|Home | About | Journals | Submit | Contact Us | Français|
Most bacteria synthesize isoprenoids through one of two essential pathways which provide the basic building block, isopentyl diphosphate (IPP): either the classical mevalonate pathway or the alternative non-mevalonate 2-C-methyl-d-erythritol 4-phosphate (MEP) pathway. However, postgenomic analyses of the Listeria monocytogenes genome revealed that this pathogen possesses the genetic capacity to produce the complete set of enzymes involved in both pathways. The nonpathogenic species Listeria innocua naturally lacks the last two genes (gcpE and lytB) of the MEP pathway, and bioinformatic analyses strongly suggest that the genes have been lost through evolution. In the present study we show that heterologous expression of gcpE and lytB in L. innocua can functionally restore the MEP pathway in this organism and confer on it the ability to induce Vγ9Vδ2 T cells. We have previously confirmed that both pathways are functional in L. monocytogenes and can provide sufficient IPP for normal growth in laboratory media (M. Begley, C. G. Gahan, A. K. Kollas, M. Hintz, C. Hill, H. Jomaa, and M. Eberl, FEBS Lett. 561:99-104, 2004). Here we describe a targeted mutagenesis strategy to create a double pathway mutant in L. monocytogenes which cannot grow in the absence of exogenously provided mevalonate, confirming the requirement for at least one intact pathway for growth. In addition, murine studies revealed that mutants lacking the MEP pathway were impaired in virulence relative to the parent strain during intraperitoneal infection, while mutants lacking the classical mevalonate pathway were not impaired in virulence potential. In vivo bioluminescence imaging also confirmed in vivo expression of the gcpE gene (MEP pathway) during murine infection.
To date, more than 30,000 isoprenoid molecules are known and have been shown to be involved in diverse processes. Well-known examples include the sterols of eukaryotes, which act as membrane stabilizers and are precursors of bile acids, carotenoids, which form part of the photosynthetic machinery of plants, and bactoprenols, which act as carbohydrate carriers in the biosynthesis of bacterial peptidoglycan (4).
Despite their diversity of structure and function, all isoprenoids derive from the five-carbon building unit isopentyl diphosphate (IPP) and its isomer, dimethylallyl diphosphate (DMAPP). IPP and DMAPP are synthesized by one of two distinct pathways: either the classical mevalonate pathway or the alternative non-mevalonate 2-C-methyl-d-erythritol 4-phosphate (MEP) pathway. These pathways have been the focus of several elegant studies in recent years, and the enzymes involved have been elucidated (4, 17, 27). Biochemical studies have revealed that an intermediate of the MEP pathway, namely, (E)-4-hydroxy-3-methyl-but-2-enyl pyrophosphate (HMB-PP), is a potent activator of human Vγ9Vδ2 T cells (1, 2, 9, 15, 26). Elevated levels of this subset of T cells in the peripheral blood of patients have been documented in a variety of microbial infections, such as salmonellosis, listeriosis, brucellosis, and tuberculosis (7, 19). Their precise role is unknown, but it has been postulated that they play a role in clearance and protective immunity or may be involved in immune evasion of the pathogen and establishment of chronic disease (10). As humans use only the mevalonate pathway, the MEP pathway is considered a novel target for the development of antibacterial drugs (18, 28).
Analysis of the bacterial genomes sequenced to date reveals that the gram-positive pathogen Listeria monocytogenes and certain Streptomyces species are unusual in containing genes for both pathways. In all other organisms examined thus far these pathways are mutually exclusive (reviewed in reference 4). Listeria innocua, which is closely related to L. monocytogenes but is nonpathogenic, possesses the first five of the seven genes of the MEP pathway but lacks genes encoding the final two enzymes (GcpE and LytB). In the current communication we describe our in depth in silico analyses of the gcpE and lytB genomic regions in both Listeria species and the possibility of restoring the MEP pathway in L. innocua, and the effect on Vγ9Vδ2 T-cell activation is examined. The essentiality of isoprenoid biosynthesis pathways in L. monocytogenes is investigated through the creation of a double pathway mutant. Finally, the contribution of the two pathways to the pathogenesis of L. monocytogenes is examined using the mouse model of infection.
The following genome sequences were accessed through the National Center for Biotechnology Information (NCBI) Genome database (http:/www.ncbi.nlm.nih.gov/): L. monocytogenes EGDe (accession number NC_003210), L. monocytogenes 10403S (accession number NZ_AARZ00000000), L. monocytogenes strain 4b F2365 (accession number NC_002973), L. monocytogenes strain 1/2a F6854 (accession number NZ_AADQ00000000), L. monocytogenes strain 4b H7858 (accession number NZ_AADR00000000), L. monocytogenes HPB2262 (accession number NZ_AATL00000000), L. monocytogenes J2818 (accession number NZ_AARX00000000), L. monocytogenes F6900 (accession number NZ_AARU00000000), L. monocytogenes FSL N3-165 (accession number NZ_AARQ00000000), L. monocytogenes LO28 (accession number NZ_AARY00000000), L. innocua CLIP11262 (accession number NC_003212), and Bacillus subtilis subsp. subtilis strain 168 (accession number NC_000964). Sequences were analyzed using various Web-based programs by using the ExPASy server (http://www.expasy.ch/tools/). The Web-based program δρ-Web was used for compositional bias of dinucleotide frequency analysis (http://deltarho.amc.nl). Genomic dissimilarity (δ*) values were calculated. A high δ* value between an input sequence and the host genome sequence indicates a heterologous origin of the input sequence. Predicted promoter regions were analyzed using BPROM (http://www.softberry.com/berry.phtml?topic=bprom&group=programs&subgroup=gfindb). Alignments were performed with CLUSTALW (http://www.ebi.ac.uk/Tools/clustalw/index.html). Phylogenetic and molecular evolutionary analyses were conducted using MEGA4 (http://www.megasoftware.net/). Phylogenetic trees were constructed using neighbor-joining algorithms, and bootstrap analyses were performed with 1,000 replicates.
The strains, plasmids, and primers used in this study are listed in Table Table1.1. Escherichia coli strain DH10B (Invitrogen, Paisley, United Kingdom) and Bacillus subtilis 168 (20) were grown aerobically at 37°C in Luria-Bertani medium (29). Listeria strains were grown aerobically at 37°C in brain heart infusion (BHI) medium (Oxoid, Basingstoke, United Kingdom). When appropriate, antibiotics were added to the medium: for E. coli, chloramphenicol (15 μg/ml) and erythromycin (250 μg/ml); for L. monocytogenes, chloramphenicol (7.5 μg/ml) and erythromycin (5 μg/ml); for L. innocua, chloramphenicol (7.5 μg/ml).
Plasmid DNA was isolated from E. coli using the QIAprep Spin Miniprep kit according to the manufacturer's instructions (Qiagen, Crawley, United Kingdom). Genomic DNA was isolated from L. monocytogenes EGDe and B. subtilis 168 using the Genelute bacterial genomic DNA kit (Sigma, Steinheim, Germany) following the recommendations of the manufacturer. Primers were purchased from Sigma Genosys (Haverhill, United Kingdom). PCR products required for cloning were obtained with KOD hot start high-fidelity DNA polymerase (Merck, Nottingham, United Kingdom) using 10 ng genomic or plasmid DNA as template. Standard procedures were applied for DNA manipulations in E. coli (29). Restriction endonucleases (Roche Diagnostics, Mannheim, Germany), T4 DNA ligase (Roche), and 2× PCR mix (Promega, Madison, WI) were used following the recommendations of the manufacturers. Transformations of L. monocytogenes and L. innocua were performed as described previously (25).
A mutant in which 0.9 kb of the coding region of lmo1441 (gcpE) was deleted from L. monocytogenes EGDe was created using a construct generated by the splicing by overlap extension (SOE) procedure (16). Primer pairs were designed (lmo1441SOEA-lmo1441SOEB and lmo1441SOEC-lmo1441SOED) to amplify 0.35-kb products flanking the lmo1441 (gcpE) gene. The resulting products were mixed in a 1:1 ratio and reamplified using primer lmo1441SOEA and lmo1441SOED. The resulting 0.7-kb product was digested with PstI and XbaI (restriction sites introduced on the lmo1441SOEA and lmo1441SOED primers) and cloned into the temperature-sensitive plasmid pKSV7 (32), previously digested with the same restriction enzymes. The resulting plasmid was designated pKSV7-lmo1441 and was electroporated into L. monocytogenes EGDe at 30°C. Chromosomal integration of the plasmid was established by overnight growth at 42°C in BHI broth containing chloramphenicol and spread plating onto BHI agar containing chloramphenicol. Single colonies were selected, and plasmid excision and curing were brought about by continuous passaging at 30°C in BHI broth without chloramphenicol, followed by spread plating onto BHI plates. Chloramphenicol-sensitive colonies were identified and a PCR, using template DNA obtained directly from the colonies by a microwave treatment (3 min, 600 W) and the primer pair lmo1441out and lmo1441outII confirmed the deletion event. One colony was selected for further analysis and was designated EGDeΔgcpE.
To create a double pathway mutant in which both the MEP and the mevalonate pathways are blocked, a similar SOE procedure (16) was employed to replace 1.2 kb of the coding region of lmo0825 (hmgR) in EGDeΔgcpE with the erythromycin resistance gene (ery). The primer pairs lmo0825A-lmo0825B and lmo0825C-lmo0825D were used to amplify 0.4-kb flanking regions of lmo0825, and the primer pair eryF-eryR was used to amplify a 1.1-kb fragment harboring ery from pRV300 (23) template DNA. The three PCR products were mixed and reamplified using the primer pair lmo0825A and lmo0825D. The resulting 1.9-kb fragment was digested with XbaI-HindIII and cloned into similarly digested pKSV7. The resulting plasmid was designated pKSV7-lmo0825ery and transformed into EGDeΔgcpE. Plasmid integration, excision, and curing were established as described above. However, erythromycin was added to all media during these procedures, allowing direct selection against plasmid excision, resulting in wild-type chromosomal organization. Moreover, mevalonate was added at all times, providing a route to synthesize the essential molecule IPP in the double pathway mutant. Candidate mutants were tested for their chloramphenicol phenotype, and sensitive colonies were checked directly in a PCR, using the primer pair lmo0825out and lmo0825outII. One colony for which the PCR confirmed the replacement event was designated EGDe-ΔhmgRΔgcpE.
The L. monocytogenes double pathway mutant was complemented using the listerial site-specific integration vector pPL2a (5, 22). lmo0825 was amplified from L. monocytogenes EGDe chromosomal DNA using the primer pair 0825F-0825R. The 1.5-kb amplicon was digested with XbaI-SalI and ligated to similarly digested pPL2a. The resulting plasmid was designated pPL2-lmo0825 and was transformed into EGDe-ΔhmgRΔgcpE. Chromosomal integration of the plasmid was confirmed by a colony PCR using the primer pair PL95-PL102 (22).
The luciferase-based reporter system pPL2lux (5) was utilized to investigate whether the lmo0010 (mvk), lmo0825 (hmgR), lmo1441 (gcpE), and lmo1451 (lytB) genes are expressed during infection. Chromosomal DNA from L. monocytogenes EGDe was used as template in PCRs using the primer pairs lmo0010luxF and -R, lmo0825F and -R, lmo1441F and -R, and lmo1451F and -R (Table (Table1).1). The resulting PCR products, harboring the 0.5-kb promoter regions immediately upstream of the corresponding genes, were digested with XhoI and cloned into SwaI-SalI-digested pPL2lux, as described previously (5). The resulting plasmids (Table (Table1)1) harbor exact translational fusions of the promoter regions to the luxABCDE operon, allowing analysis of promoter activities by monitoring bioluminescence levels. Subsequently, the pPL2lux derivatives were transformed into L. monocytogenes EGDe, and candidate integrants were checked for site-specific integration by colony PCR, using the primer pair PL95-PL102. From each transformation, one colony possessing the anticipated genotype was selected for further analysis (Table (Table11).
The gcpE and lytB homologues of B. subtilis 168 were introduced into L. innocua using plasmid pNZ44 (24). Chromosomal DNA of B. subtilis 168 and the primer pairs gcpEF-gcpER and lytBF-lytBR were used to amplify 1.2- and 1.1-kb fragments harboring BG11671 (gcpE homologue) and BG11662 (lytB homologue), respectively. Firstly, pNZ44 was digested with Ecl136II and blunt end ligated with the BG11671 amplicon. The ligation mixtures were digested with Ecl136II prior to transformation to E. coli DH10b (Invitrogen) to reduce the number of transformants harboring back-ligated vector DNA. After 40 h, full-grown colonies were used directly as templates in a PCR using the primer pair P44F-gcpER. This PCR generates a product only when an insert is cloned in the orientation that allows transcription of BG11671 through the constitutive, lactococcal P44 promoter. One clone that generated the expected amplicon was designated pNZ44-gcpE. This plasmid was digested with KpnI and ligated to the BG11662 PCR product previously digested with the same enzyme. The appropriate orientation in which the BG11662-BG11671 transcription is driven from the P44 promoter was selected for by a PCR strategy using the primer pair P44F-lytBR, yielding pNZ44-lytB-gcpE. Both pNZ44-gcpE and pNZ44-lytB-gcpE were electroporated into L. innocua (Table (Table11).
Growth of the double pathway mutant in the presence and absence of mevalonate was investigated. A stock solution of 1 M mevalonate was prepared by hydrolyzing dl-mevalonolactone (M4667; Sigma Aldrich, Dublin, Ireland) with 1 M NaOH at 37°C for 1 h (9). Mevalonate at 1 mM was added to broth when required. The cells were pelleted by centrifugation (7,000 × g for 5 min), washed three times with phosphate-buffered saline (PBS), and diluted 1:100 in fresh BHI medium with or without mevalonate. Subsequently, growth at 37°C was monitored over time in triplicate for all strains, using flat-bottom 96-well plates (Sarstedt, Newton, MA), and measurements of the optical density at 600 nm (OD600) in a Genios microtiter plate reader (Tecan, Salzburg, Austria).
Overnight cultures of the L. monocytogenes strains were pelleted by centrifugation (7,000 × g for 5 min), washed twice with PBS, and used to inoculate 8- to 12-week-old BALB/c mice intraperitoneally with 2 × 105 CFU in 200 μl of PBS. Three days postinfection, the mice were sacrificed by cervical dislocation. Livers and spleens were homogenized in PBS, and serial dilutions were plated onto BHI agar, followed by overnight incubation at 37°C. Full-grown colonies were counted, and the results were used to determine the bacterial load in the organs. When bioluminescent strains were used, bioluminescent imaging was conducted on the individual organs using 5-min exposure times at a binning value of 4 on an IVIS 100 system (Xenogen, Alameda, CA). Murine studies were carried out according to institutional guidelines.
J774.A1 mouse macrophage cells were routinely grown in antibiotic-free Dulbecco modified Eagle medium (DMEM) plus Glutamax (Gibco) supplemented with 10% fetal calf serum (Sigma) in T75 flasks (Sarstedt) and incubated at 37°C with 5% CO2. For infection assays, macrophage cells were grown to confluence in 24-well tissue culture plates. One ml of overnight bacterial culture was washed once in PBS and diluted 1/100 in DMEM to a final concentration of 3 × 107 CFU/ml. DMEM was removed from each tissue culture well, and 1 ml of DMEM containing bacteria was added. Plates were centrifuged at 1,500 × g for 10 min to increase contact between the bacteria and macrophages and incubated for 1 h in 5% CO2 at 37°C. Gentamicin (15 μg/well; Sigma) was added for 30 min. Bacterial counts were determined at this point (T0) and after 6 h (T6) by washing the cells twice with 1 ml PBS followed by lysis using 250 μl of cold sterile distilled water containing 0.1% Triton X-100. A 100-μl aliquot was serially diluted and plated onto BHI agar plates, which were incubated at 37°C overnight. The double pathway mutant was starved of mevalonate prior to experiments by inoculating cultures which were grown overnight in BHI supplemented with mevalonate into fresh BHI without mevalonate and incubated shaking at 37°C until the OD600 remained steady (approximately 6 h; see Fig. Fig.4B4B below). The double mutant was plated onto BHI agar supplemented with 1 mM mevalonate at all times.
RNA was isolated from 10 ml of full-grown cultures of L. innocua(pNZ44), L. innocua(pNZ44-gcpE), and L. innocua(pNZ44-lytB-gcpE) using the protocol described previously (5). Subsequently, 2 μg of the isolated RNA was subjected to DNase I treatment according to the manufacturer's instructions (Ambion, Huntingdon, United Kingdom). cDNA was synthesized in a 40-μl reaction mixture containing 0.2 nmol of random hexamer primer, 10 mM dithiothreitol, 1× Expand reverse transcriptase buffer, 50 U Expand reverse transcriptase (Roche Diagnostics, Mannheim, Germany), each deoxynucleoside triphosphate at a concentration of 0.25 mM, and 1.5 μg DNase I-treated RNA. Reverse transcription was initiated for 10 min at 30°C, followed by 45 min at 42°C and inactivation of the reverse transcriptase by incubation at 95°C for 2 min. Subsequently, 1 μl of the synthesized cDNA was used in PCR mixtures containing 1× PCR mix (Promega) and 1 pmol of primer pairs (IM109-IM110 for 16S RNA, BG11662F-BG11671R for gcpE-lytB, and BG11671F-BG11671R for gcpE) (Table (Table11).
γδ T-cell stimulation assays were performed essentially as described previously (3). Low-molecular-weight (LMW) extracts were prepared by sonication of grown cultures and subsequent ultrafiltration over Centriprep 3-kDa filters (Millipore, Eschborn, Germany). The extracts were tested for stimulation of human Vγ9Vδ2 T cells in peripheral blood mononuclear cell preparations by using flow cytometry. The results are expressed as the percentage of Vγ9+ cells among total CD3+ cells.
Genes for both pathways of isoprenoid biosynthesis are present in the L. monocytogenes EGDe genome (Fig. (Fig.1).1). All genes are also present in the genomes of the other 10 L. monocytogenes strains examined (genomes and accession numbers are listed in Materials and Methods). Analysis of the genome of L. innocua revealed that this closely related but nonpathogenic species possesses all mevalonate pathway loci but has the genes encoding only the first five enzymes of the MEP pathway (1-deoxy-d-xylulose 5-phosphate [DOXP]; MEP DOXP synthase, DOXP reductoisomerase, 4-diphosphocytidyl-2-C-methyl-d-erythritol [CDP-ME] synthase, CDP-ME kinase, and 2-C-methyl-d-erythritol 2,4-cyclopyrophosphate [MEcPP] synthase). Genes encoding the final two enzymes, HMB-PP synthase (GcpE) and HMB-PP reductase (LytB), are absent from its genome (Fig. (Fig.2A2A).
Further in silico analysis was carried out in order to determine whether L. monocytogenes may have acquired the two genes or, alternatively, whether the nonpathogenic species may have deleted them. The GC content of the genes was found to be similar to the overall chromosomal content (39.76% for lmo1451 and 39.11% for lmo1441, compared to a genome average of 39%). There is no evidence of a transposon insertion site or transposon-related repeat structures in this region of the chromosome. Genomic dissimilarity values (δ* values, calculated by δρ-Web) for both genes were low, suggesting that these genes are not likely to have been acquired by horizontal gene transfer. Furthermore, the computer-based approaches employed by Hain et al. (14) to detect genes of foreign origin in the L. monocytogenes genome did not designate either gene as “alien.”
As no evidence of gene acquisition by L. monocytogenes was found, efforts focused on examining the two gene sequences and surrounding genomic regions for evidence of gene loss by L. innocua. Close examination of the nucleotide sequence between lin1479 and lin1480 revealed that L. innocua possesses the putative promoter and ribosomal binding site of gcpE. Vestigial sequence encoding the first 3 amino acids of GcpE were also identified; however, these codons were followed by a stop codon, and the remainder of the gene was not present (Fig. (Fig.2B).2B). Inspection of the lytB region of the genome indicates that L. innocua still possesses the putative promoter region, as predicted by BPROM; however, none of the coding sequence could be identified (Fig. (Fig.2B).2B). Finally, phylogenetic analyses of both lytB and gcpE across the sequenced L. monocytogenes strains suggest that these loci have evolved at the same rate as the native housekeeping gene gyrB while the same genes in Bacillus subtilis have significantly diverged through evolution (data not shown).
To conclude, our bioinformatics analyses provide very strong evidence that L. innocua has lost gcpE and lytB through evolution yet has retained the remaining five genes of the MEP pathway.
We have previously shown that L. innocua is not capable of activating Vγ9Vδ2 T cells (3). The possibility of restoring the MEP pathway and introducing the ability to activate Vγ9Vδ2 T cells was assessed by introduction of plasmid pNZ44 containing gcpE (pNZ44-gcpE) or gcpE together with lytB (pNZ44-lytB-gcpE). L. innocua was also transformed with an empty vector to serve as a control. To verify the production of the anticipated transcripts, RNA was isolated from all three strains, followed by cDNA synthesis. Control PCRs were performed with the 16S RNA primer pair and confirmed that cDNA concentrations were comparable in all samples. PCRs with gcpE primers generated an amplicon only when cDNA derived from the L. innocua(pNZ44-gcpE) or L. innocua(pNZ44-lytB-gcpE) cultures was used as template, whereas the gcpE/lytB primer pair generated a PCR product only when the cDNA corresponding to L. innocua(pNZ44-lytB-gcpE) was used (Fig. (Fig.3A).3A). These results demonstrate that the introduction of the pNZ44-gcpE and pNZ44-lytB-gcpE plasmids into L. innocua results in the functional production of a gcpE and lytB-gcpE transcript, respectively. Subsequently, the strains were tested in a Vγ9Vδ2 T-cell assay. With LMW extracts derived from the control strain L. innocua(pNZ44), no significant Vγ9Vδ2 T-cell activity was detected, corroborating our earlier findings which suggested that L. innocua is not capable of activating Vγ9Vδ2 T cells (3). Using LMW extracts isolated from L. innocua(pNZ44-gcpE), the Vγ9Vδ2 T-cell activity was even detectable at extremely high dilutions (1:2,430,000) (Fig. (Fig.3B).3B). This experiment indirectly indicates the synthesis of very large amounts of HMB-PP and suggests the presence of a functional set of enzymes leading to the substrate of HMB-PP synthase, namely, MEcPP. The absence of HMB-PP reductase prevents the conversion of HMB-PP into IPP, resulting in HMB-PP accumulation and, as a consequence, very strong Vγ9Vδ2 T-cell activation. With the LMW extracts derived from L. innocua-pNZ44-lytB-gcpE, activity was observed which was detectable at the lower dilutions (1:10,000 and 1:30,000) (Fig. (Fig.3B).3B). The lower activities achieved may be explained by the fact that this strain now contains a complete MEP pathway and the HMB-PP produced by HMB-PP synthase is likely to be rapidly consumed by HMB-PP reductase into IPP. It can be inferred from these results that the first set of enzymes of the MEP pathway present in L. innocua are functional and we can resume the discussion that the intermediates of the MEP pathway are involved in other functions beside the synthesis of isoprenoid precursors. Furthermore, these results provide proof of principle that it is feasible to reintroduce both the MEP pathway and Vγ9Vδ2 T-cell activation into a bacterium lacking a complete MEP pathway.
We have previously mutated two genes of each pathway in L. monocytogenes, namely, lmo0010 (mvk), lmo0825 (hmgR), lmo1441 (gcpE), and lmo1451 (lytB) (3). These mutants were not impaired in growth in BHI medium, suggesting that both pathways are functional and that disruption of one pathway can be complemented by the action of the other. In order to analyze the cellular requirement for both pathways in the creation of the essential compound IPP, we created a double mutant in which the two pathways were inactivated in a single strain. The temperature-sensitive plasmid pKSV7 was used to create a deletion in the lmo1441 gene of the MEP pathway (see Materials and Methods). Subsequently, the lmo0825 gene of the mevalonate pathway was deleted in this background, resulting in the double pathway mutant EGDe-ΔhmgRΔgcpE. This double pathway mutant could be created only when mevalonate was added to the medium, and the mutant relied on an exogenous supply of mevalonate for growth at all times (Fig. (Fig.4A).4A). Restoration of the mevalonate pathway by introduction of lmo0825 (hmgR) on a plasmid permitted growth in the absence of mevalonate (Fig. (Fig.4B).4B). These results formally prove that L. monocytogenes requires at least one pathway of isoprenoid biosynthesis for growth under in vitro laboratory conditions and that further means of isoprenoid biosynthesis do not exist in this organism.
The growth characteristics of the double pathway mutant were examined in more detail in broth with and without added mevalonate. In medium without added mevalonate, the double pathway mutant was capable of moderate growth and could reach an OD600 of only approximately 0.2, after which growth was arrested (Fig. (Fig.4B).4B). It is possible that intracellular stores of IPP allowed this initial growth, and in support of this hypothesis, medium inoculated with a higher dilution (1:1,000) of overnight culture did not show any measurable growth (data not shown). Although growth of the mutant was arrested, cells remained viable as determined by the ability to form colonies on agar plates. Indeed, the addition of mevalonate after 6 or 10 h after the beginning of the experiment resulted in an immediate regeneration of growth (Fig. (Fig.4B).4B). However, it was observed that both the growth rate and final optical density reached were lower than for the mutant culture to which mevalonate had been added at the start of the experiment. Also, mevalonate addition after 10 h resulted in more pronounced growth defects compared to mevalonate addition after 6 h.
However, in the presence of mevalonate, the double pathway mutant displayed virtually identical growth characteristics as the complemented strain to which no mevalonate was added (Fig. (Fig.4B).4B). Moreover, growth of these strains under the conditions tested was similar to the wild-type L. monocytogenes EGDe strain, demonstrating that either the addition of mevalonate or the restoration of the mevalonate pathway results in full complementation of the double pathway mutant.
As mice do not have the Vγ9Vδ2 subset of T cells (10), the mouse model of infection cannot be used to examine the direct immunomodulatory role of HMP-BB during infection by L. monocytogenes. However, as the nonpathogenic species L. innocua lacks a complete MEP pathway and this species cannot grow intracellularly, the mouse model can be employed to investigate whether the MEP pathway contributes to growth of L. monocytogenes in this environment. The wild-type L. monocytogenes EGDe, and previously described single mutants in both pathways (EGDe-mvk, EGDe-gcpE, and EGDe-lytB mutants) (3) were used to intraperitoneally infect groups of four mice. Three days postinfection the mice were sacrificed and the numbers of bacteria present in the spleens and livers were determined (Fig. (Fig.5A).5A). The bacterial numbers observed in livers and spleens derived from the groups of mice infected with the wild-type L. monocytogenes EGDe strain and the mevalonate pathway mutant (EGDe-mvk) were similar. In contrast, the numbers obtained for both MEP pathway mutants (EGDe-gcpE and EGDe-lytB) were significantly lower (1 to 2 logs) than the wild type in both the livers and the spleens, suggesting that the MEP pathway contributes to the fitness and survival of L. monocytogenes during the infection process. Complementation of the gcpE deletion mutant by transformation of the strain with pNZ44-gcpE restored virulence back to wild-type levels (data not shown).
The survival of the double pathway mutant during murine infection was also investigated in a similar manner. This mutant was not recovered from the livers and spleens, suggesting that in vivo mevalonate and/or IPP levels are too low to allow growth of a strain that cannot synthesize its own supply (Fig. (Fig.5B).5B). Cloning hmgR from the classical pathway into the double mutant restored virulence to wild-type levels. This was unexpected, as this strain still lacks a functional MEP pathway. Experiments by Sauret-Güeto et al. (30) in E. coli have shown that spontaneous mutations arise at a high frequency to block the otherwise-lethal obstruction of isoprenoid biosynthesis. We suggest that the restoration of virulence of our L. monocytogenes double mutant may be due to mutations or alterations in transcription of a gene(s) that arose during the creation of the mutant.
In order to examine expression of genes in vivo, a luciferase-based reporter system was employed. The promoter regions of the genes (lmo0010 [mvk], lmo0825 [hmgR], lmo1441 [gcpE], and lmo1451 [lytB]) were cloned into pPL2lux and the constructed plasmids were integrated into the chromosome of L. monocytogenes EGDe, allowing direct assessment of gene expression levels by monitoring bioluminescence levels (5). All constructs were stable, and in vitro bioluminescence imaging determined that all constructs are functional (Luria-Bertani, pH 7.4) (data not shown). Mice infected with the constructed strains were sacrificed 3 days postinfection, and bioluminescence imaging was conducted on the livers (Fig. (Fig.5C).5C). Of the four genes tested, only gcpE expression (MEP pathway) was detectable in the livers of mice. Cell numbers recovered from the livers of the mice infected with the different strains were very similar, excluding the possibility that the observed differences in expression levels are caused by differences in infection efficacies.
As L. monocytogenes virulence defects observed following murine infection via the intraperitoneal route can often be attributed to an inability to survive in macrophages, the capacity of strains to survive in murine J774 macrophages was examined as outlined in Materials and Methods. It was observed that single mutants were unaffected in uptake or in intracellular survival, as the numbers of bacteria recovered were comparable to the wild type over the short time course of this experiment (data not shown). The double mutant that was starved of mevalonate prior to infection (see Materials and Methods for details) was also recovered in similar numbers to the wild type, suggesting that levels of mevalonate and/or IPP in the macrophages were sufficiently high to permit growth.
We have previously noted that although L. monocytogenes possesses all genes for both pathways of isoprenoid biosynthesis, the closely related nonpathogenic species L. innocua lacks the last two genes of the MEP pathway (3). The present study was initiated to further explore this phenomenon and to identify a possible role for the genes in L. monocytogenes.
Initial work focused on using a bioinformatics approach to examine the gcpE and lytB genomic regions, and results obtained provide very strong evidence that these genes were lost in L. innocua during adaptation to a nonpathogenic life cycle. As this species also lacks the PrfA virulence cluster required for intracellular survival, we postulated that a complete MEP pathway may be required by L. monocytogenes during this stage of infection. Furthermore, previous studies imply that the MEP pathway may be important for the virulence potential of other pathogens. An in vivo expression technology screen revealed that dxr (DOXP reductoisomerase) was induced in vivo during Klebsiella pneumoniae infection of mice (21), the Brucella abortus dxs (DOXP synthase) gene was shown to be activated in RAW macrophages 4 h postinfection (12), and experiments by Shin et al. (31) revealed that a gcpE transposon mutant of Mycobacterium tuberculosis displayed a significant decrease in tissue colonization (liver, intestine) following intraperitoneal infection of BALB/c mice.
Experiments using the mouse model of infection support our hypothesis, as MEP pathway mutants were impaired in infection of mice via the intraperitoneal route. The precise role of the MEP pathway in bacterial survival in the host environment is unknown, but as mice do not have Vγ9Vδ2 T cells, the virulence defect observed cannot be attributed to immune modulation by HMB-PP. MEP mutants were not affected in macrophage culture-based experiments and remained unaffected in early (24-h) systemic growth in mice, indicating resistance to initial macrophage defenses in vivo (data not shown). In addition, microarray experiments performed by Chatterjee et al. (6) did not show altered expression of genes of either pathway of isoprenoid biosynthesis in L. monocytogenes cells isolated from infected macrophages compared to cells grown in broth. We suggest that the longer-term deficiencies in growth in vivo seen in our experiments may reflect an as-yet-unidentified shift in the nutritional environment of the pathogen during infection that necessitates an operational MEP pathway. Further work will be necessary to prove this hypothesis. The double pathway mutant was not recovered from the livers and spleens of mice 3 days postinfection, suggesting that the in vivo mevalonate and/or IPP levels are not sufficient to allow growth of a strain that cannot undergo isoprenoid biosynthesis.
Although L. innocua lacks a complete MEP pathway, our ability to restore the pathway by heterologous expression of gcpE and lytB demonstrates that the enzymes responsible for the synthesis of the first four intermediates of the MEP pathway are still functional. It is possible that these enzymes may feed into other pathways rather than having a role in isoprenoid biosynthesis. Taking all of our observations into account, we suggest that L. innocua has devolved to express only a single pathway for isoprenoid biosynthesis and has lost the functional MEP pathway required for in vivo growth. This may represent niche adaptation by the nonpathogenic species, which no longer involves colonization of the host. It is noteworthy that other researchers have also provided evidence that L. monocytogenes and L. innocua evolved from a common pathogenic ancestor but L. innocua has lost many of the genes required for infection (8, 33).
Furthermore, our experiments suggest that the stimulation of Vγ9Vδ2 T cells by L. monocytogenes but not L. innocua is conferred by the presence of a complete MEP pathway (gcpE and lytB) in the former species. It has been suggested that the function of Vγ9Vδ2 T cells is to initiate and regulate the mucosal immune system in order to limit a detrimental host response toward commensal bacteria (10). Indeed, a metagenomic approach by Gill et al. (13) revealed that MEP pathway genes are abundant in the indigenous microflora of the human distal gut microbiome, and genome sequence analyses have revealed that the common intestinal bacteria species of Bacteroides, Bifidobacterium, Fusobacterium, E. coli, and Clostridium all possess genes of the MEP pathway (10). It is therefore entirely possible that L. monocytogenes may deliberately exploit this surveillance system, thereby inducing a state of tolerance. Since L. innocua does not need to persist in the intestinal tract, for this bacterium the ability to interact with the hosts immune system may not be essential. As mice lack Vγ9Vδ2 T cells, rhesus macaques (Macaca mulatta) or night monkeys (Aotus nancymaae) will have to be employed to explore this hypothesis (10).
We acknowledge the funding received from the Irish Government under the National Development Plan 2000-2006 and the funding of the Alimentary Pharmabiotic Centre by the Science Foundation of Ireland Centres for Science Engineering and Technology scheme. H.J. received a grant from the Deutschen Forschungegesellschaft.
We also thank Jerry Reen and Pauline Scanlan for helpful discussions concerning the bioinformatics analyses.
Editor: V. J. DiRita
Published ahead of print on 2 September 2008.