Mycobacterium Tuberculosis Eis Protein Initiates Suppression of Host Immune Responses by Acetylation of DUSP16/MKP-7

Mycobacterium tuberculosis Eis protein initiates suppression of host immune responses by acetylation of DUSP16/MKP-7

Kyoung Hoon Kima, Doo Ri Anb, Jinsu Songa, Ji Young Yoona, Hyoun Sook Kima, Hye Jin Yoona,HaNaImb, Jieun Kimb, Do Jin Kima, Sang Jae Leea, Ki-Hye Kimc, Hye-Mi Leec, Hie-Joon Kima, Eun-Kyeong Joc, Jae Young Leed, and Se Won Suha,b,1

Departments of aChemistry and bBiophysics and Chemical Biology, College of Natural Sciences, Seoul National University, Seoul 151-742, Korea; cDepartment of Microbiology, Chungnam National University School of Medicine, Daejeon 301-747, Korea; and dDepartment of Life Science, Dongguk University-Seoul, Seoul 100-712, Korea

Edited by David Eisenberg, University of California, Los Angeles, CA, and approved April 3, 2012 (received for review December 9, 2011) The intracellular pathogen Mycobacterium tuberculosis (Mtb) components and play an essential role in the regulation of innate causes tuberculosis. Enhanced intracellular survival (Eis) protein, immune signaling during mycobacterial infection (6). Because secreted by Mtb, enhances survival of Mycobacterium smegmatis intracellular survival of Mtb plays a central role in its pathogenesis (Msm) in macrophages. Mtb Eis was shown to suppress host im- (7), it is important to understand the survival strategies of this mune defenses by negatively modulating autophagy, inﬂamma- bacterium within macrophages. Mtb has evolved a number of tion, and cell death through JNK-dependent inhibition of reactive highly effective survival strategies inside the macrophage (8). The oxygen species (ROS) generation. Mtb Eis was recently demon- best-characterized survival mechanism of Mtb is the inhibition of strated to contribute to drug resistance by acetylating multiple phagosomal maturation and autophagy, between which a func- amines of aminoglycosides. However, the mechanism of enhanced tional overlap was suggested (8–11). Both processes involve sev- intracellular survival by Mtb Eis remains unanswered. Therefore, eral steps, including fusion with lysosomes, and a number of we have characterized both Mtb and Msm Eis proteins biochemi-

protein factors, such as Beclin 1 and vacuolar sorting protein 34 BIOCHEMISTRY cally and structurally. We have discovered that Mtb Eis is an effi- ɛ (VPS34), the class III phosphatidylinositol 3-kinase (12). The cient N -acetyltransferase, rapidly acetylating Lys55 of dual- identification and characterization of mycobacterial proteins that specificity protein phosphatase 16 (DUSP16)/mitogen-activated play a role in facilitating intracellular survival remain a priority for protein kinase phosphatase-7 (MKP-7), a JNK-specific phosphatase. α the development of new antituberculosis drugs. In contrast, Msm Eis is more efficient as an N -acetyltransferase. The Rv2416c gene of Mtb H37Rv strain was found to enhance We also show that Msm Eis acetylates aminoglycosides as readily intracellular survival of Mycobacterium smegmatis (Msm) in the as Mtb Eis. Furthermore, Mtb Eis, but not Msm Eis, inhibits LPS- human macrophage-like cell line U-937, and thus it was desig- induced JNK phosphorylation. This functional difference against nated as eis (enhanced intracellular survival) (7). The expression DUSP16/MKP-7 can be understood by comparing the structures of of its protein product directly correlated with the enhanced two Eis proteins. The active site of Mtb Eis with a narrow channel mycobacterial survival in U-937 cells (7). The Mtb Eis protein is fi seems more suitable for sequence-speci c recognition of the pro- produced during human tuberculosis infection and is released tein substrate than the pocket-shaped active site of Msm Eis. We into the culture medium (3). The sigma factor SigA was shown to propose that Mtb Eis initiates the inhibition of JNK-dependent bind to the eis promoter in the W-Beijing strain of Mtb, and the autophagy, phagosome maturation, and ROS generation by acety- activation of the Mtb eis gene correlated with increased SigA lating DUSP16/MKP-7. Our work thus provides insight into the levels and enhanced intracellular survival (13). Treatment of T mechanism of suppressing host immune responses and enhancing cells with Mtb Eis inhibited ERK1/2, JAK pathway, and sub- mycobacterial survival within macrophages by Mtb Eis. sequent production of TNF-α and IL-4 (14). Mtb Eis negatively regulated the secretion of TNF-α and IL-10 by primary human Rv2416c | lysine acetylation | antituberculosis drug monocytes in response to infection with the pathogen (15). Recently Mtb Eis was shown to suppress host innate immune early one-third of the world’s population is infected with defenses by negatively modulating inflammation, autophagy, and NMycobacterium tuberculosis (Mtb). This pathogenic bacte- cell death in a redox-dependent manner (16). The reported data rium causes tuberculosis, which claims the lives of millions of indicate that Mtb Eis plays an essential role in regulating both the people every year (1). Tuberculosis has also become a global early generation of reactive oxygen species (ROS) and in- health issue owing to the increased incidences of multidrug-re- flammatory responses in macrophages (16). It was also found that sistant and extensively drug-resistant strains of Mtb (2). This makes a search for targets of new antituberculosis drugs urgent. Mtb is a highly successful human pathogen, surviving and mul- Author contributions: K.H.K., D.R.A., E.-K.J., J.Y.L., and S.W.S. designed research; K.H.K., tiplying within the human macrophage cells of the infected D.R.A., J.S., J.Y.Y., H.S.K., H.J.Y., H.N.I., J.K., D.J.K., S.J.L., K.-H.K., and H.-M.L. performed people (3). Therefore, treatment of tuberculosis is difficult, re- research; J.S., H.-J.K., E.-K.J., and S.W.S. contributed new reagents/analytic tools; K.H.K., quiring many months of taking a combination of antibiotics. Mtb J.S., J.Y.Y., H.S.K., H.J.Y., H.-J.K., E.-K.J., J.Y.L., and S.W.S. analyzed data; and K.H.K., J.S., E.-K.J., J.Y.L., and S.W.S. wrote the paper. has the ability to persist in the form of a long-term asymptomatic The authors declare no conflict of interest. infection, referred to as latent tuberculosis (4). Latent tuberculosis becomes activated when the body’s immune system is This article is a PNAS Direct Submission. weakened. As a result, tuberculosis is the major cause of death Data deposition: The crystallography, atomic coordinates, and structure factors reported in this paper have been deposited in the Protein Data Bank, www.pdb.org [PDB ID code among immuno-compromised AIDS patients (5). 3RYO (M. tuberculosis Eis, in complex with acetyl CoA), 3UY5 (M. tuberculosis Eis, apo), In mycobacterial infection, host innate immune responses may and 3SXN (M. smegmatis Eis, in complex with CoA)]. play a crucial role in early protection against Mtb infection, leading 1To whom correspondence should be addressed. E-mail: [email protected]. to establishment of effective adaptive immunity to tuberculosis This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. (6). Additionally, MAPK pathways are activated by Mtb or its 1073/pnas.1120251109/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1120251109 PNAS Early Edition | 1of6 Downloaded by guest on September 27, 2021 abrogated production of both ROS and proinflammatory cyto- aminoglycosides as quickly as, or more rapidly than, Mtb Eis (Fig. kines by Mtb Eis depends on its N-acetyltransferase domain in S2). Steady-state kinetic parameters, as measured by Km and kcat the N terminus (16). Enhanced macrophage survival by Mtb Eis values (Table S1), indicate that the aminoglycoside acetyltrans- was found to occur through the regulation of ROS signaling, ferase activity of Msm Eis is comparable to or higher than that of which was JNK-dependent but was not p38- or ERK1/2-de- Mtb Eis. This result cannot explain the enhanced intracellular pendent (16). Forced expression of dual-specificity protein survival of mycobacteria by Mtb Eis. phosphatase 16 (DUSP16), also called MAPK phosphatase- To understand the observed catalytic properties, we have de- 7(MKP-7), suppressed activation of MAPKs in COS-7 cells in the termined and compared the crystal structures of both Mtb and order of selectivity, JNK >> p38 > ERK, suggesting that Msm Eis proteins (Table S2 and SI Results and Discussion). The DUSP16/MKP-7 works as a JNK-specific phosphatase (17). crystal structure of selenomethionine-substituted Mtb Eis in the A bioinformatic analysis predicted that the Mtb Eis protein con- acetyl CoA-bound form was determined by de novo phasing using tains a single acetyltransferase domain of the GCN5-related N- the single anomalous diffraction data to 2.80 Å. This model was acetyltransferase (GNAT) superfamily in the amino terminus (15). used to solve the structures of Mtb Eis in the apo form at 2.46 Å The acetyltransferase domain is predicted to cover residues 9−160 of and Msm Eis in the CoA-bound form at 2.03 Å by molecular re- the 408-residue protein and contains a variant of the characteris- placement. The overall monomeric and hexameric structures of tic sequence motif (V/I-x-x-x-x-Q/R-x-x-G-x-G/A) for acetyltrans- Mtb and Msm Eis proteins are similar to each other (Fig. S3). That ferases at positions between 93 and 103 (93VAPTHRRRGLL103) is, each monomer of both Eis proteins comprises three “structural” (Fig. S1) (15, 18). Our sequence numbering of Mtb Eis follows domains, and six subunits are associated to form a hexamer of 32 the current EXPASY UniProtKB/Swiss-Prot database; six residues symmetry. “Structural” domain 1 adopts the GNAT fold, as pre- 1MPQSDS6 at the amino terminus are missing from other databases, dicted. Unexpectedly, “structural” domain 2 is also folded into the owing to a different translation initiation at Val7. Increased ex- GNAT structure (Fig. S4) despite an apparent lack of sequence pression of Mtb Eis due to a mutation in the promoter region of similarity to other GNAT enzymes, including the Eis domain 1. the eis gene conferred resistance to aminoglycoside kanamycin, a “Structural” domain 3, containing a putative peroxisome targeting second-line antituberculosis drug (19). More recently, Mtb Eis signal type 2 (Fig. S5), resembles sterol carrier protein-2. Either was demonstrated to have an unprecedented ability to acetylate acetyl CoA or CoA is observed to be bound to “structural” domain multiple amines of many aminoglycosides, and its reaction mech- 1only,butnotto“structural” domain 2, in the ligand-bound anism was proposed on the basis of structural and mutational structures of both Eis proteins (Fig. S3). The active sites of Mtb and studies (20). However, the fundamental question about the mecha- Msm Eis proteins are large and deep enough to accommodate nism of enhanced survival of the Msm clone, which contains the aminoglycosides. However, they display distinct structural features extra Mtb eis gene in addition to its own eis gene, still remains that may explain the observed functional difference against peptide unanswered. Nonpathogenic Msm contains a homologous eis gene substrates (as discussed below). While we were preparing this ar- (MSMEG_3513) that encodes a homolog of Mtb Eis (58% amino ticle, the structure of Mtb Eis bound with acetyl CoA was reported acid sequence identity). by another group (20) [Protein Data Bank (PDB) code 3R1K]. Our To better understand how Mtb Eis enhances mycobacterial structure of acetyl CoA-bound Mtb Eis is highly similar to the survival in macrophages, we have carried out functional and reported one, with the rmsds being 0.34 Å for 396 Cα atoms in structural characterization of Eis proteins from both Mtb and a monomer (for chains A) and 0.67–0.75 Å for 2,376 Cα atoms in Msm.Wefind that both Eis proteins can acetylate aminoglyco- a hexamer. A minor difference is in the modeled residues (residues sides efficiently. We have discovered that Lys55 within the 8–161 and 167–408 in our structure; residues 9–57 and 62–408 in docking domain (also called the kinase interaction motif) of the reported structure). DUSP16/MKP-7 is readily acetylated by Mtb Eis, but not by Msm Eis. Furthermore, we show that Mtb Eis, but not Msm Eis, sup- Identification of DUSP16/MKP-7 as the Nɛ-Acetylation Target of Mtb presses LPS-induced JNK phosphorylation and proinflammatory Eis. Besides aminoglycosides, Mtb Eis was previously shown to cytokine production in macrophages. The observed functional acetylate free histone proteins, but not the histone proteins in difference between these Eis proteins against DUSP16/MKP-7, a nucleosomal complex (21). None of these acetylation targets can a JNK-specific phosphatase, can be understood by comparing explain the enhanced intracellular survival of mycobacteria by their active site features. The overall monomeric and oligomeric Mtb Eis; Mtb Eis likely has other unidentified protein targets for structures of both Mtb Eis and Msm Eis are highly similar to each acetylation. Therefore, we were interested in identifying physio- other. The most notable structural difference between them is the logically more important protein acetylation targets of Mtb Eis. presence of a narrow channel for potential sequence-specific Mtb Eis enhances macrophage survival through the regulation of binding of the acetylation target peptide in the active site of Mtb JNK-dependent ROS signaling (16), whereas DUSP16/MKP-7 Eis but not in Msm Eis. On the basis of these findings, we propose works as a JNK-specific phosphatase in vivo (17). Acetylation at that Mtb Eis initiates the inhibition of autophagy and phagosome Lys57 of DUSP1/MKP-1, a nuclear-localized phosphatase that ɛ maturation in infected macrophages by N -acetylating Lys55 of inactivates MAPK members by dephosphorylation, promoted the DUSP16/MKP-7 to suppress host immune responses for in- interaction of DUSP1/MKP-1 with its substrate p38 MAPK and tracellular survival of mycobacteria. inhibited innate immune signaling (22, 23). Three major sub- families of MAPKs are ERKs, p38 MAPKs, and JNKs. These Results MAPKs are activated by MAPK kinases, which are in turn acti- Eis Proteins from both Mtb and Msm Acetylate Aminoglycosides vated by a set of MAPK kinase kinases. The MAPK pathways that Efficiently and Have Similar Overall Structures. Recently Mtb Eis mediate innate immune signaling include MAPK kinases 3/4/6, was shown to acetylate multiple amines of many aminoglycosides, p38, and JNKs (ref. 22 and references therein). including the second-line injectable antituberculosis drugs kana- On the basis of these reports, we speculated that Mtb Eis may mycin and amikacin (19, 20). It was also argued that kanamycin acetylate component(s) of the MAPK signaling pathways, such as may not be a natural substrate of Mtb Eis because of a high Km MAPK phosphatases, to negatively control autophagy, phagosome value with it (19). To examine whether there is a functional dif- maturation, and ROS generation, ultimately leading to the sup- ference between Eis proteins from Mtb and Msm in aminoglyco- pression of host immune responses. Therefore, we examined side acetylation, we compared acetylation activities of these two whether Mtb Eis may acetylate DUSP16/MKP-7, which is known Eis proteins against amikacin, kanamycin A, tobramycin, and tobeaJNK-specific phosphatase (17). We tested a peptide paromomycin. Interestingly, Msm Eis acetylated the tested within the MAPK-docking domain of DUSP16/MKP-7:

2of6 | www.pnas.org/cgi/doi/10.1073/pnas.1120251109 Kim et al. Downloaded by guest on September 27, 2021 53LMKRRLQQDKVLIT66 [MKP-7(53−66)] (22). The underlined Our data thus establish that Eis proteins from both Mtb and Msm Lys62 was predicted to be the acetylation site in the DUSP16/MKP- are catalytically active as aminoglycoside N-acetyltransferases, but ɛ 7 docking domain (22). We also tested a peptide within the MAPK- only Mtb Eis acts as an efficient N -acetyltransferase that can docking domain of DUSP1/MKP-1: 50TIVRRRAKGAMGLE63 acetylate DUSP16/MKP-7. [MKP-1(50−63)] (22). The underlined lysine was established as the We further tested the protein N-acetyltransferase activity using acetylation site in the DUSP1/MKP-1 docking domain (22). With recombinant human DUSP16/MKP-7. We could express DUSP/ these peptides as possible substrates, we performed in vitro acety- MKP-7(1−153), which covers the rhodanese-like domain and lation assays using either Mtb Eis or Msm Eis. encompasses the docking domain, and DUSP/MKP-7(1−303), The MALDI-TOF mass spectra of the MKP-7(53−66) peptide which additionally contains the phosphatase domain, whereas the show that the unmodified peptide peak at 1,741 Da is shifted to full-length DUSP16/MKP-7 was not expressed in Escherichia coli. 1,783 Da after in vitro acetylation for 5 min by Mtb Eis (Fig. 1A). When DUSP16/MKP-7(1−153) and DUSP/MKP-7(1−303) were The observed increase of 42 Da in the peptide mass corresponds to incubated with the wild-type Mtb Eis in the presence of [14C]- covalent attachment of an acetyl group (CH3CO−). On the other acetyl CoA, they were efficiently acetylated by Mtb Eis (Fig. 2A). hand, the mass (1,558 Da) of the MKP-1(50−63) peptide before However, their acetylation by Msm Eis was much less efficient and after in vitro acetylation for 30 min by Mtb Eis agrees with the (Fig. 2A). We next performed mutation experiments to confirm predicted mass of the unmodified peptide (Fig. 1A). These results the acetylation site(s) in DUSP16/MKP-7(1−153). Each of Lys52, indicate that Mtb Eis quickly attaches a single acetyl group to the Lys55, and Lys62 in DUSP16/MKP-7(1−153) was replaced by MKP-7(53−66) peptide, whereas it does not readily acetylate the arginine, and the acetylation reaction was carried out. Mutation of MKP-1(50−63) peptide. The MKP-7(53−66) peptide has three Lys55 drastically reduced the level of acetylation by Mtb Eis, α potential N-acetylation sites: one N -acetylation site at the amino whereas mutations of Lys52 and Lys62 showed no such drastic ɛ terminus and two N -acetylation sites at Lys55 and Lys62. To reduction (Fig. 2B). With Msm Eis, mutation of Lys55 had little identify which of the three possible acetylation sites was actually effect on the level of acetylation, suggesting that acetylation by modified, we performed de novo sequencing of the acetylated Msm Eis likely occurred at the amino terminus, but not at Lys55 MKP-7(53−66) peptide by MALDI tandem mass spectrometry (Fig. 2B). These results strongly suggest that Mtb Eis functions as ɛ (Fig. 1B). Peaks of unacetylated a1, a2, y10, and y11 ions are a protein N -acetyltransferase toward human DUSP/MKP-7. present at 86.009, 217.157, 1,214.653, and 1,369.811 Da, respectively, whereas peaks of acetylated b-ion series b3*, b4*, b5*, Mtb Eis Structure Reveals a Narrow Channel for Peptide Recognition. and b8* are present at 415.398, 571.565, 727.680, and 1,096.718 Our biochemical studies indicated that both Mtb and Msm Eis BIOCHEMISTRY Da, respectively. This result identifies Lys55 of the MKP-7(53−66) proteins could readily acetylate aminoglycosides, with Msm Eis ɛ peptide as the N -acetylation site. By analogy with DUSP1/MKP-1 being a marginally better catalyst as an aminoglycoside N-acetyl- ɛ (22), we suggest that N -acetylation of DUSP16/MKP-7 at Lys55 transferase (Fig. S2). In contrast, their substrate preferences by Mtb Eis may increase the interactions between DUSP16/MKP-7 against MKP-7(53−66) and MKP-1(50−63) peptides were differ- and JNK by neutralizing the positive charge within the docking ent. Mtb Eis acetylated an internal lysine (Lys55) of the MKP-7 domain of DUSP16/MKP-7. Msm Eis acetylated both MKP-7(53 (53−66) peptide rapidly, whereas Msm Eis preferentially acety- −66) and MKP-1(50−63) peptides very quickly in approximately 5 lated the terminal amino group of these peptides (Fig. 1 and Fig. min, resulting in a mass increase by 42 Da (Fig. S6). We have S6). To understand the observed difference in peptide acetylation confirmed by MALDI tandem mass spectrometry that these pep- site preferences, we have compared the active site features of these α tides are N -acetylated at the amino terminus by Msm Eis (Fig. S6). Eis proteins. Both Eis active sites show negative electrostatic

Fig. 1. Acetyltransferase assay of Mtb Eis using MKP-7(53−66) and MKP-1(50−63) peptides by mass spectrometry. (A) Mass spectra of the MKP-7(53−66) peptide before and after acetylation reaction by Mtb Eis (Upper). The observed increase in the peptide mass by 42 Da indicates that the peptide is acetylated at a single site. Mass spectra of the MKP-1(50−63) peptide before and after acetylation reaction by Mtb Eis (Lower) indicate that this peptide is not acetylated by Mtb Eis. (B) MALDI MS/MS spectrum of the MKP-7(53−66) peptide acetylated by Mtb Eis. The fragments marked with an asterisk (*) are +42 Da-shifted ions, compared with the counterparts that would be generated from the unmodiﬁed peptide. The acetylated MKP-7(53−66) fragmentation notation using the scheme of Roepstorff and Fohlman (45) is given above the spectrum. The acetyl group of modiﬁed Lys55 is highlighted by enclosing in a red box.

Kim et al. PNAS Early Edition | 3of6 Downloaded by guest on September 27, 2021 contact between α2 and α6 helices in Msm Eis is stabilized by hydrogen bondings of Gln33-Arg200 and Thr34-Asp195 pairs (Fig. S7A). In addition, the side chain of Trp42 in Mtb Eis is located between the side chains of Trp19 (on α1) and Phe90 (on β4) in “structural” domain 1, whereas the corresponding residue in Msm Eis, Trp38, is located on the other side of Phe84 and makes van der Waals contacts with the side chains of Met41 (on α2), Leu192 (on α6), and Tyr400 (Fig. S7A). The elongated substrate- binding channel in Mtb Eis seems to be suitable not only for accommodating aminoglycosides but also for recognizing the poly- peptide substrate in a sequence-specific manner. This channel is Fig. 2. Acetyltransferase activity assay of Mtb Eis and Msm Eis using the also present in the reported structure of Mtb Eis (20). The deep, recombinant human DUSP/MKP-7(1−153) and DUSP/MKP-7(1−303) proteins round-shaped substrate-binding pocket in Msm Eis seems more as potential substrates. (A) Time-course acetyltransferase activity assay of Mtb suitable for accommodating aminoglycosides and the terminal Eis and a comparison of Mtb Eis and Msm Eis activities. The wild-type DUSP16/ amino group of peptides than sequence-specific recognition − − 14 MKP-7(1 153) and DUSP/MKP-7(1 303) were incubated with [ C]-labeled of polypeptides. acetyl CoA and Eis for the indicated duration at 37 °C. The reaction products were separated by 15% (wt/vol) SDS/PAGE, and the acetylated protein bands Mtb Eis, but Not Msm Eis, Inhibits LPS-Induced JNK Phosphorylation were visualized using a Bioimage analyzer. (B) Acetyltransferase activity of fl Mtb Eis and Msm Eis toward the wild-type and mutants of DUSP16/MKP-7(1− and Proin ammatory Cytokine Production. Because DUSP16/MKP- 153). All reactions were carried out at 37 °C for 30 min. 7 was reported to inhibit JNK activation in macrophages in response to LPS by dephosphorylating (24–28), we examined whether Mtb Eis or Msm Eis affects LPS-induced JNK phosphorylation in bone potential surfaces (Fig. 3), and key residues around the thiol group marrow-derived macrophages (BMDMs). Mtb Eis significantly re- of acetyl CoA (or CoA) are identical (Asp32/Tyr132/Phe408 in duced the level of phosphorylated JNK in BMDMs in response to Mtb; Asp28/Tyr126/Phe402 in Msm). Interestingly, we noticed LPS, whereas Msm Eis did not (Fig. 4). Furthermore, Mtb Eis sig- a difference in the shape of the predicted substrate binding sites nificantly inhibited production of TNF-α and IL-6 in BMDMs in adjacent to the bound ligand in their active sites. Mtb Eis has response to LPS, whereas Msm Eis did not (Fig. 4). Inhibition of a deep and narrow channel, whereas Msm Eis has a deep and JNK phosphorylation was reported to reduce production of these round pocket (Fig. 3). This structural difference in the substrate proinflammatory cytokines (29, 30). These data, together with our binding sites is mainly due to different arrangements of helix α2 finding that Mtb Eis acetylates DUSP16/MKP-7 at Lys55, indicate (Gly35–Leu45 in Mtb Eis; Glu32–Met41 in Msm), part of “struc- that Mtb Eis suppresses host innate immune responses through tural” domain 1, and helix α6 (Gln200–Cys210 in Mtb Eis; inactivation of JNK via acetylation of DUSP16/MKP-7. Asp190–Ala198 in Msm) and the following loop, which are part of “structural” domain 2 (Fig. S7A). The narrow channel of Mtb Eis is Discussion formed by limited hydrophobic interactions between these helices. Survival of Mtb inside human macrophages is central to tubercu- In comparison, the round pocket-shaped substrate binding site of losis infection, latency, disease activation, and transmission (31). Msm Eis is formed by extensive interactions between helix α2in The most important survival strategies of the pathogen are the “structural” domain 1 and helix α6 followed by an additional 310 inhibition of phagosomal maturation and autophagy (32). helix α6’ in “structural” domain 2 (Figs. S1 and S7A). The 310 helix Autophagy has been recognized as innate and adaptive immune α6’ is present in Msm Eis only (Figs. S1 and S7A). The close defense mechanisms (32). A functional overlap was suggested

Fig. 3. Comparison of Mtb and Msm Eis monomers. (A) Elec- trostatic potential at the surface of Mtb Eis monomer, with an enlarged view from an orthogonal angle. (B) Electrostatic potential at the surface of Msm Eis monomer, with an enlarged view from an orthogonal angle. Blue and red correspond to positive and negative potentials, respectively. Red circles indicate possible substrate binding sites of Mtb and Msm Eis.

4of6 | www.pnas.org/cgi/doi/10.1073/pnas.1120251109 Kim et al. Downloaded by guest on September 27, 2021 Fig. 4. Effects of Mtb Eis on JNK activation and cytokine production in BMDM cells. (A) Mtb Eis, but not Msm Eis, suppresses JNK activation in BMDM cells upon LPS stimulation. Cells were treated with or without Mtb Eis (or Msm Eis) (5, 10, or 20 μgmL−1) for 1 h, followed − by stimulation with LPS (100 ng mL 1)for30 min. Cells were then harvested, lysed, and subjected to Western blot analysis using anti- bodies raised to phospho-JNK and β-actin. Data shown are representative of three independent experiments that all yielded similar results. Ex- pression of phospho-JNK and β-actin in cyto- plasmic extracts of BMDMs was quantiﬁed densitometrically (Right). Data represent the mean ± SD of three independent experiments. ***P < 0.001 vs. LPS-stimulated condition. U, LPS-untreated; L, LPS-treated condition without Mtb Eis or Msm Eis. (B) Mtb Eis, but not Msm Eis, suppresses proinﬂammatory cytokine production in BMDM cells upon LPS stimulation. Cells were treated with or without Mtb Eis (or − Msm Eis) (5, 10, or 20 μgmL 1) for 1 h, followed − by stimulation with LPS (100 ng mL 1) for 18 h. Supernatants were harvested, and the levels of TNF-α andIL-6weremeasuredbyELISA.

between phagosome maturation and autophagy, both of which docking domain of DUSP16/MKP-7. We have also shown that depend on Beclin 1 and VPS34 as key players (32). Beclin 1, a pro- Mtb Eis, but not Msm Eis, significantly down-regulated the LPS- autophagy BH3 domain-containing protein, plays a central role in induced JNK phosphorylation. On the basis of these findings, we BIOCHEMISTRY autophagosome formation by interacting with several other factors propose that acetylation of DUSP16/MKP-7 by Mtb Eis is the key to promote the formation of Beclin 1–VPS34–VPS15 core com- initial event in the JNK-dependent inhibition of autophagy, plexes (12, 33). In mammalian cells under nonstarvation con- phagosome maturation, and ROS generation, which ultimately ditions, the antiapoptotic protein Bcl-2 binds to Beclin 1 and contributes to enhanced survival of Mtb within the macrophage inhibits its autophagy function (34). Starvation induces Bcl-2 dis- cells. Definitely establishing the proposed mechanism of en- sociation from Beclin 1, via phosphorylation of Bcl-2, and hanced intracellular survival would require functional studies with autophagy activation. Furthermore, JNK1, but not JNK2, was an engineered pathogen in a suitable host cell. found to mediate starvation-induced Bcl-2 phosphorylation (34). It has been well established that protein lysine acetylation Recently, JNK signaling was shown to mediate amplification of critically regulates gene transcription by targeting histones as well ROS production during multiple stresses, including infection (35). as a variety of transcription factors in the nucleus (40). Numerous Cellular stress altered mitochondria, causing JNK to translocate to proteins located outside the nucleus have also been demonstrated the mitochondria and to amplify up to 80% of the ROS generated to be acetylated (40). Indeed, protein lysine acetylation is largely by complex I. ROS activates JNK via a sequence of events emerging as a major mechanism by which key proteins are regu- for JNK mitochondrial signaling (35). There exists a molecular lated in many physiological processes (40). Recent reports also cross-talk between autophagy and apoptosis, because Beclin 1 can link lysine acetylation to heterochromatin assembly, sister chro- be cleaved by caspases, and its proautophagic activity is lost (36). matid cohesion, cytoskeleton dynamics, autophagy, receptor sig- Moreover, the resulting C-terminal fragment of Beclin 1 could naling, RNA processing, and metabolic control (41). Proteomics amplify mitochondrion-mediated apoptosis (36, 37). studies indicate that the complexity of the acetylome potentially The kinase activity of MAPKs such as JNK is negatively regu- rivals that of the phosphoproteome (42). Therefore, it is not lated by DUSPs (also called MKPs). DUSP16/MKP-7 was found surprising to find that Mtb uses Eis to acetylate the host signaling to work as a JNK-specific phosphatase in vivo, because forced protein DUSP16/MKP-7 in suppressing immune responses for its expression of DUSP16/MKP-7 suppressed activation of MAPKs in survival in macrophages. Mtb might have evolved in such as way the order of selectivity, JNK >> p38 > ERK (17). When expressed that its eis gene product has retained much of the aminoglycoside in mammalian cells, DUSP16/MKP-7 was localized exclusively in N-acetyltransferase activity, whereas it has gained a significantly ɛ the cytoplasm (17). Acetylation of the components of MAPK higher protein lysine N -acetyltransferase activity to disrupt the pathways on serine/threonine and lysine residues was previously cellular signaling pathway for intracellular survival. reported to serve as a regulatory mechanism in biological signal- Similarly to Mtb Eis, protein kinase G (PknG) of Mtb inhibits ing. Yersinia YopJ was shown to act as an acetyltransferase to phagosome–lysosome fusion and mediates intracellular survival of modify serine and threonine residues in the activation loop of mycobacteria by disrupting the host cellular signaling (43). PknG is MAPK kinase-6, thereby blocking phosphorylation and sub- one of the 11 eukaryotic-like Ser/Thr protein kinases encoded by the sequent activation of the kinase activity (38, 39). A nuclear- Mtb genome and is secreted within macrophages. Mtb Eis, like Mtb localized DUSP1/MKP-1 is acetylated by p300 on Lys57 within its PknG, could be an excellent target for the development of drugs that docking domain, resulting in deactivation of Toll-like receptor induce mycobacterial death inside macrophages. An advantage of inflammatory signaling and inhibition of innate immune signaling targeting Eis or PknG is that it does not kill the bacteria per se but (22). p300 contains a histone acetyltransferase domain. Acetyla- instead facilitates the macrophage to carry out its natural antibac- tion of DUSP1/MKP-1 enhanced its interaction with p38, thereby terial activity, delivering intracellularly surviving mycobacteria to increased its phosphatase activity, and interrupted MAPK sig- lysosomes for destruction (44). Another potential advantage of naling cascade (22, 23). In this study, we have discovered that Mtb targeting a secreted protein such as Eis or PknG is that its inhibitors ɛ Eis acts as an N -acetyltransferase to acetylate Lys55 within the are not required to be transported through the extremely

Kim et al. PNAS Early Edition | 5of6 Downloaded by guest on September 27, 2021 impermeable mycobacterial cell wall. This may greatly improve the The acetylation site of the in vitro acetylated peptides was identified by bactericidal activity of the compounds (44). Mtb Eis has an extra MALDI MS/MS analysis. advantage, because existing aminoglycoside drugs, such as kana- ACKNOWLEDGMENTS. We thank the beamline staff at Photon Factory, Japan mycin and amikacin, can be made effective by inhibiting its amino- (BL-5A, BL-17A, NW12), SPring-8, Japan (BL38B1), and Pohang Light Source, glycoside N-acetyltransferase activity. Our structural information Korea (BL-4A) for assistance during X-ray diffraction experiments; Prof. Seung would be useful in structure-based discovery of peptidomimetic or Bum Park for the use of the Synergy HT Multi-Mode Microplate Reader small-molecule inhibitors that target the active site of Mtb Eis. (BioTek Instruments); and Dong-Min Shin and Jin Kyung Kim (both at Chungnam National University) for excellent technical assistance. This work was funded by the Korea Ministry of Education, Science, and Technology Materials and Methods (MEST), National Research Foundation (NRF) of Korea, Korea-New Zealand Detailed methods of protein expression/purification, in vitro acetylation Cooperative Research Grant, Basic Science Outstanding Scholars Program assay, mass spectrometry, crystallization, X-ray data collection, structure (2008-0093867), World-Class University Program (Grant R31-10032), Grant determination, cell culture, Western blotting, and ELISA are provided in SI R11-2007-107-00000-0 from the Innovative Drug Research Center for Meta- bolic and Inflammatory Disease, and by Grant A092006 from the Korea Materials and Methods. Briefly, the genes encoding Mtb and Msm Eis pro- Ministry of Health, Welfare and Family Affairs, Korea Healthcare Technology teins were cloned into pET-28b(+) and were expressed in E. coli Rosetta II R&D Project (S.W.S.); and by Korea MEST NRF Grant 2012-0005763 through (DE3)pLysS cells. The recombinant Eis proteins with an N-terminal fusion tag the Infection Signaling Network Research Center at Chungnam National Uni- were purified, crystallized, and characterized by in vitro acetylation assays. versity (to E.-K.J.).

1. Dye C, Williams BG (2010) The population dynamics and control of tuberculosis. Sci- 23. Chi H, Flavell RA (2008) Acetylation of MKP-1 and the control of inflammation. Sci ence 328:856–861. Signal 1:pe44. 2. Chiang CY, Centis R, Migliori GB (2010) Drug-resistant tuberculosis: Past, present, 24. Tanoue T, Yamamoto T, Maeda R, Nishida E (2001) A Novel MAPK phosphatase MKP- future. Respirology 15:413–432. 7 acts preferentially on JNK/SAPK and p38 alpha and beta MAPKs. J Biol Chem 276: 3. Dahl JL, Wei J, Moulder JW, Laal S, Friedman RL (2001) Subcellular localization of the 26629–26639. Iitracellular survival-enhancing Eis protein of Mycobacterium tuberculosis. Infect Im- 25. Han SY, Kim SH, Heasley LE (2002) Differential gene regulation by specific gain-of- mun 69:4295–4302. function JNK1 proteins expressed in Swiss 3T3 fibroblasts. J Biol Chem 277: 4. Lin PL, Flynn JL (2010) Understanding latent tuberculosis: A moving target. J Immunol 47167–47174. 185:15–22. 26. Matsuguchi T, Musikacharoen T, Johnson TR, Kraft AS, Yoshikai Y (2001) A novel 5. Getahun H, Gunneberg C, Granich R, Nunn P (2010) HIV infection-associated tuber- mitogen-activated protein kinase phosphatase is an important negative regulator of – culosis: The epidemiology and the response. Clin Infect Dis 50(Suppl 3):S201 S207. lipopolysaccharide-mediated c-Jun N-terminal kinase activation in mouse macro- 6. Jo EK, Yang CS, Choi CH, Harding CV (2007) Intracellular signalling cascades regu- phage cell lines. Mol Cell Biol 21:6999–7009. lating innate immune responses to Mycobacteria: Branching out from Toll-like re- 27. Weston CR, Davis RJ (2007) The JNK signal transduction pathway. Curr Opin Cell Biol – ceptors. Cell Microbiol 9:1087 1098. 19:142–149. fi 7. Wei J, et al. (2000) Identi cation of a Mycobacterium tuberculosis gene that enhances 28. Kondoh K, Nishida E (2007) Regulation of MAP kinases by MAP kinase phosphatases. – mycobacterial survival in macrophages. J Bacteriol 182:377 384. Biochim Biophys Acta 1773:1227–1237. 8. Meena LS, Rajni (2010) Survival mechanisms of pathogenic Mycobacterium tubercu- 29. Assi K, Pillai R, Gómez-Muñoz A, Owen D, Salh B (2006) The specific JNK inhibitor – losis H37Rv. FEBS J 277:2416 2427. SP600125 targets tumour necrosis factor-α production and epithelial cell apoptosis in 9. Ehrt S, Schnappinger D (2009) Mycobacterial survival strategies in the phagosome: acute murine colitis. Immunology 118:112–121. – Defence against host stresses. Cell Microbiol 11:1170 1178. 30. Jang S, Kelley KW, Johnson RW (2008) Luteolin reduces IL-6 production in microglia 10. Vergne I, Chua J, Singh SB, Deretic V (2004) Cell biology of Mycobacterium tuber- by inhibiting JNK phosphorylation and activation of AP-1. Proc Natl Acad Sci USA 105: culosis phagosome. Annu Rev Cell Dev Biol 20:367–394. 7534–7539. 11. Vergne I, et al. (2006) Autophagy in immune defense against Mycobacterium tu- 31. Russell DG (2001) Mycobacterium tuberculosis: Here today, and here tomorrow. Nat berculosis. Autophagy 2:175–178. Rev Mol Cell Biol 2:569–577. 12. Kang R, Zeh HJ, Lotze MT, Tang D (2011) The Beclin 1 network regulates autophagy 32. Gutierrez MG, et al. (2004) Autophagy is a defense mechanism inhibiting BCG and and apoptosis. Cell Death Differ 18:571–580. Mycobacterium tuberculosis survival in infected macrophages. Cell 119:753–766. 13. Wu S, et al. (2009) Activation of the eis gene in a W-Beijing strain of Mycobacterium 33. Funderburk SF, Wang QJ, Yue Z (2010) The Beclin 1-VPS34 complex—at the crossroads tuberculosis correlates with increased SigA levels and enhanced intracellular growth. of autophagy and beyond. Trends Cell Biol 20:355–362. Microbiology 155:1272–1281. 34. Wei Y, Pattingre S, Sinha S, Bassik M, Levine B (2008) JNK1-mediated phosphorylation 14. Lella RK, Sharma C (2007) Eis (enhanced intracellular survival) protein of Mycobac- of Bcl-2 regulates starvation-induced autophagy. Mol Cell 30:678–688. terium tuberculosis disturbs the cross regulation of T-cells. J Biol Chem 282: 35. Chambers JW, LoGrasso PV (2011) Mitochondrial c-Jun N-terminal kinase (JNK) sig- 18671–18675. naling initiates physiological changes resulting in amplification of reactive oxygen 15. Samuel LP, et al. (2007) Expression, production and release of the Eis protein by species generation. J Biol Chem 286:16052–16062. Mycobacterium tuberculosis during infection of macrophages and its effect on cy- 36. Djavaheri-Mergny M, Maiuri MC, Kroemer G (2010) Cross talk between apoptosis and tokine secretion. Microbiology 153:529–540. – 16. Shin DM, et al. (2010) Mycobacterium tuberculosis eis regulates autophagy, in- autophagy by caspase-mediated cleavage of Beclin 1. Oncogene 29:1717 1719. flammation, and cell death through redox-dependent signaling. PLoS Pathog 6: 37. Wirawan E, et al. (2010) Caspase-mediated cleavage of Beclin-1 inactivates Beclin-1- e1001230. induced autophagy and enhances apoptosis by promoting the release of proa- 17. Masuda K, Shima H, Watanabe M, Kikuchi K (2001) MKP-7, a novel mitogen-activated poptotic factors from mitochondria. Cell Death Dis 1:e18. protein kinase phosphatase, functions as a shuttle protein. J Biol Chem 276: 38. Mukherjee S, et al. (2006) Yersinia YopJ acetylates and inhibits kinase activation by – 39002–39011. blocking phosphorylation. Science 312:1211 1214. fi 18. Neuwald AF, Landsman D (1997) GCN5-related histone N-acetyltransferases belong to 39. Mukherjee S, Hao YH, Orth K (2007) A newly discovered post-translational modi - — a diverse superfamily that includes the yeast SPT10 protein. Trends Biochem Sci 22: cation the acetylation of serine and threonine residues. Trends Biochem Sci 32: – 154–155. 210 216. 19. Zaunbrecher MA, Sikes RD, Jr., Metchock B, Shinnick TM, Posey JE (2009) Over- 40. Close P, et al. (2010) The emerging role of lysine acetylation of non-nuclear proteins. expression of the chromosomally encoded aminoglycoside acetyltransferase eis con- Cell Mol Life Sci 67:1255–1264. fers kanamycin resistance in Mycobacterium tuberculosis. Proc Natl Acad Sci USA 106: 41. Kim GW, Yang XJ (2011) Comprehensive lysine acetylomes emerging from bacteria to 20004–20009. humans. Trends Biochem Sci 36:211–220. 20. Chen W, Biswas T, Porter VR, Tsodikov OV, Garneau-Tsodikova S (2011) Unusual re- 42. Norris KL, Lee JY, Yao TP (2009) Acetylation goes global: The emergence of acetyla- gioversatility of acetyltransferase Eis, a cause of drug resistance in XDR-TB. Proc Natl tion biology. Sci Signal 2:pe76. Acad Sci USA 108:9804–9808. 43. Walburger A, et al. (2004) Protein kinase G from pathogenic mycobacteria promotes 21. Crossman DK (2007) Characterization of a novel acetylatransferase found only in survival within macrophages. Science 304:1800–1804. pathogenic strains of Mycobacterium tuberculosis. PhD thesis (Univ of Alabama, 44. Nguyen L, Pieters J (2009) Mycobacterial subversion of chemotherapeutic reagents Tuscaloosa, AL). and host defense tactics: Challenges in tuberculosis drug development. Annu Rev 22. Cao W, Bao C, Padalko E, Lowenstein CJ (2008) Acetylation of mitogen-activated Pharmacol Toxicol 49:427–453. protein kinase phosphatase-1 inhibits Toll-like receptor signaling. J Exp Med 205: 45. Roepstorff P, Fohlman J (1984) Proposal for a common nomenclature for sequence 1491–1503. ions in mass spectra of peptides. Biomed Mass Spectrom 11:601.

6of6 | www.pnas.org/cgi/doi/10.1073/pnas.1120251109 Kim et al. Downloaded by guest on September 27, 2021