Research Article

A Role for SIRT2-Dependent Histone H3K18 Deacetylation in Bacterial Infection

See allHide authors and affiliations

Science  02 Aug 2013:
Vol. 341, Issue 6145, 1238858
DOI: 10.1126/science.1238858

Structured Abstract

Introduction

Posttranslational modification of histones is a well-documented mechanism by which the chromatin structure is modulated to regulate gene expression. Increasing evidence is revealing the strong impact of bacterial pathogens on host chromatin. However, our knowledge of the mechanisms underlying pathogen-induced chromatin changes and the impact of histone modifications and chromatin modifiers on infection is still in its infancy.

Embedded Image

Mechanism and consequence of SIRT2 activation by L. monocytogenes. Listeria induces SIRT2 relocalization from cytoplasm to chromatin, where SIRT2 deacetylates H3K18. The consequences of this cascade are control of host transcription, as illustrated by representative genes regulated by SIRT2, and control of infection, as assessed by staining cells for the secreted bacterial factor InlC (red), which is overexpressed in the cytosol, and host actin, which is polymerized into comet tails by bacteria (green). Error bars indicate SEM; **P < 0.001. Ac, acetyl; deAc, deacetylase.

Methods

We used the model bacterium Listeria monocytogenes and analyzed the mechanisms underlying a specific histone modification, deacetylation of histone H3 on lysine 18 (H3K18). Through immunoblotting, mass spectrometry, and chromatin immunoprecipitation, we studied how infection affected this modification, both in vitro and in vivo. We used a combination of chemical inhibitors, small interfering RNA (siRNA), and knockout mice to discover the key role of the host histone deacetylase sirtuin 2 (SIRT2) and determine its effect on infection. We performed microarray analysis to identify how infection and SIRT2 modulated host transcription.

Results

L. monocytogenes induces deacetylation of H3K18. This modification is mediated by the host deacetylase SIRT2. Upon infection, SIRT2 translocates from the cytosol to the chromatin of the host at the transcription start sites of a subset of genes that are repressed. We find that this process is dependent on activation, by the bacterial protein InlB, of the cell surface receptor Met and downstream phosphatidylinositol 3-kinase (PI3K)/AKT signaling. Finally, infecting cells in which SIRT2 activity was blocked (by pharmacological agents, treatment with siRNA, or the use of SIRT2–/– mice) resulted in a significant impairment of bacterial infection, showing that activity of SIRT2 is necessary for infection, both in vitro and in vivo.

Discussion

Our study identifies a stimulus, infection by L. monocytogenes, that leads to nuclear localization of SIRT2, a deacetylase previously shown to be mainly cytoplasmic. In fact, only upon infection and SIRT2 translocation from the cytoplasm to the chromatin does this deacetylase have a role in transcriptional repression. This mechanism of host subversion could be common to other invasive pathogens that induce deacetylation of histones, and it defines a target for potential therapeutic treatment.

Bacterial Subversion Tactics

Intracellular bacterial pathogens such as Listeria monocytogenes can change host cell transcription programs to promote infection. Eskandarian et al. (1238858) found that during infection, the Listeria effector protein InlB promoted the movement of a host protein deacetylase, SIRT2, from its normal location in the cytosol to the nucleus. In the nucleus, SIRT2 helped to repress a number of host cell genes by deacetylating one of their associated histones. In mice, reduced levels of SIRT2 impaired bacterial infection.

Abstract

Pathogens dramatically affect host cell transcription programs for their own profit during infection, but in most cases, the underlying mechanisms remain elusive. We found that during infection with the bacterium Listeria monocytogenes, the host deacetylase sirtuin 2 (SIRT2) translocates to the nucleus, in a manner dependent on the bacterial factor InlB. SIRT2 associates with the transcription start site of a subset of genes repressed during infection and deacetylates histone H3 on lysine 18 (H3K18). Infecting cells in which SIRT2 activity was blocked or using SIRT2−/− mice resulted in a significant impairment of bacterial infection. Thus, SIRT2-mediated H3K18 deacetylation plays a critical role during infection, which reveals an epigenetic mechanism imposed by a pathogenic bacterium to reprogram its host.

Chromatin is a dynamic and highly regulated structure crucial for compacting DNA into the nucleus without compromising vital functions such as gene expression, cell differentiation, or cell division. The basic unit of chromatin is the nucleosome, which is composed of 147 base pairs of DNA wrapped around an octamer of histone proteins H2A, H2B, H3, and H4 (1). One well-documented mechanism by which the chromatin structure is modulated is posttranslational modification of histones (2). Each histone can be modified on different residues by phosphorylation, acetylation, methylation, etc. The cellular outcome of such modifications depends on the residue targeted and its surrounding context. Increasing evidence shows the impact of histone modifications on host immunity and bacterial pathogenesis (3, 4).

Acetylation and deacetylation are important histone modifications that regulate many cellular processes, including nucleosome assembly, chromatin compaction, and gene expression. Acetylation is mediated by histone acetyltranferases (HATs), which use acetyl–coenzyme A as a cofactor to catalyze the transfer of an acetyl group to a lysine residue, thereby reducing its charge (5). This modification lowers the affinity between histones and DNA, generally allowing chromatin to adopt a more relaxed structure and the transcription machinery to be recruited. Deacetylation, which is mediated by histone deacetylases (HDACs), counteracts the effects of acetylation and, in most cases, is associated with transcriptional repression. Despite their names, both HATs and HDACs have also been shown to target nonhistone proteins (6).

In humans, 18 HDACs are grouped into four classes. The 10 members of classes I (HDAC1, -2, -3, and -8) and II (HDAC4, -5, -6, -7, -9, and -10) share considerable similarity to each other in their catalytic core that uses zinc as a cofactor, but differ in size and structural organization (7). Class I is homologous to the yeast deacetylase RPD3 and class II with yeast HDA1 (8). HDAC11 is sometimes categorized as a class I member based on homology to RPD3, but phylogenetic analysis indicates that this deacetylase belongs to a separate class, referred to as class IV (9). Class III HDACs have no sequence similarity to classes I and II, are homologous to the yeast transcriptional repressor Sir2, and use nicotinamide adenine dinucleotide as a cofactor (10, 11).

The seven class III members, also called sirtuins, have been found in a wide variety of subcellular locations and are involved in different cellular processes (12). Human sirtuin 1 (SIRT1) localizes to the nucleus and has been shown to be involved in transcriptional repression (13). In contrast, the activities of SIRT2 and SIRT3 have been mainly characterized in extranuclear compartments, the cytosol and the mitochondria, respectively (10, 14). The role of HDACs and, more specifically, sirtuins in bacterial infections has not been investigated to date.

Listeria monocytogenes is a foodborne pathogen, which mainly causes disease in immunocompromized patients and pregnant women. Major advances in the understanding of infection have come from the characterization of the mechanisms by which this facultative intracellular pathogen invades host cells, evades killing, and exploits cellular functions through the activity of its numerous virulence factors (15). The two proteins InlA and InlB promote entry into nonphagocytic cells. Once internalized, L. monocytogenes escapes the host vacuole by secretion of the pore forming toxin listeriolysin O (LLO) and phospholipases PlcA and PlcB (16). In the cytosol, L. monocytogenes grows and divides by exploiting host nutrients, including glucose-6-phosphate that is uptaken by the Hpt permease (17). Listeria perpetuates its intracellular lifestyle by polymerizing host actin using ActA and spreading to neighboring cells (15). To escape recognition, InlC dampens the host immune pathway NF-κB, and InlK and ActA counteract autophagy (1820).

Increasing evidence is uncovering the strong influence of bacterial pathogens on host chromatin (2123). However, our knowledge of the impact of histone modifications and chromatin modifiers on infection is in its infancy. In this study, we reveal a central role of SIRT2 and deacetylation of H3 in the transcriptional reprogramming imposed by the bacterial pathogen L. monocytogenes. We show that SIRT2 can be targeted to the chromatin of host cells, where it deacetylates histone H3 on lysine 18 (H3K18) at a subset of genes. SIRT2 activity is crucial during a listerial infection, defining H3K18 deacetylation as a key mechanism that allows host subversion.

Results

L. monocytogenes Induces Deacetylation of Lysine 18 at the Histone H3 N-Terminal Tail

Our previous study of histone modifications induced during L. monocytogenes infection of host epithelial cells showed a deacetylation on the N-terminal tail of histone H3 (3). To characterize this infection-induced histone modification, our first aim was the identification of the lysine residue(s) that were deacetylated. Antibodies raised against specific H3 acetyl-lysine residues were used to assess the level of acetylation at lysines 9, 14, and 18 in cells infected with L. monocytogenes. Deacetylation was specifically observed to occur on H3K18 by 3 hours of infection of HeLa cells and continued to decrease through 24 hours of infection (Fig. 1A). Within the first 5 hours of infection, no deacetylation was observed at lysine 9 (K9), K14 of histone H3, or K16 of histone H4, suggesting that under the conditions tested, only K18 is modified during infection (Fig. 1A).

Fig. 1 Infection induces deacetylation of H3K18.

(A) Acetylation levels in uninfected HeLa (-) and L. monocytogenes–infected cells (+), as detected by immunoblotting. (B) Quantification of acetylated H3K18 immunoblots in HeLa cells (n ≥ 3 experiments). Error bars represent SEM. Statistical significance was calculated using a Student’s t test. *P < 0.05; **P < 0.001. (C) The peptides detected by MS are indicated above the graphs (62). YQKprSTELLIR peptide, which is near the C terminus of histone H3 (pr, propionyl group), is used as control. The lysine residue at position 56 was exclusively identified in its light (L) (m/z = 653.8721, EGD) or heavy (H) (m/z = 655.3769, uninfected) propionylated form, indicating that this residue was not modified. The observed ion ratio is roughly 1/1, demonstrating equal mixing of both samples. The second peptide, KacQLATKprAAR (ac, acetyl group) carries acetylated K18 at its N terminus. Doublet peaks at m/z 542.8272 and 544.3319 correspond to forms of the peptide with light or heavy propionylated K23, derived from the infected and uninfected sample, respectively. The lower presence of this peptide in the infected sample indicated deacetylation of K18 with 25.4%. The ratio of light peptides (corresponding to the infected sample) to the heavy peptides (the uninfected sample) is shown under the graph (n = 2).

We had previously characterized two other histone modifications induced by L. monocytogenes—dephosphorylation of serine 10 on histone H3 and deacetylation of histone H4—showing that they were induced by one virulence factor, the toxin LLO (3). However, as shown in fig. S2, a mutant Listeria lacking the gene encoding LLO, Δhly, induces the same deacetylation of H3K18 as a wild-type (WT) strain. Therefore, H3K18 deacetylation is independent of the histone modifications previously observed in Listeria and occurs through a different mechanism that we characterize below.

Deacetylation of H3K18 was then assessed in vivo. The spleens of Balb/c mice were collected after intravenous infection with L. monocytogenes and compared with those of uninfected mice. Similar to what is observed in in vitro infection, deacetylation of H3K18, but not H3K9 or H3K14, occurred in the spleens harvested after 72 and 96 hours of infection (Fig. 2, A and B).

Fig. 2 Infection induces H3K18 deacetylation in vivo.

(A) Balb/c mice were sacrificed at the indicated times after the start of infection. The spleens were collected, and the level of acetylated H3K18 was analyzed by immunoblot. (B) Quantification of H3K9 and H3K14 acetylation levels in spleens of mice uninfected or infected with L. monocytogenes (Lm) for 72 hours. n ≥ 4 mice per time point. Error bars indicate SEM. Statistical significance was calculated using a Student’s t test. *P < 0.05; **P < 0.001.

SIRT2 Is Necessary for H3K18 Deacetylation During Infection

To identify the host factor involved in infection-induced H3K18 deacetylation, we blocked the activity of HDAC classes with specific chemical inhibitors and tested their effect on H3K18 acetyl levels in infected cells. The activity of HDAC classes I and II was blocked with trichostatin A (TSA), and that of class III, the sirtuins, was blocked with nicotinamide (NIC). TSA treatment did not inhibit infection-induced deacetylation, whereas NIC treatment completely blocked H3K18 deacetylation, suggesting a role for sirtuins in deacetylation (Fig. 3A). Next, we used specific inhibitors of the sirtuin family. Whereas a SIRT1 inhibitor, 6-chloro-2,3,4,9-tetrahydro-1H-carbasole-1-carboxamide (CTCC), had no effect, a SIRT2 inhibitor, 2-cyano-3-[5-(2,5-dichlorophenyl)-2-furanyl]-N-5-quinolinyl-2-propenamide (AGK2), blocked infection-induced deacetylation of H3K18 (Fig. 3A). With the use of small interfering RNA (siRNA), we also knocked down SIRT1, SIRT2, SIRT6, or SIRT7, which have reported deacetylase activity and are localized either in the cytoplasm or nucleus. In agreement with results obtained with chemical inhibitors, only the SIRT2 siRNA blocked H3K18 deacetylation upon infection, suggesting that SIRT2 is the HDAC responsible for this modification (Fig. 3B).

Fig. 3 The catalytic activity of SIRT2 is necessary for H3K18 deacetylation.

HeLa cells were either uninfected (-) or infected (+) with L. monocytogenes for 5 hours. (A) Cells were pretreated with HDAC inhibitors before infection. TSA, trichostatin A (class I and II inhibitors); NIC, nicotinamide (class III inhibitor); CTCC, SIRT1 inhibitor; AGK2, SIRT2 inhibitor. (B) Cells were knocked down for the expression of sirtuins by siRNA. (C) Cells were transfected with either scramble or SIRT2 siRNA and with a plasmid expressing WT SIRT2, a siRNA-insensitive SIRT2 (WT), or a catalytically inactive siRNA-insensitive SIRT2 (N168A). H3K18 acetylation levels were quantified by immunoblot (n ≥ 3). Error bars denote SEM. Statistical significance was calculated using a Student’s t test. *P < 0.05; **P < 0.001.

To definitively determine that the catalytic activity of SIRT2 was necessary for H3K18 deacetylation during infection, we complemented cells knocked down for SIRT2 by siRNA by using two plasmids expressing the wild type or a catalytically inactivated SIRT2 [Asn168→Ala168 (N168A)] (24) insensitive to siRNA. Transfecting SIRT2 knockdown cells with a siRNA-insensitive SIRT2 resulted in H3 deacetylation upon infection (Fig. 3C). This effect was observed only when transfecting with the siRNA-insensitive WT SIRT2. Deacetylation did not occur when complementing cells with the siRNA-insensitive N168A SIRT2. Furthermore, deacetylation occurred normally upon infection in control experiments using scramble siRNA, showing that N168A SIRT2 is not dominant negative (Fig. 3C). Thus, the deacetylase activity of SIRT2 is important for H3K18 deacetylation during infection.

SIRT2 Is Targeted to the Nucleus and Is Associated to Chromatin upon Infection

To determine the site of action of SIRT2, we examined its localization during infection. The activity of SIRT2 has mainly been characterized in the cytoplasm of resting cells (25). We first assessed the subcellular localization of endogenous SIRT2 by immunofluorescence. Although SIRT2 in uninfected cells was detected in both the cytosol and the nucleus, cells infected with L. monocytogenes showed a clear nuclear labeling, suggesting that infection induces relocalization of this protein (Fig. 4A).

Fig. 4 SIRT2 translocates to the nuclear and chromatin fractions upon infection.

(A) Endogenous SIRT2 was detected by immunofluorescence in HeLa cells uninfected (-) or infected (+) for 5 hours. Scale bars, 10 μm. DAPI, 4′,6-diamidino-2-phenylindole. (B) Uninfected cells (-) or 5-hour infected cells (+) were fractionated and immunoblotted for the indicated proteins. Experiments represent n ≥ 3. (C) Immunoprecipitation of SIRT2-FLAG from cytosolic, nuclear, and chromatin fractions was analyzed by MS. The numbers of SIRT2 fragments detected by MS/MS are indicated in the table; numbers of the corresponding SIRT2 peptides are in parentheses. A graphical representation of the MS/MS spectra is also shown.

The subcellular localization of SIRT2 was further characterized by cell fractionation. Cells were fractionated into cytosolic, nuclear soluble, and chromatin fractions. Control proteins known to localize to these cellular compartments were used to confirm the purity of each cell fraction. Detection of SIRT2 in these cellular fractions showed that SIRT2 (the two major splice variants) was present in the cytosol and in the soluble nuclear fraction in both uninfected and infected cells. Notably, in infected cells, SIRT2 was mostly detected in the chromatin fraction, where H3K18 deacetylation was observed (Fig. 4B). It is important to note that the activity of SIRT2 appears to be restricted to the chromatin fraction, as the cytoplasmic target tubulin was not deacetylated (Fig. 1A).

In parallel to immunoblotting cell fractions for SIRT2, we immunoprecipitated SIRT2 from each of the cytosolic, nuclear soluble, and chromatin fractions and performed mass spectrometry (MS) on the recovered material. The results confirmed that there was 10- to 20-fold more SIRT2 in the chromatin and nuclei of infected cells compared with that of uninfected cells (Fig. 4C).

We further assessed whether retaining SIRT2 in the nucleus with leptomycin B (without infection) was sufficient to deacetylate H3K18. Leptomycin B caused SIRT2 accumulation in the nucleus but did not lead to H3K18 deacetylation (fig. S3, A and B). In addition, a catalytically inactive SIRT2, which relocalized to the nucleus upon infection, did not induce H3K18 deacetylation (Fig. 3C and fig. S3C). Thus, infection induces activation and targeting of SIRT2 to the chromatin fraction where deacetylation of H3K18 occurs.

H3K18 Deacetylation and SIRT2 Nuclear Targeting Are Triggered by the Listeria Virulence Factor, InlB

To identify the bacterial factor(s) necessary for inducing H3K18 deacetylation, we screened L. monocytogenes mutants defective for infection. One mutant, ΔinlB, which is defective for invasion of HeLa cells, did not exhibit H3K18 deacetylation, suggesting that either the InlB protein itself or the entry of bacteria into cells is important for inducing H3K18 deacetylation (Fig. 5A). InlB is a well-expressed surface protein of Listeria that, upon interaction with the cell-surface receptor c-Met, mediates entry of bacteria or beads into nonphagocytic cells. We tested whether a noninvasive species, L. innocua (which, when engineered to express InlB, is able to enter into HeLa cells), can induce H3K18 deacetylation. Although L. innocua had no effect, L. innocua expressing InlB led to H3K18 deacetylation levels similar to those induced by L. monocytogenes (Fig. 5A), suggesting that no other virulence factor besides InlB is required for H3K18 deacetylation. Similar results were obtained with polystyrene beads coated with InlB, which induced H3 deacetylation, whereas uncoupled beads had no effect (Fig. 5A). Furthermore, purified InlB was sufficient to induce H3K18 deacetylation (Fig. 5A). When treating cells with the natural c-Met ligand hepatocyte growth factor (HGF), deacetylation was also observed to similar levels and with the same kinetics as with purified InlB (Fig. 5A and fig. S4A). In contrast, cells treated with epidermal growth factor (EGF), which binds the EGF receptor, had no effect (Fig. 5A and fig. S4A). The correlation between deacetylation of H3K18 and SIRT2 relocalization was also assessed with purified InlB, HGF, and EGF. Nuclear accumulation of SIRT2 was observed in all conditions where deacetylation occurred; accumulation did not occur upon EGF treatment (Fig. 5B). Thus, through the c-Met receptor, InlB or HGF are sufficient to induce SIRT2 nuclear translocation and H3K18 deacetylation.

Fig. 5 H3K18 deacetylation and SIRT2 translocation is mediated by the cellular receptor Met.

(A) Quantification of acetylated H3K18 immunoblots in HeLa cells infected with L. innucua (Li) or L. monocytogenes (Lm) mutant stains or treated with purified proteins (n ≥ 3). Error bars represent SEM. Statistical significance was calculated using a Student’s t test. *P < 0.05; **P < 0.001. (B) Endogenous SIRT2 was detected by immunofluorescence of HeLa cells treated with purified proteins and/or pretreated with chemical inhibitors. Scale bars, 10 μm. (C) HeLa cells are pretreated with wortmannin or LY2940002 before infection (+). (D) Plasmids expressing either WT PI3K or a dominant-negative PI3K were transfected into HeLa cells, which were infected with L. monocytogenes (+). Error bars represent SEM. Statistical significance was calculated using a Student’s t test. *P < 0.05; **P < 0.001.

Met-Induced PI3K/Akt Signaling Is Necessary for H3K18 Deacetylation and SIRT2 Nuclear Recruitment

We next addressed the effect of the signaling cascade downstream of c-Met on H3K18 deacetylation. We treated cells with inhibitors of tyrosine phosphorylation, phosphatidylinositol 3-kinase (PI3K), and Akt, all known to be activated upon binding of InlB to c-Met, and assessed acetyl H3K18 levels and SIRT2 relocalization. When cells were treated with genistein to block tyrosine phosphorylation, neither InlB nor HGF induced H3K18 deacetylation or SIRT2 nuclear targeting within the time frame assayed (Fig. 5B and fig. S4A). To control for the activity of genistein, we monitored levels of P-Akt, a downstream target of the c-Met–dependent tyrosine phosphorylation. Although InlB and HGF induced an increase in the levels of phospho-Akt, this decrease was not observed in samples treated with genistein (fig. S4D).

Next, we used either chemical inhibition of PI3K, with wortmannin or LY294002, or expression of a dominant-negative p85 regulatory subunit to show that PI3K activity was necessary for SIRT2 nuclear accumulation and H3K18 deacetylation (Fig. 5B and fig. S4, A and B). We further assessed the role of Akt by using the chemical inhibitor 1-l-6-hydroxymethyl-chiro-inositol 2-(R)-2-methyl-3-O-octadecylcarbonate (HIMO). This compound blocked InlB- and HGF-dependent H3K18 deacetylation and nuclear relocalization of SIRT2 (Fig. 5B and fig. S4C). Together, these results establish that the signaling cascade mediated by the cell receptor c-Met and the downstream signaling factors PI3K and Akt is one essential pathway linking L. monocytogenes to SIRT2 and H3K18 deacetylation.

L. monocytogenes Infection Induces SIRT2-Dependent Modulation of Host Gene Expression

To further characterize the role of SIRT2 during infection, we searched for genes modulated during infection in a SIRT2-dependent manner. Transcriptome analyses were carried out comparing four different conditions: uninfected HeLa cells, cells infected for 5 hours with L. monocytogenes, and AGK2 pretreated cells with or without infection. When comparing uninfected cells, treated or not with AGK2, we did not identify any genes that were differentially regulated in a significant manner, suggesting that SIRT2 has no effect on resting cells (Fig. 6A). This observation is consistent with our cell fractionation data, which show that in resting cells only a small fraction of SIRT2 is bound to chromatin. In contrast, AGK2 had a significant effect on gene transcription induced by infection. In the absence of AGK2, infection with L. monocytogenes led to activation of 158 genes and repression of 272 genes (table S5). Notably, pretreatment with AGK2 significantly decreased the number of infection-induced activated and repressed genes to 30 and 1, respectively (Fig. 6A). Using these data, we categorized genes as SIRT2-independent if AGK2 pretreatment had no effect on their expression and SIRT2-dependent if AGK2 affected their expression, and we validated these results by quantitative polymerase chain reaction (PCR) on a subset of genes (fig. S5). Because AGK2 displayed the most significant effect on repressed genes and because most genes repressed during infection are dependent on SIRT2 for their expression pattern, we conclude that SIRT2 has a key role in gene repression during infection.

Fig. 6 SIRT2 regulates gene transcription during infection.

(A) Heat-map representation of the mean fold change in gene expression, as determined by transcriptome analysis of Caco2 cells infected for 5 hours (n ≥ 2). Red represents gene activation; blue indicates gene repression. (B and C) ChIP using antibodies targeting SIRT2, H3K18Ac, and H3 was quantified by qPCR (n ≥ 3). H3K18Ac qPCRs are normalized to H3 qPCRs, and SIRT2 qPCR results are represented as percent of the input. Error bars denote SEM. Statistical significance was calculated using a Student’s t test. *P < 0.05; **P < 0.001.

We further tested whether the genes identified by transcriptome analysis were similarly regulated by HGF and the bacterial factor InlB, as with infection. Both treatments with InlB and HGF resulted in the same down-regulation of genes identified in the SIRT2-dependent signature (fig. S6). These results suggest that L. monocytogenes is hijacking SIRT2 to impose a transcriptional control on the host, and this response is mediated by the Met receptor.

H3K18 Deacetylation and SIRT2 Recruitment Occur at the TSS of SIRT2-Regulated Genes

Previous studies in T cells have shown that H3K18 acetylation levels are enriched at transcriptional start sites (TSSs) and enhancers of active mammalian genes (26, 27). We thus probed the TSSs of SIRT2-regulated genes for SIRT2 recruitment and H3K18 deacetylation by chromatin immunoprecipitation (ChIP). Whereas SIRT2-independent and -dependent activated genes exhibited an increase in H3K18 acetylation levels and a loss of SIRT2 at transcription start sites, SIRT2-dependent repressed genes showed an opposite phenotype (Fig. 6, B and C). All tested SIRT2-dependent repressed genes (MYLIP, EHHADH, SYDE2, ERCC5, and LEF1) exhibited a more than 10-fold increase in recruitment of SIRT2 and a significant decrease in the level of acetylated H3K18 upon infection (Fig. 6C). SIRT2 recruitment and H3K18 deacetylation were observed only at the TSSs and not at exon 2, strongly supporting a role for SIRT2 in transcriptional regulation (fig. S7). We further verified that other histone residues were not modified upon SIRT2 recruitment. ChIP experiments with antibodies against AcH3K9, AcH3K14, and AcH4K16 (anti-AcH3K9, anti-AcH3K14, and anti-AcH4K16, respectively) showed that none of the corresponding residues were deacetylated at the genes where SIRT2 was recruited (fig. S8). Because the histone acetyl transferase CBP is displaced by infection with adenovirus, which culminates in a threefold reduction in cellular H3K18 acetylation, we asked whether the same mechanism was acting during a Listeria infection. Therefore, we verified whether CBP was being relocalized by infection with L. monocytogenes as with adenovirus (28). There was no change in the level of CBP binding at SIRT2-repressed genes, where H3K18 deacetylation occurs (fig. S8). Therefore, our data support a model in which infection targets SIRT2 to a subset of genes, where it specifically imposes H3K18 deacetylation and gene repression.

SIRT2 Is Critical for Infection with L. monocytogenes

We next assessed the impact of SIRT2 on infection. We first used either a siRNA approach or a pharmacological approach to perform experiments in tissue culture cells. Infection was quantified by either immunoblot—measuring the levels of a secreted bacterial factor, InlC, which accumulates during infection (18)—or fluorescence-activated cell sorting (FACS) analysis of host cells infected with green fluorescent protein (GFP)–expressing L. monocytogenes. Treating cells with siRNA against SIRT1, -6, or -7 or a SIRT1 chemical inhibitor, CTCC, had no effect on infection (Fig. 7, A and B). In contrast, cells treated with a SIRT2 siRNA or an inhibitor, AGK2, were significantly less infected than untreated cells (Fig. 7, A and B). AGK2 is not toxic to bacteria, nor does it affect the host’s cell cycle (fig. S9). Although the initial stages of infection progressed similarly in the presence or absence of AGK2, at later times of infection, the levels of InlC or the GFP fluorescence were significantly greater in untreated cells than in cells treated with AGK2 (Fig. 7A). Furthermore, Listeria comet tails were not affected by AGK2 treatment, suggesting that loss of SIRT2 activity has no affect on motility (Fig. 7C and fig. S10). In addition, SIRT2 activity was important for infection of the listerial mutant ΔactA, which is defective in cell-to-cell spreading (Fig. 7A). Thus, SIRT2 is required for the late stages of a listerial infection, probably for bacterial replication, in cultured cells.

Fig. 7 SIRT2 is essential for infection by L. monocytogenes.

(A) The level of L. monocytogenes infection is measured by FACS analysis of Caco2 cells pretreated (or not) with AGK2, where the geometric mean (y axis) represents the number of intracellular bacteria (10,000 cells measured; n ≥ 3) at 3, 5, 8, and 24 hours after the start of infection. Error bars denote SEM. *P < 0.05, as measured with a Student’s t test. (B) Quantification of immunoblots detecting the bacterial protein InlC in 5-hour–infected HeLa cells knocked down for SIRTs 1, 2, 6, and 7 by siRNA or in cells treated with chemical inhibitors to deacetylases. Error bars indicate SEM. Statistical significance was calculated using a Student’s t test. *P < 0.05; **P < 0.001. (C) Caco2 cells were treated (or not) with AGK2 and infected with WT L. monocytogenes for 5 hours. Listeria comet tails are visualized by fluorescence imaging of phalloidin and InlC by immunofluorescence (see text). Images are shown as negatives for better visualization. Scale bars, 50 μm. Large inserted boxes show 5× digital magnifications of the smaller boxes.

The impact of SIRT2 on infection was also determined in vivo in Sirt2tm1a(EUCOMM)Wtsi mice (Sirt2−/−) generated at the Sanger Institute. WT or Sirt2−/− mice were infected intravenously with L. monocytogenes and the spleens and livers were collected for bacterial enumeration. In agreement with our in vitro data, the spleens of Sirt2−/− mice were significantly less infected than those of WT mice, confirming the crucial role of SIRT2 on infection in vivo (Fig. 8A). We further tested the role of InlB by comparing infection of a ΔinlB mutant in both WT and Sirt2−/− mice. In contrast to WT L. monocytogenes, a ΔinlB mutant infected both WT and Sirt2−/− mice strains similarly (Fig. 8A), confirming the importance of InlB in hijacking SIRT2 in vivo. We further assessed the levels of H3K18 acetylation in the spleens of infected mice. Whereas deacetylation took place in WT mice, it did not occur in Sirt2−/− mice or in WT mice infected with a ΔinlB mutant (Fig. 8B). Thus, activity of SIRT2 on H3K18 is important for infection in vivo, and InlB is the bacterial factor triggering this activity.

Fig. 8 The activity of SIRT2 is important for in vivo infection by L. monocytogenes.

(A) Colony forming units in spleens of SIRT2+/+ or SIRT2−/− mice infected for 72 hours with L. monocytogenes (Lm) or ΔinlB. Each symbol represents one mouse. (B) Immunoblot analysis of H3K18 acetylation levels in mice spleens. Error bars indicate SEM. Statistical significance was calculated using a Student’s t test. *P < 0.05; **P < 0.001; ns, not significant.

Discussion

Here, we have shown that L. monocytogenes hijacks the host HDAC, SIRT2, to impose a transcriptional program on the host. In addition, we have uncovered a nuclear function for SIRT2 in deacetylating H3 specifically on lysine 18 in response to infection and to activation of the PI3K/AKT signaling cascade.

Histone Deacetylation upon Infection

The study of pathogen-induced histone modifications is a recent field of research. Three bacteria were reported to modulate the levels of acetylated histones during infection, but the underlying mechanisms have remained unknown. During a Mycobacterium tuberculosis infection, histone deacetylation occurs at the promoters of specific genes—HLA-DRα, HLA-DRβ, and CIITA—correlating with transcriptional repression of these genes (29, 30). Infection of gastric epithelial cells by Helicobacter pylori was shown to induce deacetylation of H3K23 but had no effect on H3K9, H3K14, H3K18, H3K23, H3K9K14, or H4K8 (31). Finally, Anaplasma phagocytophilum was reported to activate the expression of genes encoding HDAC1 and -2, correlating with transcriptional repression of key immunity genes and a decrease in a general H3 acetylation levels at the promoter of these same genes (32).

Through binding and sequestering the HATs CBP and p300 to a specific subset of host genes, adenovirus infection causes a threefold reduction in total cellular histone H3K18 acetylation (28). The resulting effect was to stimulate cell cycling and inhibit antiviral responses and cellular differentiation (33). Although the mechanism of H3K18 deacetylation appears different for adenovirus and Listeria, the cellular outcome could be the same. Thus, together with our results, these studies point to histone deacetylation as a common strategy used by pathogens during infection.

An Additional Nuclear Function for SIRT2

The role of SIRT2 has mainly been characterized in the cytoplasm, where it regulates microtubule dynamics through deacetylation of α-tubulin (34) and NF-κB gene expression through deacetylation of p65 (35) and also controls adipocyte differentiation (36) and autophagy (37) through deacetylation of FOXO1. However, recent studies have identified nuclear SIRT2 targets such as p53 and p300 (34). SIRT2-dependent deacetylation of H3K56 was shown to occur after DNA-damage–induced hyperacetylation (38). Additionally, in vitro studies have demonstrated that SIRT1, -2, and -3 can deacetylate H3K18 and H4K16 (39). The apparent contradiction of a protein being cytoplasmically located yet exhibiting specificity for a histone residue was resolved when SIRT2 localization was monitored during cell cycle progression. SIRT2 localizes to the cytoplasm throughout the cell cycle, with the exception of prophase at the beginning of M phase, where it translocates to the nucleus to cause deacetylation of H4K16 (40). However, in interphase cells, only one report describes an equilibrium between nuclear import and export of SIRT2, suggesting a constant shuttling of this protein (24). Our study identifies a stimulus leading to nuclear localization of SIRT2 in interphase cells. Notably, our transcriptome analysis demonstrates that SIRT2 does not regulate transcription in resting cells and plays a role only during infection. Indeed, only in infected cells, in which SIRT2 is recruited to the host chromatin, does SIRT2 have a role in transcriptional repression.

The mechanism of SIRT2 localization to the nucleus remains unknown. We noticed a shift in the migrating band size of SIRT2 upon infection-induced localization to the chromatin (Fig. 4B), suggesting a posttranslational modification of SIRT2. Phosphorylation of SIRT2 has been reported: SIRT2 is phosphorylated at serine 368 by the CDK1 kinase and dephosphorylated by the CDC14A and B phosphatases (41). A smaller SIRT2 isoform is phosphorylated by CDK2 and -5, at the serine residue corresponding to the Ser368 residue of the full-length protein described above (42). Although these phosphorylation events have been shown to affect cell cycle progression and enzymatic activity, no difference in SIRT2 subcellular localization has been reported. Our results show that infection with Listeria provides a practical stimulus that we can use to study the posttranslational modification of SIRT2 and its role in SIRT2 translocation.

PI3K/AKT Signaling and SIRT2

We have shown that the cellular receptor c-Met is important for inducing SIRT2 translocation to the nucleus. We further identified PI3K and AKT as downstream signaling intermediates to H3K18 deacetylation. To date, only one study has correlated AKT signaling with histone modifications (43). In this study, they demonstrate that activation of PI3K/AKT signaling can be a trigger for H3K27 trimethylation at many downstream target genes. The genes marked with H3K27 are different from those identified here, suggesting that the listerial response is different, even though it is mediated by PI3K/AKT. The difference in the signaling response terminating in H3K18 deacetylation is additionally illustrated by our observation that EGF, which also activates PI3K and AKT, does not induce H3 deacetylation or SIRT2 nuclear localization. Furthermore, both AKT signaling and H3K18 deacetylation are correlated with cancer development. The PI3K signaling cascade is one of the most commonly activated signaling pathways in human cancer, and H3K18 is correlated with aggressive cancer phenotypes and poor patient prognosis (44). However, although both PI3K activation and H3K18 deacetylation occur during infection with L. monocytogenes, oncogenic transformation has not been reported after listeriosis. SIRT2 has been described as having tumor-suppressing activities (45, 46). Together, additional factors that remain to be identified are important for providing specificity to the response induced by a L. monocytogenes infection.

Gene Regulation by SIRT2

Although many histone deacetylases have a role in transcriptional regulation, no such role has been directly attributed to SIRT2, as it has not previously been shown to target histones in vivo. Here, we demonstrate a role for SIRT2 in transcriptional regulation, because every down-regulated gene (except one) identified during infection is dependent on SIRT2 for repression. However, because H3 deacetylation is observed globally by immunoblot, we predict that it must also occur in other regions besides the TSSs of specific genes, perhaps in intergenic regions as reported for adenovirus infections (47).

A gene ontology analysis of the 271 repressed genes (listed in table S5) revealed that a significant number of them are DNA binding proteins (51 genes) and/or are implicated in transcriptional regulation (55 genes). A few prominent examples are SMAD1 and FOXM1, transcription factors that participate in a wide range of critical cellular processes, including proliferation, differentiation, and apoptosis; IRF2, a transcription factor important in regulating the interferon response in immune response to infection; SMARCA2, which is a member of the SWI/SNF family of proteins that alter the chromatin structure in an adenosine triphosphate–dependent manner; and SAP130, which is part of the Sin3A repressor complex important in transcriptional repression. Thus, transcriptional remodeling by SIRT2 could be greater than what we report here, because a second wave of transcriptional modulation could follow the SIRT2-dependent deregulation of genes involved in transcriptional regulation. A significant number of SIRT2 down-regulated genes are involved in immune response regulation. For instance, many genes (such as RASGRP1, MAPK14, PIK3R3, PTPNG, SOS1, VAV3, ABL1, CAMK26, MAP2K6, and LEF1) are regulators of B and T cell receptor signaling. The chemokine Cxcl12, which is strongly chemotactic for lymphocytes, and the interferon transcription factor IRF2 are also down-regulated. From these data, we suggest that L. monocytogenes is hijacking SIRT2 to impose a transcriptional control of the host, thereby manipulating many essential cellular functions to promote a listerial infection.

Ingenuity analysis of our Affymetrix (Santa Clara, California) arrays showed that SIRT2 regulates a cluster of 36 genes that increase cell survival. The intracellular lifestyle of L. monocytogenes is consistent with an active mechanism to promote survival of the host cell. Shigella flexneri and H. pylori have been shown to dampen rapid turnover of epithelial cells to prolong colonization within the intestinal epithelial cells (48, 49). Our results suggest that this might also be the case for Listeria.

Role of SIRT2 During Infection

We have shown the importance of SIRT2 in infection, both in tissue culture cells and in vivo. Furthermore, we have demonstrated that SIRT2 plays an essential role in deacetylating H3K18. Its effect on H3K18 during infection appears to be specific, as deacetylation of other known targets such as tubulin and H4K16 does not occur (Fig. 1A). We thus suggest that the main role of SIRT2 in infection is to deacetylate H3K18; however, we cannot exclude that SIRT2 is targeting other proteins during infection, in addition to H3K18. Investigations are in progress to address this issue.

H3K18 was recently found to be deacetylated by SIRT7 in cancer cells (50), suggesting that different deacetylases act on this residue under different conditions.

In summary, our study provides evidence that L. monocytogenes highjacks the host HDAC, SIRT2, to impose a transcriptional program on the host through activation of the PI3K/AKT signaling cascade. We have, therefore, uncovered a nuclear function for SIRT2 in deacetylating H3 specifically on lysine 18. We had previously reported that L. monocytogenes imposes histone modifications (3), but here we demonstrate that a histone modifier is essential for infection.

Materials and Methods

Antibodies

We used the rabbit polyclonal antibodies anti–acetyl-histone H3 K9 (Cell Signaling, catalog no. 9671), anti–acetyl-histone H3 K14 (Cell Signaling, 4318), anti–acetyl-histone H3K18 (Cell Signaling, 9675), anti–histone H3 (Cell Signaling, 9715), anti–histone H4 (AbCam, ab10158), anti–acetyl-histone H4 K16 (Millipore, 06-762), anti–trimethyl-histone H3 K9 (Upstate, 07-442), and anti-SIRT2 (Thermo Scientific, PA3-200). We used the mouse monoclonal antibodies anti–HP1-1α (Euromedex, 2HP-1H5-AS), anti–α-tubulin (Sigma, T6074), and anti-actin (Sigma, A5441).

Cell Culture, Infections, and Inhibitors

The bacterial strains that we used in this study are indicated in table S1. L. monocytogenes strains were grown in brain-heart infusion medium (Difco, Detroit, Michigan) at 37°C until the optical density at 600 nm = 1. When required, antibiotics were added (chloramphenicol, 7 or 35 μg/ml; erythromycin, 5 μg/ml).

HeLa (ATCC CCL-2) and CaCO2 (ATCC HTB-37) cells were cultured in minimum essential medium (MEM) plus GlutaMAX (GIBCO) supplemented with 1 mM sodium pyruvate (GIBCO), 0.1 mM nonessential amino acid solution (GIBCO), and 10% (HeLa) or 20% (CaCO2) fetal calf serum (FCS). RAW 264.7 cells (ATCC TIB-71) were cultured in Dulbecco’s minimum essential medium plus GlutaMAX (GIBCO) supplemented with 2 mM glutamine (GIBCO), 1 mM sodium pyruvate, and 10% FCS.

HeLa and CaCO2 were grown to semiconfluence, at which point they were serum-starved (serum low medium: MEM plus GlutaMAX, 1 mM Na2+ pyruvate, 0.1 mM nonessential amino acid solution, 0.25% FCS) for 24 hours before use in experiments. Exponential phase bacteria were washed twice in the above-mentioned serum-low medium and added to cells at a multiplicity of infection of 50:1 (HeLa and CaCO2). After 1 hour of infection, cells were washed with serum-low medium, and 10 μg/ml gentamycin was added. Infections were carried out for 5 hours unless otherwise indicated.

InlB was isolated as per Ireton et al. (51). Cells were treated with purified phosphate-buffered saline [PBS, InlB (10 ng/ml), HGF (10 ng/ml, Sigma, H5791)] or EGF (100 nM, Sigma, E9644). Polystyrene carboxyl P(S/V-COOH/1) 1.1-μm beads (Bang Laboratories, via biovalley.fr, PC04N-7740-1I) were coated with InlB, as per protocol (Bang Laboratories, Beads Above the Rest, TechNote 205, Covalent Coupling).

For experiments involving pharmacological inhibitors, cells were pretreated for 2 hours before infection with TSA (5 μM, Sigma, T8552), NIC (5 μM, Sigma, N0636), CTCC (5 μM, Enzo Life Sciences, ALX-270-437), and AGK2 (5 μM, Enzo Life Sciences, ALX-270-484) or for 30 min with PBS, dimethyl sulfoxide, wortmannin (100 nM, Sigma, W1628), genistein (10 μM, Sigma, G6776), and HIMO (10 μM, Alexis Biochemicals, ALX-270-292).

Cloning

For overexpression of WT SIRT2, we used a SIRT2-GFP construct (pEGFP-c1 backbone; EGFP, enhanced green fluorescent protein), which was a gift from B. North (24). siRNA-insensitive SIRT2 and SIRT2 N168A were constructed by PCR amplification of the SIRT2-GFP construct with primer pairs: (i) SIRT2 fwd and SIRT2 05 rev, (ii) SIRT2 05 fwd and SIRT2 rev, and (iii) SIRT2 fwd and SIRT2 rev (see tables S3 and S4 for primer sequences), followed by insertion at the EcoRI site of the pEGFP-c1 vector.

Transfections

DharmaFECT (Dharmacon) transfection was used to introduce RNA interference knockdown SIRT1 (Thermo Scientific, siGENOME SMARTpool M-003540_01_0005), SIRT2 (Thermo Scientific, siGENOME SMARTpool M-004826-02-0005, or siRNA fragment D-004826-05), or control scramble siRNA (On-TARGETplus SMARTpool). Cells were assayed 72 hours after siRNA transfection.

SIRT2-FLAG or SIRT2N168A-FLAG were rendered sensitive to siRNA by the addition of three point mutations by PCR amplification using primers SIRT2 fwd with SIRT2 rev 05 and SIRT2 fwd 05 with SIRT2 rev (see tables S3 and S4 for primer list).

Immunoblotting and Cell Fractionation

Total cell lysates were harvested by removing growth medium and adding lysis buffer [1M Tris HCl (pH 6.8 at 25°C), 10% SDS, 50% glycerol, 0.05% bromophenol blue, 10% β-mercaptoethanol]. Spleen and liver samples were collected as described in (52). Samples were boiled for 5 to 10 min, sonicated for 5 s, and loaded on a 15% acrylamide gel. A semidry transfer was conducted for 1 hour, at 32 mA per transfer, followed by blocking of the Hybond P–polyvinylidene difluoride membranes (GE Healthcare) in tris-buffered saline–Tween [Tris 50 mM (pH 8), NaCl 15mM, 0.1% Tween] supplemented with 10% milk. Transferred membranes were incubated with previously mentioned primary antibodies for 2 hours at 25°C or at 4°C. Membranes were washed and incubated with horseradish peroxidase–conjugated goat α-rabbit or α-mouse antibodies (Biosys Laboratories). Quantification of Western blots was performed using the G:box-ichemi machine (SynGene).

Cell fractionation was conducted as follows: Cells were resuspended in buffer A [20 mM HEPES (pH 7.0), 0.15 mM EDTA, 0.15 mM EGTA, 10 mM KCl]. 1% NP40 was added, followed by SR buffer [50 mM HEPES (pH 7.0), 0.25 EDTA, 10 mM KCl, 70% (m/v) saccharose]. Samples were centrifuged for 5 min at 2000 g. The supernatant was isolated as the cytosolic fraction and recentrifuged for 20 min at 20,000 g to eliminate cell nuclear debris. The pellet was washed in buffer B [10 mM HEPES (pH 8.0), 0.1 mM EDTA, 100 mM NaCl, 25% (v/v) glycerol] and centrifuged for 5 min at 2000 g. Buffers A, B, and SR were supplemented with 0.15 mM spermidine, 0.15 mM spermine, 1 mM dithiothreitol, and 1× Complete (Roche). The washed pellet was resuspended in sucrose buffer [20 mM Tris (pH 7.65), 60 mM NaCl, 15 mM KCl, 0.34 M sucrose, 0.15 mM spermine, 0.15 mM spermidine], followed by the addition of a high-salt buffer [20 mM Tris (pH 7.65), 0.2 mM EDTA, 25% glycerol, 900 mM NaCl, 1.5 mM MgCl2] to obtain a final salt concentration of 250 mM. Samples were incubated for 25 min and centrifuged for 10 min at 10,000 revolutions per minute (rpm). The supernatant was isolated as the nuclear soluble fraction from the pellet, which represents chromatin and nuclear insoluble material. The pellet was resuspended in sucrose buffer + MNase (0.0025 units/ul) and 1 mM CaCl2 and was incubated at 37°C for 10 min. 4 mM EDTA was added, and samples were sonicated using the Bioruptor (Diagenode) for 7.5 min (15 s on and 1 min off) and centrifuged for 15 min at 13,000 rpm. The supernatants represent a soluble chromatin fraction.

Immunofluorescence and FACS Analysis

Cells were grown on glass cover slides. After infection, cells were fixed in 4% paraformaldehyde and permeabilized in 0.3% triton for 15 min. Immunostaining was performed with an anti-SIRT2 antibody (Thermo Scientific, PA3-200) in 1% BSA+ 0.1% Tween 100. Infections for immunofluorescence and FACS were carried out with GFP-expressing L. monocytogenes. Infected cells were monitored with a FACScalibur (BD Bioscience), and analysis was done with Flowjo software (www.flowjo.com/).

Microarray Analysis

Total mRNA from uninfected and infected cells, pretreated or untreated with AGK2 (5 μM for 2 hours), was extracted and purified as per an RNeasy kit (Qiagen, Valencia, California). Quality assessment and normalization of the arrays was performed with the tools available in the Expression Console v1.1 (Affymetrix) and Bioconductor packages. Total RNA (200 ng) was reverse-transcribed and amplified per the manufacturer’s protocols using the Applause WTA Amp-Plus System (Nugen Technologies, 5510-24) and was fragmented and biotin-labeled using the Encore Biotin Module (Nugen Technologies, 4200-12). Gene expression was determined by hybridization of the labeled template to HuGene 1.0 ST microarrays (Affymetrix, Santa Clara, California). Hybridization cocktail and posthybridization processing were performed according to the “Target Preparation for Affymetrix GeneChip Eukaryotic Array Analysis” protocol found in the appendix of the Nugen protocol of the fragmentation kit. Arrays were hybridized for 18 hours, washed using fluidics protocol FS450_0007 on a GeneChip Fluidic Station 450 (Affymetrix), and scanned with an Affymetrix Genechip Scanner 3000, generating CEL files for each array. Three biological replicates were run for each condition.

Gene-level expression values were derived from the CEL file probe-level hybridization intensities using the model-based robust multichip average algorithm (RMA) (53). RMA performs normalization, background correction, and data summarization. An analysis was performed using the Limma t test (54), and a P value threshold of P < 0.05 was used as the criterion for expression. The estimated false discovery rate (FDR) of this analysis was calculated with the Benjamini-Hochberg approach (55) to correct for multiple comparisons.

qRT-PCR

Total mRNA was extracted using the RNeasy kit (Qiagen). Reverse transcription was performed with the iScript cDNA Synthesis kit (BioRad). Qualitative PCR (qPCR) was done using the SsoFast EvaGreen Supermix (BioRad) and run on a MyIQ device (BioRad). Data were analyzed by the ∆∆Ct method.

Chromatin Immunoprecipitation

Preparation of ChIP samples was adapted from Lebreton et al. (56). Chromatin inputs corresponded to 1.8 × 105 cells for each individual ChIP assay. All buffers were supplemented with Complete EDTA-free protease inhibitor cocktail tablets (Roche). Formaldehyde-fixed cells were washed in PBS and lysed in 10 mM Tris (pH 8), 10 mM EDTA, 0.5 mM EGTA, 0.25% Triton X-100 for 5 min on ice. The nuclear pellets were recovered by brief centrifugation at 3000 × g, and the soluble nuclear fraction was extracted with 250 mM NaCl, 50 mM Tris (pH 8), 1 mM EDTA, 0.5 mM EGTA for 30 min on ice. After brief centrifugation at 16,000 × g, chromatin pellets were resuspended in 10 mM Tris (pH 8), 1 mM EDTA, 0.5 mM EGTA, 0.5% SDS and then sonicated with a Bioruptor (Diagenode) to shear chromatin to a final size of 150 to 600 base pairs. Extracts were quantified by an absorbance of 260 nm, and material quantities were adjusted accordingly. Samples were then diluted to obtain the following immunoprecipitation (IP) buffer composition: 150 mM NaCl, 10 mM Tris (pH 8), 0.1% SDS, 1% Triton X-100, 0.1% sodium deoxycholate, 1 mM EDTA, 0.5 mM EGTA. IP was carried out in the on setting at 4°C with antibodies to SIRT2, H3, H3K18 Ac, H4, and H4K16 Ac. Immunocomplexes were recovered with Dynabeads Protein G (Invitrogen), added for 90 min at 4°C, and then washed five times in a succession of isotonic and saline buffers as described in (57). After a final wash in 10 mM Tris pH8, 1 mM EDTA, 0.01% Igepal, bound material was eluted by the addition of water containing 10% Chelex (Bio-Rad), followed by boiling for 10 min to reverse the cross-link. Samples were then incubated with proteinase K (100 μg/ml) for 30 min at 55°C with some shaking and then boiled for another 10 min. Finally, the ChIP DNA fraction was separated from beads and the Chelex matrix by centrifugation. Recovered supernatants were quantified by qualitative real-time PCR (qRT-PCR) using the ∆∆Ct method. Results for samples immunoprecipitated with AcH3K18 were normalized to samples immunopecipitated with H3. The sequences for the primers used are given in tables S3 and S4.

Sirt2 Mice

Sirt2tm1a(EUCOMM)Wtsi mice were obtained from the Sanger Center. For details, see www.informatics.jax.org/javawi2/servlet/WIFetch?page=alleleDetail&key=606707. Infections were performed by intravenous injection of 105 bacteria per animal. Experiments were performed according to the Institut Pasteur guidelines for animal experimentation.

Immunoprecipitation

Immunoprecipitation of SIRT2-FLAG was performed with M2-FLAG affinity gel (Sigma, A2220), according to the manufacturer’s protocol. Elution was performed in 0.1 M glycine HCl, pH 3.5.

SDS-PAGE and LC-MS/MS Analysis

Proteins were separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) on a 4 to 15% polyacrylamide gel (Bio-Rad) and stained by colloid coommassie blue (Invitrogen). For every sample, the gel lane was cut into five consecutive gel slices that were washed with H2O, incubated for 15 min with water/acetonitrile (1/1, v/v), and incubated for 15 min with 100% acetonitrile before drying completely in a vacuum concentrator. Sequencing-grade trypsin (0.25 μg, Promega) in 50 mM ammonium bicarbonate in water/acetonitrile (9/1, v/v) was added to the dried gel slices, and proteins were digested overnight at 37°C. Peptides eluted from every gel slice were dried completely in a vacuum concentrator and redissolved in 15 μl solvent A [0.1% formic acid in water/acetonitrile (98/2, v/v)], 5 μl of which were used for liquid chromatography–tandem mass spectrometry (LC-MS/MS) analysis on an Ultimate 3000 high-performance LC system (Dionex) in line connected to an LTQ Orbitrap Velos mass spectrometer (Thermo Electron). Trapping was performed at 10 μl/min for 4 min in solvent A on a PepMap C18 column [0.3-mm inner diameter by 5 mm (Dionex)]; after back-flushing from the trapping column, the sample was loaded on a reverse-phase column (made in house, 75-μm inner diameter by 150 mm, 3-μm beads C18 Reprosil-HD, Dr. Maisch, Entringen, Germany). Peptides were eluted by a linear increase from 2 to 55% solvent B [0.08% formic acid in water/acetonitrile (2/8, v/v)] over 30 min at a constant flow rate of 300 nl/min.

The mass spectrometer was operated in data-dependent mode, automatically switching between MS and MS/MS acquisition for the 10 most abundant ion peaks per MS spectrum. Full-scan MS spectra [mass-to-charge ratio (m/z) of 300 to 2000] were acquired at a resolution of 60,000 in the orbitrap analyzer after accumulation to a target value of 1,000,000. The 10 most intense ions above a threshold value of 5000 were isolated for fragmentation by collision-induced dissociation at a normalized collision energy of 35% in the linear ion trap (LTQ) after filling the trap at a target value of 5000 for a maximum of 50 ms. From the MS/MS data in each LC run, Mascot generic files were created using the Mascot Distiller software (version 2.4.3.3, Matrix Science). When generating peak lists, grouping of spectra was performed with m/z 0.005 tolerance on the precursor ion, a maximum intermediate retention time of 30 s, and a maximum intermediate scan count of five. A peak list was only generated when the MS/MS spectrum contained more than 10 peaks, no deisotoping was performed, and the relative signal-to-noise limit was set at two.

Generated peak lists were then searched with Mascot using the Mascot Daemon interface (version 2.3.0, Matrix Science) against the human proteins in the Swiss-Prot database (database release version of 7 July 2012 containing 20,235 human protein sequences). Variable modifications were set to oxidation of methionine residues; pyroglutamate formation of N-terminal glutamine residues; acetylation of peptide N termini and lysine residues; di-glycine modification of lysine residues; and phosphorylation of serine, threonine, and tyrosine residues. Mass tolerance of the precursor ions was set to 10 parts per million; mass tolerance of the fragment ions was set to 0.5 daltons. The peptide charge was set to 1+, 2+, or 3+, and one missed tryptic cleavage site was allowed. Also, the C13 setting of Mascot was set to 1. Only peptides that were ranked first and scored above the threshold score set at 99% confidence were withheld. For processing of all MS data, the ms_lims software platform was used [PubMed identifier (PMID) 20058248]. In addition, the C13 setting of Mascot was set to 1. Only peptides that were ranked first and scored above the threshold score set at 99% confidence were withheld. In total, 18,834 fragmentation spectra were identified (with a FDR of 0.23%) (PMID 18067246). For processing of all MS data, the ms_lims software platform was used (PMID 20058248).

Histone Purification

Histones were purified according to protocol published in (58) using trichloroacetic acid precipitation.

SDS-PAGE, in Gel Propionylation and LC-MS/MS Analysis

Deacetylation of H3K18 by differential MS was performed using an adapted version of the protocol of Garcia et al. (59). Briefly, equal amounts (100 μg) of purified histones from infected and uninfected samples were separated by SDS-PAGE on a 16.5% Tris-tricine gel (Biorad) and stained by coommassie blue. Protein bands corresponding to histone H3 were cut and modified by in-gel differential propionylation (60, 61). Histone H3 bands were cut and washed with H2O, incubated for 15 min with water/acetonitrile (1/1, v/v), and incubated for 15 min with 100% acetonitrile before being dried completely in a vacuum concentrator. Isotopically light (12C3)– or heavy (13C3)–labeled N-hydroxysuccinimide propionate (1 mg) (provided by K. Gevaert and B. Ruttens, Ghent University, Belgium) was then added in 150 μl 100 mM sodium phosphate (pH 8.0) to the gel pieces of the EGD infected and uninfected samples, respectively, and both samples were incubated for 1 hour at 30°C. To ensure quantitative labeling, this step was repeated after washing and drying the gel samples again, as described above. After this second propionylation step, gel pieces were washed twice with 200 μl 50 mM ammonium bicarbonate (pH 8.0) for 10 min to remove and quench excess reagents. Water-hydroxylamine (2 μl, 1/1, v/v, Sigma) was then added to the final wash solution, and samples were incubated for 20 min at room temperature to revert possible O-acylation of Ser, Thr, or Tyr residues. To ensure complete oxidation of methionine residues, gel samples were washed with 200 μl 0.1% trifluoroacetic acid (TFA) for 10 min at 30°C and then incubated with 200 μl 0.5% hydrogen peroxide in 0.1% TFA for 30 min at 30°C. Gel samples were washed for 10 min with 200 μl acetonitrile and dried completely in a vacuum concentrator. Sequencing-grade trypsin (0.15 μg, Promega) in 50 mM ammonium bicarbonate in water/acetonitrile (9/1, v/v) was added to the dried gel slices, and proteins were digested overnight at 37°C. Equal amounts of peptides eluted from both gel samples were mixed, dried completely in a vacuum concentrator, and redissolved in 40 μl solvent A [0.1% formic acid in water/acetonitrile (98/2, v/v)], 5 μl of which was injeced for LC-MS/MS analysis on a LTQ Orbitrap Velos mass spectrometer (Thermo Electron). LC-MS/MS analysis, peak list generation, and Mascot searches occurred with only small adjustments to the search settings: Oxidation of methionines residues was set as fixed modification (+15.994915 daltons), whereas acetylation (+42.010565 daltons) and light (+56.026215 daltons) and heavy (+59.036279 daltons) propionylation of lysine residues were set as variable modifications. Also, a maximum of two missed tryptic cleavage sites was allowed.

A recent paper has reported that the bacterium Legionella secretes a protein named RomA that trimethylates K14 of histone H3, leading to repression of a number of host genes (63).

Supplementary Materials

References and Notes

  1. Single-letter abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; H, His; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.
  2. Acknowledgments: We thank G. Dougan and S. Clare for their help with the Sirt2tm1a(EUCOMM)Wtsi mice, E. Verdin for providing the SIRT2N168 mutant, C. Becavin for his assistance with the heatmap, J. Chamot-Rooke and M. Duchateau for the use of the mass spectrometers, and K. Gevaert and B. Ruttens (VIB, Ghent University, Belgium) for providing N-hydroxysuccinimide propionate reagents. We extend a special thanks to A. Le for his time and effort during the summer of 2009. Work in the Cossart laboratory is funded by the Pasteur Institute, Inserm (U604), Institut National de la Recherche Agronomique (USC2020), the Fondation Louis Jeantet, the European Research Council (advanced grant 233348 MODELIST), the Fondation Le Roch Les Mousquetaires, and the Agence Nationale pour la Recherche grants (ERANET “LISTRESS” and EPILIS). We have received funding for this study from the French government’s Investissement d’Avenir program, Laboratoire d’Excellence “Integrative Biology of Emerging Infectious Diseases” (grant ANR-10-LABX-62-IBEID). H.A.E. is a doctoral student supported by the Fondation pour la Recherche Médicale, F.I. is a Postdoctoral Fellow of the Research Foundation Flanders (FWO-Vlaanderen), and P.C. is an international senior research scholar of the Howard Hughes Medical Institute. The transcriptome analysis data for this publication can be found in ArrayExpress with the accession number E-MEXP-3912.
View Abstract

Stay Connected to Science

Navigate This Article