icmH regulatory region.

As indicated above, the evolutionary origins of the majority of the Icm/Dot proteins are probably proteins involved in conjugation of IncI plasmids, and these Icm/Dot proteins are most likely the major components of the Icm/Dot type IV secretion complex. One interesting question concerning the seven Icm/Dot proteins that have no homologues on IncI plasmids is from where do they originate. In comparison to the other Icm/Dot proteins, the IcmH and IcmF proteins are unique in the sense that they have homologous proteins in several bacteria that live in close contact with eukaryotic cells (Fig. 1). This finding is in contrast to the other Icm/Dot proteins which do not contain homologous proteins encoded by IncI plasmids. Proteins homologous to IcmS, IcmW, and IcmQ were found only in C. burnetii, which contains all the Icm/Dot proteins except IcmR, which has no known homologue (73). In the other bacteria in which proteins homologous to IcmF and IcmH were found, these proteins were encoded by genes in a large region that contained between 10 and 15 genes; however, L. pneumophila does not contain homologues to the proteins other than IcmH and IcmF. In three cases, this large gene region was shown to participate in the interaction of the bacteria with eukaryotic cells (4, 15, 16, 21, 47). This information might indicate that the Icm/Dot proteins probably originated from at least three different evolutionary sources: (i) 18 Icm/Dot proteins probably originated from an IncI conjugative plasmid; (ii) IcmH and IcmF probably originated from a common ancestral system, from which L. pneumophila and the other bacteria obtained their icmH and icmF homologues; (iii) IcmWS and IcmRQ have unknown evolutionary origins, but since there are no homologues of these proteins in the other two systems described, it is most likely that these proteins have a third (or third and fourth) evolutionary origin.

Among the bacteria that contain proteins homologous to the IcmH and IcmF proteins, there are bacteria that use type III secretion systems (E. coli O157 and S. enterica) or type IV secretion systems (L. pneumophila and B. pertussis) for pathogenesis and there are human and animal pathogens (Y. pestis, and V. cholerae) and plant pathogens (A. tumefaciens and Pseudomonas fluorescens), as well as plant symbionts (R. leguminosarum and Mesorhizobium loti). Deeper analysis of the L. pneumophila icmH and icmF genes, as well as of homologous genes in the other bacteria, might reveal their importance for the association of these bacteria with eukaryotic cells and the relationship of these genes to the functions of the main systems that participate in the interactions of the bacteria with eukaryotic cells.