Report

Listeria Intracellular Growth and Virulence Require Host-Derived Lipoic Acid

See allHide authors and affiliations

Science  17 Oct 2003:
Vol. 302, Issue 5644, pp. 462-464
DOI: 10.1126/science.1088170

Abstract

Listeria monocytogenes is a Gram-positive intracytosolic pathogen that causes severe disease in pregnant and immunocompromised individuals. We found that L. monocytogenes lacking the lipoate protein ligase LplA1 was defective for growth specifically in the host cytosol and was less virulent in animals by a factor of 300. A major target for LplA1, the E2 subunit of pyruvate dehydrogenase (PDH), lacked a critical lipoyl modification when the ΔlplA1 strain was grown intracellularly, which suggests that abortive growth was due to loss of PDH function. Thus, the use of host-derived lipoic acid may be a critical process for in vivo replication of bacterial pathogens.

Host cells provide intracellular pathogens protection from humoral immunity and nutrients essential for growth (1, 2). Listeria monocytogenes is a facultative Gram-positive bacterium that replicates efficiently within the host cytosol, but relatively little is known about genes required for its intracellular survival and replication (3). A hexose phosphate transporter, hpt, allows pathogenic Listeria spp. to use hexose phosphates in the host cytosol as a carbon and energy source (4). Hpt mutant bacteria grow less efficiently inside cells and are 10% as virulent as wild type in mice. To identify additional genes important for intracellular growth, we performed a modified intracellular methicillin selection on a transposon insertion library of L. monocytogenes (table S1) (5). We identified three types of mutants from the methicillin selection. The first class of mutants was nonhemolytic on agar containing blood; these mutants remained in the vacuole where they were unable to replicate. The second class of mutants consisted of amino acid auxotrophs, primarily proline and threonine auxotrophs. A third class of mutants was hemolytic and grew in minimal medium. These mutants probably contained insertions in genes important for intracellular growth; one of the mutants, DP-L2214, was selected for further analysis.

DP-L2214 exhibited normal growth in both rich (Fig. 1A) and minimal bacteriological media (6). In contrast, replication of DP-L2214 in macrophages aborted at about 5 hours post infection (hpi) (Fig. 1B; fig. S1). Thus, DP-L2214 displayed a replication defect that was restricted to the intracellular environment. By sequencing the DNA adjacent to the transposon, we identified an open reading frame (ORF) disrupted by the Tn917 insertion. A BLAST search of the L. monocytogenes genome with this sequence revealed homology to the lipoate protein ligase gene (lplA) of Escherichia coli, so we termed this gene lplA1 (EGD-e lmo0931) (7). The L. monocytogenes genome sequence also revealed a second lplA-like gene, lplA2 (EGD-e lmo0764) (7). To verify that the intracellular replication defect of DP-L2214 was due to interruption of the lplA1 ORF, we constructed an in-frame deletion of the lplA1 gene. The ΔlplA1 strain was characterized in an intracellular replication assay in J774 cells (Fig. 1B). Growth of ΔlplA1 in J774 cells appeared similar to that of DP-L2214 in doubling time and kinetics. ΔlplA1 was also compared with DP-L2214 in an L2 fibroblast plaquing assay that measured intracellular growth and cell-to-cell spread over 3 days (Fig. 1C). Plaque size of both ΔlplA1 and DP-L2214 was on average 56 and 58% of the wild-type plaque size, respectively. Thus, disruption of the lplA1 ORF by Tn917 in DP-L2214 resulted in loss of function and was responsible for the abortive growth phenotype.

Fig. 1.

Growth of wild-type and LplA1 mutant strains. (A) Growth of L. monocytogenes in brain-heart infusion (BHI) broth. (B) Intracellular replication of wild type (filled squares), DP-L2214 (filled circles), or ΔlplA1 (open circles) bacteria in J774 macrophages. (C) Intracellular growth and spread of L. monocytogenes in murine L2 fibroblasts. After 72 hours, a 3-ml 1× DMEM (Dulbecco's Modified Eagle's Medium) agar overlay containing Neutral Red was added to allow visualization of necrotic foci.

The intracellular replication defect of the LplA1-deficient strains was not apparent until about 5 hpi; we hypothesized that bacteria grown in the rich laboratory medium used for the inoculum contained pools of lipoylated proteins, targets of LplA1 or LplA2, which could permit a limited number of divisions inside the macrophage before cessation of bacterial growth. To test this possibility, bacteria isolated from J774 macrophages 4 hpi were used to reinfect a fresh macrophage monolayer (Fig. 2A). The wild-type L. monocytogenes strain grew to the same extent as wild-type bacteria that were not subjected to a prior macrophage infection. In contrast, the lplA1 mutant strain was unable to replicate after reinfection. In the reinfection assay, normal bacterial growth could be restored to DP-L2214 by expression of a wild-type lplA1 gene on a plasmid using its endogenous promoter (Fig. 2B). Thus, LplA1 was sufficient to complement the transposon insertion mutant, DP-L2214, and was required for intracellular growth.

Fig. 2.

Intracellular growth of L. monocytogenes LplA1 mutants after isolation from macrophages. (A) J774 macrophages were infected by wild-type (filled squares) or ΔlplA1 (filled circles) L. monocytogenes. About 99% of macrophages were infected. At 4 hpi the monolayer was lysed with 0.1% NP-40, host nuclei were removed, and bacteria were harvested by microcentrifugation. These bacteria were used to infect a new monolayer of macrophages to assess intracellular growth. (B) J774 macrophages were infected as described in (A) by wild type + pAM401 (filled squares), DP-L2214 + pAM401 (filled circles), or DP-L2214 + pAM401lplA1 (open circles).

Lipoic acid (LA) is a disulfide-containing cofactor required for the function of the pyruvate dehydrogenase (PDH) enzyme complex and other structurally related enzyme complexes (8). LplA from E. coli ligates exogenous LA to the E2 subunit of PDH to produce E2-lipoamide, which plays a pivotal role in the aerobic metabolism of most organisms (9). B. subtilis PDH mutants are unable to grow in glucose minimal medium, and humans with point mutations that affect PDH function suffer from a severe metabolic disorder (10, 11). To identify target proteins for LplA1 in L. monocytogenes, we analyzed the profile of lipoylated proteins from bacteria grown in broth culture by immunoblot using an LA-specific antibody (Fig. 3A). The antibody against LA reproducibly recognized one dominant protein that was identified as the E2 subunit of PDH by mass spectrometry. No difference in lipoylation of E2 PDH was observed between the wild type and the ΔlplA1strain, consistent with the lack of phenotype of LplA1-deficient strains in broth culture. Thus, E2 PDH represents a target of LA modification in L. monocytogenes.

Fig. 3.

Analysis of E2 PDH modification in wild-type and ΔlplA1 strains. (A) Total protein preparations were made from bacteria grown in BHI overnight (22). The protein samples were analyzed by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and by immunoblot using a polyclonal LA-specific antibody. (B) J774 macrophages were infected with wild type or the ΔlplA1mutant. At 4 hpi, the monolayer was lysed in 1× PBS + 0.1% NP-40, and host cell nuclei were removed. Bacteria from the remaining supernatant, along with host cell debris, were harvested by microcentrifugation. Bacterial cell pellets were lysed, and total bacterial protein was obtained as described in (A). Protein samples were analyzed by immunoblot using an LA-specific antibody or an antibody to an unrelated bacterial protein, ActA.

We next determined the lipoylation state of E2 PDH in L. monocytogenes during intracellular growth. J774 macrophages were infected with wild-type L. monocytogenes or the ΔlplA1mutant strain at a high multiplicity of infection such that most cells contained one or more bacteria. At 4 hpi total cell lysates were prepared from intracellular bacteria for immunoblot analysis (Fig. 3B). Although lipoylated E2 was observed in wild-type bacteria grown in macrophages, modified E2 PDH was not present in ΔlplA1 lysates. The pool of modified E2 in the bacterial inoculum after overnight culture in rich medium probably allowed the ΔlplA1 strain to undergo about four rounds of division in the host cell over 5 hours before functional E2 was depleted. When lipoylated E2 was depleted after several rounds of replication in the cytosolic environment, ΔlplA1 mutant bacteria were not able to establish a productive infection (Fig. 2A). Thus, despite the presence of a second lipoate protein ligase in the genome, LplA1 performs a critical and nonredundant function during intracellular growth that involves modification of E2 PDH.

We next tested the virulence of the ΔlplA1mutant in an intravenous mouse model of infection by determining the median lethal dose (LD50). In the BALB/c background, the lplA1mutant strains had LD50 values that were 1/250th to 1/300th as virulent as the wild-type L. monocytogenes strain: 5 to 6 × 106 and 2 × 104 colony-forming units, respectively (table S1). Although the lplA1 mutants were less virulent, a heterologous antigen secreted by a ΔlplA1 mutant strain stimulated a robust CD8+ T cell response in vaccinated mice, which suggests that this strain is a promising candidate for vaccine development (12). Our in vivo experiments highlight the important contribution of LplA1 and host-derived LA to infection by L. monocytogenes.

Like L. monocytogenes, E. coli also has two lipoate protein ligase genes, lplA and lipB (13, 14). E. coli LplA ligates scavenged LA to E2 PDH, whereas LipB uses de novo synthesized LA. However, L. monocytogenes has no lipB homolog and requires an external source of LA for growth in minimal medium (15). Studies of LA metabolism have shown that little free LA exists in the mammalian cytosol (16). Thus, LplA1 may not be important in replication of L. monocytogenes under conditions where free LA is available, but it is required in the host cell where lipoyl groups may be scavenged from host molecules such as peptides. Some other bacterial pathogens, like Chlamydia trachomatis, Staphylococcus aureus, and Streptococcus pyogenes, also have two lplA-like genes but no lipB homolog, which suggests that utilization of host-derived LA may play a critical and underappreciated role in the replication of many bacterial pathogens.

The metabolic function of lipoylated PDH is important for aerobic growth, consistent with the abortive growth phenotype we have observed in the ΔlplA1 mutant. The ability of microorganisms to adapt their metabolism to a restrictive environment has undoubtedly contributed to their success as pathogens, as demonstrated in Candida albicans and Mycobacterium tuberculosis, both of which require an enzyme of the glyoxylate cycle for virulence (17, 18). However, recent studies have revealed novel functions for E2 PDH that appear independent of the metabolic function of PDH holoenzyme, including resistance to oxidative stress and transcriptional regulation (10, 1921). The possibility that E2-lipoamide, and thus LplA1, may regulate additional bacterial processes during an L. monocytogenes infection remains to be explored.

Supporting Online Material

www.sciencemag.org/cgi/content/full/302/5644/462/DC1

Materials and Methods

Fig. S1

Table S1

References and Notes

References & Notes

View Abstract

Navigate This Article