An acetyltransferase effector conserved across Legionella species targets the eukaryotic eIF3 complex to modulate protein translation

Abstract

The survival of Legionella spp. as intracellular pathogens relies on the combined action of protein effectors delivered inside their eukaryotic hosts by the Dot/Icm (defective in organelle trafficking/intracellular multiplication) type IVb secretion system. The specific repertoire of effector arsenals varies dramatically across over 60 known species of this genera with Legionella pneumophila responsible for most cases of Legionnaires' disease in humans encoding over 360 Dot/Icm effectors. However, a small subset of "core" effectors appears to be conserved across all Legionella species raising an intriguing question of their role in these bacteria's pathogenic strategy, which for most of these effectors remains unknown. L. pneumophila Lpg0103 effector, also known as VipF, represents one of the core effector families that features a tandem of Gcn5-related N-acetyltransferase (GNAT) domains. Here, we present the crystal structure of the Lha0223, the VipF representative from Legionella hackeliae in complex with acetyl-coenzyme A determined to 1.75 Å resolution. Our structural analysis suggested that this effector family shares a common fold with the two GNAT domains forming a deep groove occupied by residues conserved across VipF homologs. Further analysis suggested that only the C-terminal GNAT domain of VipF effectors retains the active site composition compatible with catalysis, whereas the N-terminal GNAT domain binds the ligand in a non-catalytical mode. We confirmed this by in vitro enzymatic assays which revealed VipF activity not only against generic small molecule substrates, such as chloramphenicol, but also against poly-L-lysine and histone-derived peptides. We identified the human eukaryotic translation initiation factor 3 (eIF3) complex co-precipitating with Lpg0103 and demonstrated the direct interaction between the several representatives of the VipF family, including Lpg0103 and Lha0223 with the K subunit of eIF3. According to our data, these interactions involve primarily the C-terminal tail of eIF3-K containing two lysine residues that are acetylated by VipF. VipF catalytic activity results in the suppression of eukaryotic protein translation in vitro, revealing the potential function of VipF "core" effectors in Legionella's pathogenic strategy.IMPORTANCEBy translocating effectors inside the eukaryotic host cell, bacteria can modulate host cellular processes in their favor. Legionella species, which includes the pneumonia-causing Legionella pneumophila, encode a widely diverse set of effectors with only a small subset that is conserved across this genus. Here, we demonstrate that one of these conserved effector families, represented by L. pneumophila VipF (Lpg0103), is a tandem Gcn5-related N-acetyltransferase interacting with the K subunit of human eukaryotic initiation factor 3 complex. VipF catalyzes the acetylation of lysine residues on the C-terminal tail of the K subunit, resulting in the suppression of eukaryotic translation initiation factor 3-mediated protein translation in vitro. These new data provide the first insight into the molecular function of this pathogenic factor family common across Legionellae.

Keywords: Legionella pneumophila; acetyltransferase; effectors; pathogenesis.

Fig 1

VipF is a tandem GNAT acetyltransferase. (A) Domain organization is shown on the top and crystal structure of Lha0223 (PDB 6WQB) where the N-terminal domain is colored cyan while the C-terminal domain is colored orange and both splayed β-strands are colored magenta. (B) General acetyltransferase substrate screen using the 5´5´-dithiobis(2-nitrobenzoic acid) (DTNB)-based assay. The experiments were performed in three independent experiments with similar results. Statistical analysis for the DTNB assays was performed with a two-tailed t-test (***, P < 0.004; ****, P < 0.0001). (C and D) A comparison of the residues (cyan) involved in coordinating acetyl coenzyme A (AcCoA) (magenta) between the C-terminal and N-terminal domains. (E) Coomassie stain of purified wild-type (WT) enzymes and their mutants used for 1B and 1F. (F) Alanine substitutions of Asp251 and Tyr263 in both Lpg0103 and Lha0223 were tested for activity against the histone H3 peptide using the DTNB-based detection assay.

Fig 2

The human putative protein targets of VipF. (A) The putative human protein targets of VipF were identified using Affinity purification followed by mass spectrometry (AP-MS) with U937 cell lysate. The heat map representation of the AP-MS results is displayed as the average total spectrum peptide counts in at least nine MS runs. (B) A yeast two-hybrid (Y2H) experiment was performed for all subunits of the eIF3 complex and DHX57 in order to determine direct interactions between VipF and candidate human protein targets, confirming that VipF interacts with eIF3-K. (C) A maximum-likelihood phylogenetic tree of 37 VipF orthologs was generated using the amino acid sequence. The phylogenetic tree was constructed with IQ-TREE and visualized using iTOL. (D) Y2H experiment of VipF orthologs with low to high sequence similarity to Lpg0103, demonstrating that orthologs of VipF can interact with eIF3-K. The AP-MS and Y2H experiments were performed in three independent experiments with similar results.

Fig 3

Mutational analysis of highly conserved residues that mediate interaction with human eIF3-K. (A) Mapping of highly conserved residues (orange) located in the interdomain cleft of the Lha0223 crystal structure. (B) Y2H experiment of the alanine substitution of highly conserved residues residing in the interdomain cleft of VipF. Double distilled water (DDW) was spotted on plates to ensure growth was not due to contamination. The Y2H experiment was performed in three independent experiments with similar results.

Fig 4

VipF interacts with the C-terminal tail of eIF3-K and is capable of acetylating eIF3-K in vitro. (A) AlphaFold2 model of the VipF (light gray)-eIF3-K (gold) complex. VipF residues essential for interaction and catalysis are highlighted in red. (B) Y2H analysis of pairwise interaction between VipF WT and either eIF3-K WT or eIF3-K (1-184). (C) 5´5´-Dithiobis(2-nitrobenzoic acid-based detection of the acetylation of eIF3-K WT with VipF WT, VipF Y263A, or VipF I217S. (D) Determination of the acetylation of two lysine residues present in the C-terminal tail of eIF3-K WT, eIF3-K (1-184), eIF3-K K197A, eIF3-K K199A, or eIF3-K K197A + K199A. (E) Western blot analysis of the acetylation of eIF3-K WT in the presence of VipF WT. For the samples with full-length eIF3-K, eIF3-K K197A, and eIF3-K K199A, we also observed an acetylated protein band corresponding to Lpg0103, which may be a result of self-catalysis promoted by the presence of eIF3-K. The amount of protein used for immunoblotting analysis is indicated by the Coomassie blue stain gel (bottom). The enzymatic assays and western blot experiments were performed in three independent experiments, rendering similar results.

Fig 5

VipF dampens protein translation in a eukaryotic cell lysate. (A) Translation of luciferase mRNA in RRLs incubates with either RalF, Lgt1, VipF WT, VipF I217S, or VipF Y263A. Lysate was incubated with either 0.1 µM (circle symbol) or 1 µM (triangle symbol) of purified effector protein. (B) Translation of luciferase mRNA in RRLs incubated with either Lgt1, VipF WT, VipF I217S, or VipF Y263A. Lysates were incubated with 1 µM of purified effector protein. RRLs were also incubated in the absence (circle symbol) or presence of acetyl coenzyme A (AcCoA) (triangle symbol = 25 µM, square symbol = 50 µM, and diamond symbol = 75 µM). The data are the representation of three independent experiments that had similar results. The asterisks represent the statistical significance generated using the t-test (***, P < 0.001; ****, P < 0.0001; n.s., not significant).