Report

Host Proteasomal Degradation Generates Amino Acids Essential for Intracellular Bacterial Growth

See allHide authors and affiliations

Science  16 Dec 2011:
Vol. 334, Issue 6062, pp. 1553-1557
DOI: 10.1126/science.1212868

Abstract

Legionella pneumophila proliferates in environmental amoeba and human cells within the Legionella-containing vacuole (LCV). The exported AnkB F-box effector of L. pneumophila is anchored into the LCV membrane by host-mediated farnesylation. Here, we report that host proteasomal degradation of Lys48-linked polyubiquitinated proteins, assembled on the LCV by AnkB, generates amino acids required for intracellular bacterial proliferation. The severe defect of the ankB null mutant in proliferation within amoeba and human cells is rescued by supplementation of a mixture of amino acids or cysteine, serine, pyruvate, or citrate, similar to rescue by genetic complementation. Defect of the ankB mutant in intrapulmonary proliferation in mice is rescued upon injection of a mixture of amino acids or cysteine. Therefore, Legionella promotes eukaryotic proteasomal degradation to generate amino acids needed as carbon and energy sources for bacterial proliferation within evolutionarily distant hosts.

Although bacterial acquisition of nutrients in vivo is one of the most fundamental aspects and is a prerequisite for bacterial infections (1, 2), it is not known whether intracellular pathogens possess specific virulence strategies to modify cellular processes to obtain sources of carbon, nitrogen, and energy. In the amoeba host and in human alveolar macrophages, the Legionella pneumophila–containing vacuole (LCV) is remodeled into a rough endoplasmic reticulum (RER)–derived vacuole that evades lysosomal fusion (3). The Dot/Icm type IV secretion system delivers ~300 effectors into the host cell to remodel it into a proliferative niche (4, 5). Host-mediated farnesylation anchors the AnkB effector of strains AA100/130b of L. pneumophila to the cytosolic leaflet of the LCV membrane, where the effector functions as a bona fide F-box protein that assembles polyubiquitinated proteins on the LCV (611). This effector is required for intracellular proliferation within human cells and amoeba and for intrapulmonary bacterial proliferation in mice (68, 12). Despite the severe defect of intravacuolar proliferation of the ankB mutant, its LCV evades lysosomal fusion and is remodeled by the RER, similar to the LCV harboring the wild–type (WT) strain (12). How the AnkB-mediated assembly of polyubiquitinated proteins promotes bacterial proliferation within amoeba and human cells is unclear (6, 9, 12).

Polyubiquitination is a conserved eukaryotic posttranslational modification that covalently polymerizes ubiquitin monomers to a target protein through one of the seven lysine (K) residues of ubiquitin. One of the best characterized is K48-linked polyubiquitination, which is a hallmark for 26S proteasomal degradation (13), generating short peptides that are rapidly degraded to amino acids by aminopeptidases (1416).

To determine the lysine linkages in polyubiquitinated substrates assembled by AnkB, we transfected human embryonic kidney (HEK) 293 cells with hemagglutinin (HA)-tagged native ubiquitin or one of its variants with a single Lys-Arg substitution of the seven lysine residues (K6R, K11R, K27R, K29R, K33R, K48R, and K63R) or with ubiquitin variants harboring a single lysine (6K, 11K, 27K, 29K, 33K, 48K, and 63K) (17). Cells were infected with WT L. pneumophila, its isogenic ankB mutant, or the translocation-defective dotA mutant control. At 2 hours postinfection, >95% of the infected cells harbored a single bacterium. Overexpression of any of the 14 ubiquitin variants in which K48 was intact had little or no effect on decorating the WT strain–containing LCV with polyubiquitinated proteins, and intravacuolar proliferation of L. pneumophila was not affected (Fig. 1). However, when K48 was substituted, decoration of the LCV with polyubiquitinated proteins and bacterial proliferation was reduced (Student’s t test P < 0.001) up to 48 hours postinfection, similar to reduced decoration of the LCV by ubiquitin with no lysine residues (0K) (Fig. 1) and decoration after infection by the dotA or ankB mutants (table S1). Thus, AnkB-mediated preferential decoration of the LCV with K48-linked polyubiquitinated proteins is critical for intracellular proliferation.

Fig. 1

Preferential docking of K48-linked polyubiquitinated proteins to the LCV and its role in intravacuolar proliferation of L. pneumophila. Representative confocal microscopy images of untransfected or transfected HEK293 cells with HA-tagged native ubiquitin (Ub) or one of its variants with a single Lys-Arg substitution (K6R, K11R, K27R, K29R, K33R, K48R, and K63R) or six of the Lys residues of ubiquitin were substituted and a single Lys was left intact (6K, 11K, 27K, 29K, 33K, 48K, and 63K). The cells were infected with L. pneumophila (Lpn, green) at multiplicity of infection (MOI) of 20 (20 to 30% infection efficiency), and recruitment of polyubiquitinated proteins (HA-UB, red) to the LCV was examined at 2 (A and C) and 12 (B and D) hours postinfection. Arrowheads indicate marked colocalization of polyubiquitinated proteins with the LCVs. Numbers in the merged images in (A) and (C) are quantification of % of LCVs positive for recruitment of polyubiquitinated proteins. (B and D) The % cells with replicative LCVs harboring ≥five bacteria at 12 hours, whereas ~95% of LCVs harbored a single bacterium at 2 hours (A and C). For all assays, 100 infected cells were analyzed from multiple cover slips. The results are representative of three independent experiments performed in triplicate.

We tested the hypothesis that proteasomal degradation of the K48-linked polyubiquitinated proteins generated amino acids that promoted intracellular growth of L. pneumophila. HEK293 cells and human monocytes-derived macrophages (hMDMs) were pretreated with proteasomal inhibitors, MG-132 or PS-341, in the presence or absence of various doses (0.25 to 1 mM) of a mixture of amino acids supplement to the rich tissue culture media. Neither inhibitor affected viability of the host cells, but both inhibitors restricted intracellular replication of L. pneumophila, and the restriction was bypassed upon supplementation with amino acids in a dose-dependent manner (Student’s t test P < 0.001) (Fig. 2A and fig. S1). Three chemical inhibitors of aminopeptidases that degrade the proteasome-generated short peptides (1416) inhibited intracellular proliferation in a dose-dependent manner, which was also relieved upon supplementation with amino acids (Fig. 2B). Supplementation of amino acids also restored bacterial proliferation within K48R ubiquitin-overexpressing HEK293 cells (Figs. 1B and 3, A and B).

Fig. 2

The effect of inhibition of host proteasomes or aminopeptidases on intracellular bacterial proliferation in hMDMs is bypassed by amino acid supplementation. (A) Kinetics of intracellular bacterial proliferation was assessed at several time points postinfection of untreated or MG-132-treated hMDMs infected by WT L. pneumophila in presence or absence of excess 2.5× (0.25 mM) or 5× (0.5 mM) amino acids (AA) supplement of the rich RPMI media. (B) Kinetics of intracellular bacterial proliferation was assessed at several time points postinfection of untreated hMDMs compared with cells treated with each of the three different inhibitors of aminopeptidases in the presence or absence of 5× (0.5 mM) amino acid (AA) supplementation. The results are representative of three independent experiments performed in triplicate, and error bars represent standard deviation.

Fig. 3

Docking of K48-linked polyubiquitinated proteins to the LCV is essential for averting amino acid starvation by intravacuolar L. pneumophila but is bypassed by supplementation of amino acids. Representative confocal microscopy images of HEK293 cells overexpressing native HA-tagged Ub (HA-UB, red) or its K48R and 48K variants without or with 5× minimal essential media (MEM) excess amino acid (AA) supplement. (A) The cells were infected with WT L. pneumophila (Lpn, green) to assess formation of replicative vacuoles at 12 hours postinfection, and quantification is shown in (B) (see Fig. 1B legend). (C) The cells were infected with Pfla-gfp-expressing WT L. pneumophila (Lpn, blue), and expression of GFP was examined at 8 hours postinfection. Arrowheads indicate heavy colocalization of the LCVs with polyubiquitinated proteins. Quantification of Pfla-dependent GFP expression by L. pneumophila is shown in (D). All analyses were performed on 100 infected cells analyzed from multiple cover slips. The results are representative of three independent experiments performed in triplicate, and error bars represent standard deviation.

A standard reporter system for amino acid starvation in L. pneumophila is the flaA promoter fused to green fluorescent protein (PflaA-gfp) (18). Expression of PflaA-gfp was triggered at 8 hours within cells overexpressing any of the 15 ubiquitin variants in which K48 was substituted (such as K48R) but not when K48 was intact (Student’s t test P < 0.002) (fig. S2). In a dose-dependent manner, the amino acid supplement blocked triggering of PflaA-gfp in cells overexpressing K48R ubiquitin compared with cells expressing native or 48K ubiquitin (Student’s t test P < 0.001) (Fig. 3, C and D).

Similar to the WT strain, the LCV harboring the ankB mutant evades lysosomal fusion and is remodeled by the RER (12). Thus, the sole defect of the ankB null mutant appeared to be failure to generate amino acids from proteasomal degradation of the AnkB-assembled K48-linked polyubiquitinated proteins. Comparative measurements of free amino acids in cellular lysates showed that, at 8 hours postinfection, the WT strain triggered a significant rise in the levels of each of the free amino acids in an AnkB-dependent manner (Fig. 4A, fig. S3, and table S2). Amino acid supplementation specifically rescued the ankB but not the dotA mutant for the severe defect in intravacuolar proliferation in hMDMs (Fig. 4B) and amoeba (fig. S4A) (Student’s t test P < 0.001), and the rescue was similar to genetic complementation by the native ankB allele (6, 8, 12). Upon addition of excess amino acids at 6, 12, or 24 hours postinfection, intracellular proliferation of the ankB mutant was triggered (Fig. 4B).

Fig. 4

The critical role of AnkB-mediated generation of amino acids in intracellular proliferation and manifestation of pulmonary disease by L. pneumophila is bypassed by supplementation of an amino acid mixture or by cysteine. (A) The levels of cellular free amino acids were determined and are expressed as the -fold ratio of infected/uninfected HEK293 cells. The analyses were performed on three wells of cells, and the data shown are one of two representative experiments. All free amino acids were significantly elevated (Student’s t test, P < 0.001) in WT strain–infected compared to ankB mutant–infected cells. (B) Intracellular growth kinetics within of hMDMs to assess the effect of AA supplementation when added before infection or at 6, 12, and 24 hours after infection by the ankB mutant. (C and D) Rescue of the ankB mutant for intracellular proliferation in hMDMs or HEK293 cells by 5× (0.5 mM) supplement of individual amino acids (C) or by pyruvate or citrate (D). (E) Intrapulmonary bacterial proliferation in intratracheally inoculated mice injected intraperitoneally at 6-hour intervals with 0.5 mM amino acid mixture (AA) or with 0.5 mM Cys. The results are representative of three independent experiments performed in triplicate, and error bars represent standard deviation.

Nutrient limitation in L. pneumophila triggers RelA and SpoT, which are the amino acid and fatty acid starvation sensors, respectively (18), and one of their downstream inducible targets is flaA (18). Real-time quantitative analyses of transcription of intracellular bacteria at 8 hours postinfection showed that, compared with expression in the WT strain, the expression of relA, spoT, and flaA was dramatically up-regulated by the ankB mutant within amoeba (fig. S4B) and hMDMs (fig. S4C), similar to up-regulation of the three genes upon starvation of WT L. pneumophila in vitro (fig. S4D). Expression of relA was the most induced (100- to 170-fold) in the ankB mutant in both host cells, and the induction was reversed by amino acid supplementation (fig. S4, B and C). However, our data did not exclude that other nutritional and nonnutritional stresses were also exhibited by intracellular Legionella, despite rescue of the ankB mutant for its in vivo amino acid starvation response through amino acid supplementation.

Supplementation of each of the 20 individual amino acids showed that cysteine or serine specifically rescued the ankB but not the dotA mutant in both human cells (Fig. 4C) and amoeba (fig. S5, A and B), similar to the amino acids mixture. Other amino acids exhibited differences in effects between the two host cells, which likely reflects different levels of intracellular free amino acids available to Legionella and the bacterial efficiency to generate carbon and energy from these amino acids within macrophages and amoeba (fig. S6). Upon injection of A/J mice at 6-hour intervals with the mixture of amino acids or with cysteine alone, the ankB but not the dotA mutant was specifically rescued for intrapulmonary proliferation, similar to complementation by the native ankB allele (Fig. 4E).

The tricarboxylic acid (TCA) cycle is a primary route of carbon assimilation and energy production from amino acids by Legionella, and serine is a primary energy and carbon source in vitro and intracellularly, although carbohydrates have a minor role (19) (fig. S6) (19, 20). Serine and cysteine are converted by Legionella to pyruvate, followed by acetyl coenzyme A (CoA) that feeds the TCA cycle (fig. S6). Similar to supplementation of cysteine and serine, supplement of pyruvate or citrate specifically rescued the ankB but not the dotA mutant for intracellular proliferation in human cells (Fig. 4D) and amoeba (fig. S5C). Thus, rescue of the ankB mutant upon amino acid supplementation was direct and was not due to a secondary effect on host cells by the excess amino acids. Rescue of the ankB mutant for intracellular proliferation by cysteine, serine, pyruvate, or citrate indicates that the ankB mutant has sufficient levels of nutrients in the host cell with the exception of higher levels of a needed major source of carbon and energy production through the TCA cycle.

The high demand for amino acids by Legionella is demonstrated by the presence of ~12 classes of adenosine triphosphate–binding cassette transporters, amino acid permeases, and many proteases and by the requirement to supplement the rich medium for in vitro growth of Legionella with 3.3 mM cysteine (21, 22). Our data suggest that cysteine is a major source of carbon and energy for Legionella within amoeba and human cells. To satisfy its high demands for amino acids within amoeba and human cells, Legionella uses the AnkB bona fide F-box effector to exploit the conserved eukaryotic processes of K48-linked polyubiquitination and the proteasome machineries to generate amino acids. The generated amino acids are imported to the LCV through SLC1A5 and other host amino acid transporters (20, 22). However, generation of amino acids through promoting proteasomal degradation may not be the only virulence strategy dedicated to obtain carbon and energy sources from the host, because the type II–secreted degradation enzymes (21) may provide an additional strategy to generate sources of carbon and energy. Our work shows a microbial strategy that exploits conserved eukaryotic proteasomal degradation of K48-linked polyubiquitinated proteins to generate amino acids for use as sources of carbon and energy in vivo. This microbial strategy is required for intracellular bacterial growth in evolutionarily distant host cells and for manifestation of disease in mammals.

Supporting Online Material

www.sciencemag.org/cgi/content/full/science.1212868/DC1

Materials and Methods

Figs. S1 to S6

Tables S1 and S2

References (2338)

References and Notes

  1. Acknowledgments: We thank C. Sasakawa for the generous gift of providing all the ubiquitin mutant plasmid DNA and J. Cox at the University of Utah Metabolomics Core Facility for the free amino acids analyses. Y.A.K. is supported by Public Health Service awards R01AI43965 and R01AI069321 from National Institute of Allergy and Infectious Diseases and by the commonwealth of Kentucky Research Challenge Trust Fund.
View Abstract

Navigate This Article