Chemical modulation of SQSTM1/p62-mediated xenophagy that targets a broad range of pathogenic bacteria

ABSTRACT The N-degron pathway is a proteolytic system in which the N-terminal degrons (N-degrons) of proteins, such as arginine (Nt-Arg), induce the degradation of proteins and subcellular organelles via the ubiquitin-proteasome system (UPS) or macroautophagy/autophagy-lysosome system (hereafter autophagy). Here, we developed the chemical mimics of the N-degron Nt-Arg as a pharmaceutical means to induce targeted degradation of intracellular bacteria via autophagy, such as Salmonella enterica serovar Typhimurium (S. Typhimurium), Escherichia coli, and Streptococcus pyogenes as well as Mycobacterium tuberculosis (Mtb). Upon binding the ZZ domain of the autophagic cargo receptor SQSTM1/p62 (sequestosome 1), these chemicals induced the biogenesis and recruitment of autophagic membranes to intracellular bacteria via SQSTM1, leading to lysosomal degradation. The antimicrobial efficacy was independent of rapamycin-modulated core autophagic pathways and synergistic with the reduced production of inflammatory cytokines. In mice, these drugs exhibited antimicrobial efficacy for S. Typhimurium, Bacillus Calmette–Guérin (BCG), and Mtb as well as multidrug-resistant Mtb and inhibited the production of inflammatory cytokines. This dual mode of action in xenophagy and inflammation significantly protected mice from inflammatory lesions in the lungs and other tissues caused by all the tested bacterial strains. Our results suggest that the N-degron pathway provides a therapeutic target in host-directed therapeutics for a broad range of drug-resistant intracellular pathogens. Abbreviations: ATG: autophagy-related gene; BCG: Bacillus Calmette–Guérin; BMDMs: bone marrow-derived macrophages; CALCOCO2/NDP52: calcium binding and coiled-coil domain 2; CFUs: colony-forming units; CXCL: C-X-C motif chemokine ligand; EGFP: enhanced green fluorescent protein; IL1B/IL-1β: interleukin 1 beta; IL6: interleukin 6; LIR: MAP1LC3/LC3-interacting region; MAP1LC3/LC3: microtubule associated protein 1 light chain 3; Mtb: Mycobacterium tuberculosis; MTOR: mechanistic target of rapamycin kinase; NBR1: NBR1 autophagy cargo receptor; OPTN: optineurin; PB1: Phox and Bem1; SQSTM1/p62: sequestosome 1; S. Typhimurium: Salmonella enterica serovar Typhimurium; TAX1BP1: Tax1 binding protein 1; TNF: tumor necrosis factor; UBA: ubiquitin-associated.

Upon infection, intracellular bacteria reside within phagosomes to escape from the surveillance of host defense systems [18][19][20][21][22][23][24]. A subpopulation of bacteria is exposed to the cytosol during which the secretory system on bacterial membranes secretes bacterial effector proteins to modulate the functions of host cells [23,[25][26][27]. Once cells sense bacteria in the cytosol, specific proteins on bacterial membranes are assembled with ubiquitin chains for targeting to autophagic membranes and lysosomal degradation [28,29]. Several E3 ligases were identified to ubiquitinate specific receptors on bacterial membranes, including PRKN/parkin that mediates K63-linked ubiquitination to target Mycobacterium tuberculosis (Mtb) to autophagic membranes via SQSTM1 and CALCOCO2 [30] and SMURF1 that mediates K48-linked ubiquitination for Mtb and Listeria monocytogenes [31]. In addition, the ubiquitination of Salmonella enterica serovar Typhimurium (S. Typhimurium) involves the cooperative activities of LRSAM1, ARIH1, and RNF31/HOIP (ring finger protein 31) [32][33][34]. Each type of bacteria may be marked by distinct ubiquitin codes, which are defined by the biochemical specificity of E3 enzyme [34,35]. Ubiquitinated bacteria are recognized by autophagy receptors to be targeted to autophagosome for degradation [36].
Various pathogens have evolved unique strategies to avoid, subvert, or co-opt host defense systems to aid their survival [8]. To inhibit bacterial infection, several generations of antibiotics have been developed, contributing to reducing disease incidence caused by bacterial pathogens [37,38]. However, as the incidences of multidrug resistance of pathogenic bacteria rise and threaten the ability to treat bacterial infection, it is increasingly urgent to develop fundamentally different approaches [39]. In principle, there are several antimicrobial strategies such as phage therapy, vaccines, antibodies, probiotics, antimicrobial peptides, and host-directed therapy [38][39][40][41]. Host-directed therapy utilizes small molecule drugs or proteins to augment host defense mechanisms or modulate excessive inflammation [42]. Among host-directed therapies, the pharmaceutical modulation of autophagy is emerging as a universal means to eradicate intracellular bacteria using lysosomal hydrolases [42,43]. To develop xenophagy-inducing drugs, several chemicals were identified to inhibit the infectivity of intracellular bacteria by activating autophagy. Rapamycin enhances colocalization of Mtb with LC3 and acidification of mycobacterial phagosomes in vitro at 50 μM 2 h after infection in a manner independent of autophagy [44,45]. In another study, however, rapamycin facilitated the infectivity of Mtb when treated at 1 μM for 3 days in vitro [44]. The antidiabetic drug metformin upregulates lipidated LC3 form and shows an antimicrobial effect in vitro at 1 mM and in mice at 500 mg/kg (q.o.d. injection, intraperitoneal [i.p.]) [46]. The AMPK activator AICAR inhibits the infectivity of Mtb in vitro at 50-100 μM and in mice at 500 mg/kg i.p. by activating autophagic pathways [47]. Resveratrol, a SIRT1 (sirtuin 1) activator, shows moderate efficacy against Mtb in vitro at 10-100 μM [48]. AR-12, triclosan and D61 also exhibit antimicrobial effects against S. Typhimurium in vitro at various concentrations [49,50]. To date, there are no known xenophagy-inducing drugs with satisfactory efficacy via a clear mode of action.
The N-degron pathway is a proteolytic system that targets proteins harboring destabilizing N-terminal (Nt) residues, called N-degrons [51][52][53]. The N-degrons such as N-Arg are recognized by the N-recognins UBR1, UBR2, UBR4, and UBR5 that mediate substrate ubiquitination, leading to proteasomal degradation via the ubiquitinproteasome system (UPS) [54,55]. Recently, we have shown that the N-degron pathway regulates autophagic proteolysis, in which the Nt-Arg is post-translationally conjugated on endoplasmic reticulum (ER)-residing molecular chaperones [56][57][58][59]. This autophagic N-degron binds the ZZ domain of the autophagic receptor SQSTM1/p62 and promotes the self-polymerization of SQSTM1 in complex with protein cargoes and SQSTM1 interaction with LC3 on phagophores, leading to lysosomal degradation. We also developed chemical mimics of the autophagic N-degron as synthetic agonists to the ZZ domain of SQSTM1 [58,59]. These agonists were shown to induce in vitro lysosomal degradation of misfolded protein aggregates in neurodegeneration [58] and the autophagic turnover of the ER [59].
In this study, we show that invading bacteria and SQSTM1-dependent selective autophagy mutually counteract each other and developed xenophagy-inducing drugs (XIDs) targeting SQSTM1. These drugs exhibited antimicrobial efficacy against various infectious bacteria by inducing MTORindependent selective autophagy. Pharmaceutical activation of SQSTM1 facilitated autophagosome biogenesis and SQSTM1 association with bacteria, which in turn recruited autophagic membranes to bacterial membranes, leading to lysosomal degradation. In mice, these xenophagy modulators exhibited strong antimicrobial efficacy for S. Typhimurium, Mtb, and even multidrug-resistant Mtb and suppressed excessive production of inflammatory cytokines. We suggest that these XIDs may be developed into drugs for a broad spectrum of pathogenic bacteria.

Counteractive crosstalk between S. Typhimurium and host cells via autophagy
Several studies have characterized autophagic pathways in cells infected with bacteria [19], yet the role of SQSTM1dependent selective autophagy as a host defense system remains poorly understood. As an initial step to explore autophagy as a target for antimicrobial drugs, we assessed autophagic flux in RAW264.7 and HeLa cells infected with the Gram-negative bacteria S. Typhimurium. Immunoblotting analyses showed that the synthesis of LC3 and its lipidation into LC3-II ( Figure 1A) were strikingly suppressed in a manner depending on multiplicity of infection (MOI) ( Figure 1B and S1A) and time during 1-6 h post-infection ( Figure 1C). Such an activity of LC3 was tightly blocked at a basal level as evidenced by insensitivity to bafilomycin A 1 , an inhibitor of the lysosomal V-ATPase ( Figure 1D). In contrast to bacteria, host cells could robustly induce the synthesis and lipidation of LC3 ( Figure 1E) and the formation of LC3-positive puncta ( Figure 1F) in response to lipopolysaccharide (LPS), a pathogen-associated molecular pattern (PAMP). In addition, basal autophagy flux of host cells was dramatically reduced when S. Typhimurium was infected ( Figure 1G). These results suggest that the suppression of autophagic pathways is essential for the life cycle of S. Typhimurium.
Next, we further monitored the remaining autophagic activities in cells infected with S. Typhimurium. Quantitative realtime PCR (RT-qPCR) showed increased transcription of auto-phagy-related genes such as ATG9B, WIPI1, TMEM74, ULK2, and DRAM1 ( Figure 1H). When visualized using immunostaining analyses, bacterial membranes were associated with LC3positive autophagic membranes ( Figure S1B) as well as various autophagic receptors such as SQSTM1, NBR1, OPTN, and CALCOCO2 with distinct spatiotemporal patterns ( Figure 1I). Consistently, the intracellular growth of S. Typhimurium was markedly facilitated by knockdown of SQSTM1 ( Figure 1J), LC3B, or ATG5 ( Figure 1K). Moreover, the level of SQSTM1 and OPTN increased upon bacterial infection as compared with TAX1BP1, and CALCOCO2 ( Figure 1L). A similar induction was observed with mRNA expression of autophagy receptors including SQSTM1 and OPTN ( Figure 1M). These results suggest that the infected cells retain residual autophagic activities to degrade intracellular bacteria, providing SQSTM1-dependent selective autophagy as a drug target against a broad range of pathogenic bacteria.

Development of small molecule agonists that exert antimicrobial efficacy
We have previously developed two small molecule ligands, YTK-1105 and YOK-1104, to the ZZ domain of SQSTM1 that induce selective autophagy ( Figure S2A) [58,59]. To develop antimicrobial agents with high selectivity and efficacy, we synthesized approximately 120 derivatives of these initial compounds by employing 3D structure modeling of the SQSTM1 ZZ domain associated with SAR (structure-activity relationship). The optimization processes first focused on the activity to induce selfoligomerization of SQSTM1 and to induce the synthesis and lipidation of LC3. Following the initial screening, we assessed the antimicrobial efficacy of the resulting derivatives in the macrophage RAW264.7 and HeLa cells infected with S. Typhimurium by using colony-forming unit (CFU) assays. This screening yielded five compounds (YTK-A76, YT-6-2, YOK-1204, YTK-2205, and YOK-1109) with high antimicrobial efficacy in cultured cells (Figure 2A and S2B-D). In CFU assays of RAW264.7 cells, these ligands exhibited antimicrobial efficacy in a manner depending on the dose ( Figure 2B), time, and MOI ( Figure 2C,D). The efficacy was reproduced in various cell lines such as J774A.1, THP-1, and BMDMs macrophage cell ( Figure  S2E) as well as the HCT116 and HeLa epithelial cancer cells ( Figure 2E).
To rule out the possibility that SQSTM1 agonists directly kill intracellular bacteria, we performed disk diffusion assays ( Figure S2F). As expected, antibiotics such as ampicillin, kanamycin, and gentamicin effectively killed S. Typhimurium, and the zone of inhibition became larger in a dose-dependent manner. In contrast, SQSTM1 agonists did not form such a zone of inhibition ( Figure S2F). Moreover, bacteria normally grew in the growth medium containing as high as 15 μM of SQSTM1 agonists ( Figure 2F). During these experiments, viability of host cells also was not significantly affected at the concentrations used for CFU assays ( Figure 2G and S2G). Finally, to rule out the possibility that SQSTM1 agonists affect bacterial entry, we measured the number of intracellular bacteria at 1 hpi following pretreatment of SQSTM1 agonists. No difference was observed ( Figure S2H). These results confirm that SQSTM1 agonists exert antimicrobial efficacy by modulating the function of host cells.
In the docking simulation study, the identified compounds located in the position where the first and second residues of N-degron bound ( Figure 2H). The binding mode of YT-6-2 showed a hydrogen bond between the oxygen attached to chiral center of YT-6-2 and the side chain of Asp147, which we speculated that play an important role in the binding. To determine the direct binding of the SQSTM1 agonist to SQSTM1 ZZ domain, we performed pulldown assays using biotinylated YT-6-2 and full-length SQSTM1 ( Figure S2I). The result showed that biotinylated YT-6-2 bound wild-type SQSTM1 and that the binding was significantly reduced when a point mutation was introduced within ZZ domain (D147K) ( Figure 2I). To further support the selectivity of SQSTM1 agonists, we synthesized ATB1095 which lacks the essential amine group required for hydrogen bond with the side chain of Asp147 ( Figure S2J). As expected, this compound showed no antimicrobial effect ( Figure S2K). These results suggest that SQSTM1 agonists exert their antimicrobial efficacy through specific hydrogen bonds to the ZZ domain of SQSTM1.

SQSTM1 agonists rescue autophagic activities from suppression by S. Typhimurium
To characterize the mode of action of SQSTM1 agonists, we assessed their activities to induce degradative flux via SQSTM1-dependent selective autophagy. Immunoblotting analyses showed that YTK-A76, YT-6-2, YOK-1204, YTK-2205 and YOK-1109 enhanced the synthesis and lipidation of LC3 via non-canonical pathway ( Figure 3A, S3A-S3C) associated with the increases in LC3 + autophagic membranes, most of which colocalized with SQSTM1 + puncta ( Figure 3B and S3D). Cells treated with SQSTM1 agonists produced excessive lipidated LC3 which was attributed to the activation of autophagic flux ( Figure 3C). As an alternative assay, we also monitored the ratio of autophagosomes and autolysosomes in the macrophage BMDMs using a retrovirus expressing the mCherry-EGFP-LC3B fusion. The treatment of SQSTM1 agonists markedly increased in the number of autolysosomes (mCherry + EGFP − , red) ( Figure S3E). When the cells were co-treated with SQSTM1 agonists and bafilomycin A 1 , autophagosomes (mCherry + EGFP + , yellow) failed to turn into autolysosomes ( Figure S3E). These results demonstrate that SQSTM1 agonists enhance the degradative flux of autophagic cargoes.
Next, we determined whether SQSTM1 agonists are capable to rescue autophagic flux from suppression by S. Typhimurium. Indeed, SQSTM1 agonists restored the synthesis and lipidation of LC3 to normal levels in HeLa cells ( Figure 3D) and BMDMs ( Figure 3E) infected with S. Typhimurium. Autophagic flux assays using NH 4 Cl showed that the LC3 induction is attributed to the activation of autophagic flux ( Figure 3F). In sharp contrast, rapamycin, an inducer of MTOR (mammalian target of rapamycin)-modulated core autophagic pathways, exhibited no such efficacy ( Figure 3G). These results suggest that SQSTM1 agonists exert antimicrobial efficacy via a non-canonical pathway and independent of rapamycin-modulated bulk autophagy.

SQSTM1 agonists exert antimicrobial efficacy by binding to the ZZ domain of SQSTM1
To further validate the selectivity of SQSTM1 agonists in therapeutic efficacy, we examined the activity of autophagy with ATB1095. Indeed, ATB1095 exhibited no detectible activity to induce the synthesis and lipidation of LC3 ( Figure  S4A). Moreover, CFU assays showed that ATB1095 and another negative control compound, ATB1094, failed to inhibit the infectivity of S. Typhimurium in RAW264.7 cells ( Figure S2K). These results verify the chemical selectivity of SQSTM1 agonists in therapeutic efficacy.
Next, we determined the binding specificity of YTK-A76 to the ZZ domain of SQSTM1. CFU assay with S. Typhimurium showed that the SQSTM1 agonist lost their efficacy when SQSTM1 was depleted using siRNA ( Figure 4A). The SQSTM1 agonist regained antimicrobial efficacy when SQSTM1 knockdown HeLa cells were transfected to express wild-type SQSTM1 but not mutant SQSTM1 lacking ZZ domain that showed comparable transfection efficiency ( Figure 4B and S4B). Furthermore, the activity of the SQSTM1 agonist to induce autophagosome biogenesis was dependent on the ZZ domain of SQSTM1 ( Figure S4C).
We also tested whether SQSTM1 agonists induced the recruitment of SQSTM1 to S. Typhimurium using colocalization assays. Immunostaining analyses showed that colocalization of SQSTM1 with DAPI-positive S. Typhimurium markedly increased when cells were treated with SQSTM1 agonists ( Figure 4C and S4D). To test whether the activity to recruit SQSTM1 to the bacterial surface is dependent on the SQSTM1 ZZ domain, HeLa cells were depleted of SQSTM1 using siRNA to the 3'-untranslated region (3'-UTR) and subsequently engineered to express SQSTM1-EGFP or SQSTM1-ZZΔ-EGFP. When the cells were treated with YTK-A76, the colocalization between SQSTM1 and bacteria was ~70% with SQSTM1-EGFP and ~30% with SQSTM1-ZZΔ-EGFP ( Figure 4D and S4E). These results demonstrate that SQSTM1 agonists induce the recruitment of SQSTM1 to the membrane of intracellular S. Typhimurium through their binding to the SQSTM1 ZZ domain.
gIt is known that SQSTM1 undergoes selfpolymerization via the PB1 domain and interacts with ubiquitin chains on cargoes via the UBA domain [60,61]. To determine whether SQSTM1 oligomerization is required for the recruitment of SQSTM1 to ubiquitinated bacterial membranes, SQSTM1 was depleted in HeLa cells using siRNA targeting its 3'-UTR, followed by transient expression of wild type SQSTM1-MYC and SQSTM1-PB1Δ-MYC. Immunostaining analyses of the cells infected with S. Typhimurium showed that YTK-A76 induced the association of wild-type SQSTM1, but not SQSTM1-PB1Δ-MYC, with bacteria ( Figure 4E). Likewise, mutant SQSTM1 lacking the UBA domain (SQSTM1-UBAΔ-MYC) failed to associate with bacteria ( Figure 4E). Consistently, when ubiquitin was depleted using siRNA targeting the UBB mRNA, YTK-A76 almost lost its ability to accelerate the SQSTM1 recruitment on the bacteria ( Figure 4F). This inability of SQSTM1 to associate with bacteria in both the absence and presence of SQSTM1 agonists was attributed to the loss of antimicrobial efficacy as determined by CFU assays ( Figure 4G). These results suggest that SQSTM1 agonists facilitate the recruitment of SQSTM1 to bacteria through self-polymerization via PB1 domain and the interaction with ubiquitin chains via UBA domain.

SQSTM1 agonists induce xenophagy of S. Typhimurium by facilitating the target of the pathogens to the autophagosome
To determine whether SQSTM1 agonists act as xenophagyinducing drugs, we characterized the autophagic flux of host cells infected with S. Typhimurium ( Figure 3). Immunostaining analyses of cells treated with SQSTM1 agonists revealed that the majority of SQSTM1-positive bacteria were also positive for LC3-positive autophagic membranes ( Figure S1B). The colocalization of SQSTM1 + LC3 + autophagic membranes with DAPI-positive S. Typhimurium markedly increased by SQSTM1 agonists ( Figure 5A). Moreover, the ability of SQSTM1 agonists to eradicate bacteria was almost completely blocked by the lysosomal inhibitor NH 4 Cl ( Figure 5B). Consistently, knockdown of either ATG5 or LC3B abolished the activity of SQSTM1 agonists to induce degradation of degradation of intracellular bacteria ( Figure 5C). The indispensable role of autophagy was further confirmed by the finding that the antimicrobial efficacy of SQSTM1 agonists were not significantly impaired when proteolytic flux via the UPS was inhibited by using the proteasome inhibitor MG132 ( Figure S5A and S5B). Next, to obtain the decisive evidence for xenophagy, we performed transmission electron microscopy (TEM) in RAW264.7 cells. A significantly increased portion of bacteria were sequestered within autophagosomes in cells treated with YTK-A76 ( Figure 5D and S5C). These results demonstrate that SQSTM1 agonists induce the targeting of intracellular bacteria to autophagic membranes, leading to lysosomal degradation.
It is known that MTOR is a master regulator of various signaling pathways including bulk autophagy, the cell cycle, immune responses, and metabolism [62,63]. We therefore tested whether pharmaceutical activation of autophagy using rapamycin also inhibits intracellular bacterial growth. In sharp contrast to SQSTM1 agonists, rapamycin exhibited no antimicrobial effect ( Figure 5E,F and S5D). Furthermore, unlike SQSTM1 agonists, rapamycin failed to enhance SQSTM1 recruitment of intracellular bacteria and bacterial targeting to autophagosomes ( Figure 5G). Resveratrol, a SIRT1 activator, also showed no antimicrobial activity against S. Typhimurium in RAW264.7 cells at 10-100 μM ( Figure 5H). These results suggest that SQSTM1 agonists exert antimicrobial efficacy via selective autophagy and are independent of rapamycin-modulated canonical autophagic pathways.

SQSTM1 agonists enhance host innate immunity against Mtb and other pathogens via SQSTM1-mediated xenophagy
Although xenophagy has been implicated in the lysosomal degradation of Mtb, it has not been fully understood about the underlying mechanisms and xenophagy-inducting reagents against Mtb [14,64]. To investigate whether the SQSTM1 agonists have anti-mycobacterial activities, we treated Mtb-infected BMDMs with YTK-A76, YT-6-2, YTK-2205, and YOK-1109. CFU assays showed that the growth rate of Mtb was dramatically inhibited at 5-10 μM ( Figure 6A and S6A). Similarly, YTK-2205 decreased the intracellular Mtb growth in J774A.1 cells ( Figure S6B). Silencing of SQSTM1 in BMDMs significantly abrogated the antimicrobial response induced by YTK-2205 during Mtb infection ( Figure 6B). During these experiments, the viability of both host cells and bacterial was not significantly influenced by SQSTM1 agonists at the concentrations used in this study ( Figure S6C,D). Likewise, no difference was observed in the entry, i.e., phagocytosis of the Mtb in the absence and presence of SQSTM1 agonists ( Figure S6E). These results demonstrate the efficacy of SQSTM1 agonists to inhibit the infectivity of Mtb.
Next, we examined whether SQSTM1 agonists induce the autophagic targeting of Mtb. Co-immunostaining analyses showed that SQSTM1 agonists induced the colocalization of SQSTM1 and Mtb ( Figure 6C and S6F). Moreover, YTK-2205 treatment of Mtb-infected BMDMs led to a significant increase of colocalization between Mtb phagosomes and autophagosomes as compared with the control ( Figure 6D). Also, the level of LC3B was significantly upregulated in Mtbinfected BMDMs by YTK-2205 treatment ( Figure S6G). Treatment of Mtb-infected BMDMs with YTK-2205 significantly upregulated the colocalization between Mtb phagosomes and lysosomes ( Figure S6H), suggesting that SQSTM1 agonists enhance targeting of Mtb to autophagosomes that fuse with lysosomes for degradation. Furthermore, whereas mycobacteria were mostly detected within phagosomal structures in control cells, a significant portion of Mtb was sequestered in morphologically heterogeneous double-membrane vesicles that appeared to be autophagosomes in YTK-2205treated BMDMs ( Figure 6E). Collectively, these data strongly suggest that SQSTM1 agonist-induced xenophagic flux is required for antimicrobial responses against Mtb infection.
Next, we also determined whether SQSTM1 agonists have antimicrobial efficacy against other bacteria including Gramnegative E. coli and the Gram-positive S. pyogenes. CFU assays (MOI of 10) showed that SQSTM1 agonists efficiently inhibited the growth of both strains ( Figure 6F,G). These results suggest that SQSTM1 has the potential as a drug target for a broad range of pathogenic bacteria.

SQSTM1 agonists enhance xenophagy mediated host defense against pathogens in mice
To validate the efficacy of SQSTM1 agonists in xenophagy, we orally infected C57BL/6 male mice with S. Typhimurium (10 6 per mice), followed by i.p. injection of 20 mg/kg YTK-A76 or YTK-2205 once daily for 6 days. On day 7, the liver and spleen were harvested, and their extracts were used for CFU assays. When compared with the vehicle-treated group, mice injected with YTK-2205 showed approximately a 100-fold reduction in the number of bacteria ( Figure 7A). Although the relative efficacy was weaker, a similar inhibition of bacterial infection was observed with YTK-A76 ( Figure 7A). Histological examination of H&E-stained paraffin sections from livers revealed inflammation-associated damaged tissues in control mice. Such damages were barely observed in mice injected with YTK-2205 ( Figure 7B). These results suggest that SQSTM1 agonists inhibit the infectivity of S. Typhimurium and reduce inflammation caused by bacterial infection.
Next, we investigated whether administration of YTK-2205 or YT-6-2 promotes host antimicrobial defense during Mtb infection in mice. We infected mice with Mtb intranasally (5 × 10 4 CFU), followed by i.p. injection of 10 mg/kg YTK-2205 per 2 days (total 5 injections; for Figure 7C left) or 20 mg/kg YT-6-2 thrice per week (total 11 injections; for Figure 7C right). Notably, CFU assays of lung extracts showed that the bacterial burden was significantly reduced in mice treated with YTK-2205 or YT-6-2 ( Figure 7C). Histopathological analyses also revealed significant decreases in the area of lung inflammation in mice treated with YT-6-2 ( Figure 7D). Similarly, we infected mice with Bacillus Calmette-Guérin (BCG) (1 × 10 7 CFU) and administered YTK-2205 by i.p. injection. Treatment with YTK-2205 resulted in reduced BCG burdens in the lungs ( Figure 7E) and the number of granulomatous lung lesions and inflamed areas decreased ( Figure 7F). Additionally, to explore whether SQSTM1 agonist amplified the host defense against multidrug-resistant (MDR) strain of Mtb, mice were infected with MDR-Mtb (5 × 10 3 CFU) and then treated with YT-6-2. Relative to a vehicle, injection of YT-6-2 significantly reduced the MDR-Mtb growth in the lungs ( Figure 7G). These results demonstrate the efficacy of SQSTM1 agonists to decrease bacterial load in the lungs and pulmonary inflammation in mice.

The SQSTM1 agonists inhibit the production of proinflammatory mediators in vitro and in vivo
Host defense to bacterial infection depends on the coordinated responses of inflammatory responses induced by innate and adaptive immune cells [65,66]. Because SQSTM1 agonists reduced tissue pathology and enhanced host immune responses in S. Typhimurium -and Mtb-infected mice, we sought to evaluate the possible beneficial effects of the drugs on chronic inflammation associated with S. Typhimurium and Mtb disease. We therefore determined whether SQSTM1 agonists have efficacy to reduce the excessive production of cytokines in HeLa cells infected with S. Typhimurium. RT-qPCR analyses showed that YTK-A76 efficiently counteracted the mRNA induction of IL1B (interleukin 1 beta) and IL6 ( Figure 8A). Furthermore, to examine the role of SQSTM1 agonists in the modulation of inflammatory responses during Mtb infection, BMDMs infected with Mtb were treated with YTK-A76. The SQSTM1 agonists efficiently suppressed Mtbinduced upregulation of pro-inflammatory cytokines (Il6 and Il1b) and chemokine (Cxcl5) ( Figure 8B). Consistently, in peritoneal macrophages (PMs) and BMDMs infected with Mtb, YTK-2205 at 5 μM inhibited Mtb-induced upregulation of the pro-inflammatory cytokines and chemokine ( Figure 8C and S7). Furthermore, BCG-infected mice injected i.p. with YTK-2205 produced a reduced level of tumor necrosis factor alpha (Tnfa) mRNA in the lung tissues compared to those in the vehicle group ( Figure 8D). Collectively, these results indicate that SQSTM1 agonists inhibit the expression of inflammatory mediators during pathogenic bacterial infections.

Discussion
Selective autophagy mediates the targeting of specific cellular materials such as misfolded proteins and damaged organelles to phagophores for lysosomal degradation [1,2,5,67,68]. Recent studies showed that intracellular viruses and bacteria can be targeted for lysosomal degradation via selective autophagy [8,9,14]. Several chemical modulators of bulk and selective autophagy have been exploited to induce the targeted degradation of bacteria, yet there are no therapeutic agents available whose mode-of-actions are clearly defined with satisfactory efficacy, selectivity, and safety. In this study, we exploited small-molecule agonists to the ZZ domain of SQSTM1 as anti-microbial agents against a broad range of bacteria such as S. Typhimurium, E. coli, and S. pyogenes as well as Mtb. These agonists induced xenophagy during which SQSTM1 was associated with bacterial membranes to recruit LC3 on autophagic membranes, leading to lysosomal degradation. The anti-microbial efficacy of SQSTM1 agonists was demonstrated in vitro and using mouse models of S. Typhimurium and Mtb. Our results identify SQSTM1 as a potential drug target in host-directed therapy against a broad range of pathogenic bacteria including multi-drug resistant strains.
Our earlier work has identified the autophagic receptor SQSTM1 as an N-recognin that recognizes various types of N-degrons such as the Nt-Arg and targets the cargoes carrying N-degrons to autophagic membranes and lysosomes [56,58]. In this study, we developed novel SQSTM1 agonists that exhibit efficacy in xenophagy of various intracellular bacteria in cultured cells as well as mice. The anti-microbial efficacy involved no detectible activity to directly kill bacteria, further supporting the selective role of SQSTM1-mediated autophagy in host innate immunity. Given our earlier results, we speculate that this xenophagy process is driven by N-degron-mediated intramolecular activation of SQSTM1, which involves the exposure of PB1 domain that facilitates SQSTM1 polymerization and LIR domain that interacts with LC3. Consistently, rapamycin, an inducer of MTOR-modulated bulk autophagy exhibited no efficacy in xenophagy of S. Typhimurium, highlighting the specificity of SQSTM1 agonists to induce selective autophagy to target and eradicate intracellular bacteria. One remaining question involves the molecular mechanism by which SQSTM1 is recruited to bacterial membranes. Our previous studies showed that SQSTM1 agonists facilitated the autophagic degradation of protein aggregates as well as the ER along with its luminal contents [59], with differential preference and selectivity depending on the chemical structures. While further investigations are needed to correlate the chemical structures to the intramolecular conformational changes and the spatiotemporal arrangement of SQSTM1 associated with its cargoes, one could speculate that the oligomerization of SQSTM1 may be a common step for various cargo types. Nonetheless, it should be noted that where SQSTM1 oligomers during aggrephagy grow into large aggregates via uncontrolled formation of disulfide bond formation between SQSTM1 [58], SQSTM1 forms oligomers in an organized topology during the ER-phagy [59]. We therefore suggest that SQSTM1 during xenophagy is associated with an unknown receptor(s) on bacterial membrane in a manner similar to that in the ER-phagy. If this were true, another burning question concerns the receptor for SQSTM1 on bacterial membrane. Screening of such bacterial receptors is underway.
Various pathogens have evolved strategies to escape from or inhibit host defense systems including selective autophagy. In this study, we focused on the antimicrobial effects against S. Typhimurium and Mtb, both of which are well-known intracellular pathogens that can be counteracted by interplay between the UPS and xenophagy activation [69][70][71][72][73][74]. During infection, S. Typhimurium invade into host cells and reside in Salmonellacontaining vacuoles (SCVs), which are damaged by the type III secretion systems (T3SSs), allowing the escape and cytosolic exposure of S. Typhimurium [75][76][77]. By contrast, Mtb is a more successful pathogen that resides within phagosomes and arrests phagosomal maturation [78,79]. Mtb access into the cytosol through Mtb ESX-1 type seven secretion system via disruption of phagosomal membrane contributes to its virulence mechanism [80,81]. Once exposed to the cytosol, both bacteria are sensed and rapidly ubiquitinated to generate a xenophagy signal, leading to the recruitment of cargo receptors such as SQSTM1, OPTN, and CALCOCO2 [73,74,[82][83][84]. Although multiple anti-Mtb drugs have been developed for decades, it remains challenging to overcome the issues of antibiotic resistance for therapeutics against Mtb infection. In this study, we show that SQSTM1 agonists enhances the ability of host cells to recognize and target cytosolic Mtb to autophagic membranes. Moreover, these xenophagy inducers did not activate inflammatory immune responses, but suppressed several proinflammatory cytokines, such as IL1B/IL-1β and TNF/TNF-α. Such an activity to downregulate proinflammatory responses may represent to the function of selective autophagy as a regulator of excessive inflammatory responses ( Figure 7). Importantly, we found the antimicrobial effects of SQSTM1 agonists upon MDR-Mtb infection in vivo, highlighting the potential use of SQSTM1 agonists as host-directed therapeutics against drug-resistant TB. In addition, SQSTM1 agonists were effective against infection with BCG, which lacks genes encoding ESX-1 to promote cytosolic translocation and xenophagy [73,85]. Although one remaining question concerns whether SQSTM1 agonists facilitates the targeting of phagocytosed or exposed Mtb in the cytosol, our results suggest that SQSTM1dependent xenophagy may provide a therapeutic strategy different from the currently used one.
As multidrug resistance of pathogenic bacteria is rapidly rising, it is increasingly urgent to develop novel strategies fundamentally different from traditional antibiotics. Recent studies explored several anti-microbial strategies such as phage therapy, vaccines, antibodies, probiotics, antimicrobial peptides, and host-directed therapy [38][39][40][41]. Host-directed therapy aims to enhance host defense mechanisms or modulate excessive inflammation [42], such as autophagy. Several chemical modulators of autophagy have been used to induce targeted degradation of intracellular bacteria, including rapamycin, metformin, resveratrol, AR-12, and D61 [28,86]. These therapeutic approaches target AMP-activated protein kinase activation, vitamin D receptor signaling, innate immune activation, and sirtuin 1 pathway [19,87,88]. Despite rather extensive studies, there are no generally applicable therapeutic strategies whose mode-of-actions are clearly defined with satisfactory efficacy and safety. Our results show that the chemical activation of SQSTM1-dependent selective autophagy is generally applicable for various bacterial strains. It should be noted that whereas the traditional antibiotics strictly depend on molecular interactions, the mode-ofaction of SQSTM1-dependent xenophagy involves the recognition of (universally conserved) ubiquitin chains on bacterial membranes by SQSTM1, which in turn recruits autophagic membranes to the site of degradation. Therefore, in principle, SQSTM1 may be a potential drug target in host-directed therapy against a broad range of multi-drug resistant bacteria. The development of xenophagy-inducing drugs is under way. HCT116 cells were cultured in McCoy's 5a Medium (GIBCO, 16600). All of the media were supplemented with 10% FBS and 100 units/mL penicillin-streptomycin. Primary BMDMs isolated from C57BL/6 mice and cultured in DMEM. BMDMs were differentiated for 3-5 days in the presence of CSF2/M-CSF (colony stimulating factor 2; R&D Systems, 416-ML). For the preparation of PMs, each mouse was injected (i. p.) with 1 ml of 3% thioglycolate (Sigma-Aldrich, T0632) and after 3 days the peritoneal fluid was collected in ice cold phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 , pH 7.4) containing 3% FBS. The collected cell suspension was centrifuged and the cells were counted. To generate SQSTM1-knockout HeLa cells, oligos were synthesized and inserted into the px330puro vector (Addgene, 110403; deposited by Sandra Martha Gomes Dias) using a published protocol to generate gRNA with hCas9 protein [89]. gRNA sequences were designed using an online program provided at http://chopchop.cbu. uib.no/. gRNA sequences were as follows: SQSTM1 -TTGTAGCGGGTTCCTACCAC(-). HeLa cells were transfected with the targeting plasmid using Lipofectamine 2000 reagent according to the manufacturer's instructions (Invitrogen Life Technologies, 11668019). And then clones propagated from single cells were chosen. Immunoblotting and genomic DNA sequencing were used to confirm the depletion of target genes. All the cell lines were determined to be negative in a mycoplasma test using a MycoAlert detection kit (Lonza, LT07-118). All the culture plates and the cell lines were maintained at 37°C and 5% CO 2 in a humidified incubator.

Mice
C57BL/6 mice (sex-matched) aged 6-8 weeks with a wild-type (WT) background were purchased from Samtako Bio Korea (Gyeonggi-do, Korea). Mice were maintained under specific pathogen-free conditions. All mice experiments and maintenance were done in adherence to guidelines set forth by the Institutional Animal Care and Use Committee, Seoul National University (SNU-190531-2-1) and Chungnam National University School of Medicine (202009A-CNU-155, CNUH-A0044-1), and the Korean Food and Drug Administration. For the Salmonella model, C57BL/6 mice aged 8 weeks were given 1 × 10 6 CFU of S. Typhimurium and SQSTM1 agonist (20 mg/kg) by i.p. injection. The mice were given SQSTM1 agonist (20 mg/kg) by i.p. daily for 6 days and at which point they were killed, the liver and spleen removed, homogenized, and resuspended in PBS, followed by plating on Luria Broth agar plates. The plates were incubated overnight at 37°C and colonies were counted. For the BCG model, C57BL/6 mice aged 8 weeks were given 1 × 10 7 CFU of BCG intranasally and the SQSTM1 agonist (20 mg/kg) by i.p. injection. After infection, the mice were i.p. injected with YTK-2205 at 3-6 days post-infection (dpi) and sacrificed at 7 dpi. For the Mtb model, C57BL/6 mice aged 8 weeks were given 5 × 10 4 CFU of Mtb by intranasal injection and the SQSTM1 agonist (10-20 mg/kg) by i.p. injection. After infection, the mice were i.p. injected with YTK-2205 q.o.d. 5 times and sacrificed at day 10 dpi and YT-6-2 t.i.w. for 28 days. For MDR-TB infection model, C57BL/6 mice aged 8 weeks were given 5 × 10 3 CFU of MDR-Mtb by intranasal injection and the YT-6-2 (20 mg/ kg) by i.p. injection. After infection, the mice were i.p. injected with YT-6-2 t.i.w. for 28 days and lung tissues were collected, homogenized and re-suspended in PBS and plated on 7 H10 agar.

Bacterial infection
S. Typhimurium were inoculated at 2 mL LB broth and cultured at 37°C overnight. A suspension of S. Typhimurium (100 µl) was subcultured in 10 mL LB for 3 h. The absorbance of bacterial suspension was measured and diluted suspension of S. Typhimurium was inoculated to cultured cells with indicated MOI for 30 min. After the incubation the cells were washed with Dulbecco's Phosphate-Buffered Saline (DPBS; Thermo Fisher Scientific, 14,040,133) and the cells were treated with media containing 100 µg/ml gentamycin (Thermo Fisher Scientific, 15,710,064) for 1 h and the media was exchanged with fresh medium including 10 µg/ ml gentamycin. Cells were infected with indicated MOI of Mtb for 2 h. Extracellular bacteria were washed with PBS and infected cells were further cultured in fresh medium for the indicated time. For in vivo infection, frozen Mtb or BCG were thawed and inoculated intranasally (Mtb; 5 × 10 4 CFU/mice, BCG; 1 × 10 7 CFU/mice, MDR-Mtb; 5 × 10 3 CFU/mice). To measure the bacterial burden, lungs were harvested after sacrificing the mice after 10 days of Mtb infection, 7 days of BCG, and 10 or 28 days of MDR-Mtb infection. Lungs were homogenized in PBS and serial dilutions of the homogenates were plated in 7H10 agar plates. After 2-3 weeks, colonies were counted.

RNA extraction and qRT-PCR analysis
Total RNA from homogenized lung or cell was isolated using TRIzol reagent (Thermo Fisher Scientific, 15596-026), following the manufacturer's instructions. After RNA quantification, cDNA was synthesized by reverse transcription using the reverse transcriptase premix (Elpis Biotech, EBT-1515). Real-time PCR (qRT-PCR) was performed using SYBR green master mix (Qiagen, 330503) and primers for indicated genes, in Rotor-Gene Q 2plex system (Qiagen). Data were analyzed using 2 ΔΔ threshold cycle (Ct) method where mouse Gapdh was used for normalization. Data are expressed as relative fold changes. Primer sequences (human) are as follows: The mRNA expression level of autophagy related genes was analyzed by using AccuTarget™ qPCR screening kit (Bioneer, SH-0017-10).

Synthesis of 3,4-bis((4-fluorobenzyl)oxy)phenol (B2)
To a solution of B1 (187 g, 526.7 mmol) in DCM (2 L) was added m-CPBA (126 g, 730.4 mmol). The mixture was stirred at RT for 16 h. Then the reaction was washed with saturated sodium bicarbonate solution, concentrated under vacuum. Then the crude product is added to methanol (1.5 L) and water (200 mL) was added KOH (58.9 g, 1.05 mol). The mixture was stirred at RT for 3 h. Then the reaction was filtered and the solid to dryness under vacuum. The crude compound was purified by silica gel, eluted with EA/PE

Immunofluorescence analysis
Cells were cultured on coverslips and infected with Mtb-ERFP as described above. After the appropriate infection, cells were washed three times with PBS, fixed with 4% paraformaldehyde for 15 min, permeabilized with 0.25% Triton X-100 (Sigma-Aldrich, T8787) for 10 min, and incubated with primary antibodies for 2 hours at room temperature or overnight at 4°C. Cells were washed with PBS to remove excess primary antibodies and then incubated with secondary antibodies for 1 hour at room temperature. Each experiment was completed on duplicate coverslips and the results are expressed as the mean and standard deviation. Images of dynamic cell colocalization were recorded as vertical z-stacks. LAS X small 2.0 and Adobe Photoshop 7 (Adobe Systems) were used for image processing.

Colony-forming unit (CFU) assay
Wild type (WT) strain of Salmonella enterica serovar Typhimurium (S. Typhimurium) strains was utilized for the study. A single colony of S. Typhimurium was grown overnight at 37°C in the shaking incubator. Secondary culture (5% inoculum) was grown for three hours in microaerophilic conditions. HeLa or RAW264.7 cell lines were infected at a MOI of 10 for 30 min. The infected cells were washed with DPBS and the cells were treated with media containing 100 µg/ml gentamycin (Thermo Fisher Scientific, 15710064) for 1 h to exterminate the extracellular bacteria. The media was exchanged with fresh medium including 10 µg/ml gentamycin. The cells were further cultured with and without the chemical SQSTM1 agonist and incubated for the indicated time in figure legends. Finally, the mammalian cells were lysed using lysis buffer (0.1% Triton X-100, 1X PBS (Welgene, LB 204-01). The intracellular S. Typhimurium containing lysates were serially diluted and spread on the Luria broth (LB) agar plate. After overnight incubation at 37°C, the CFU was determined. For intracellular bacterial viability, cells were infected with Mtb for 4 h. Cells were washed with PBS and lysed in distilled water to release the intracellular bacteria. The harvested bacteria were then plated in Middlebrook 7H10 agar with OADC and incubated for 2-3 weeks, and colonies were counted.

RNA interference assay
Reagents for siRNA silencing were purchased from Life Technologies. Transfection was performed at a final concentration of 40 nM using Lipofectamine RNAiMAX reagent (Invitrogen, 13778150) according to the manufacture's protocol.

Molecular docking studies
For the docking study, five novel compounds were generated and optimized in Cresset Flare software. The available crystal structure of the SQSTM1 ZZ domain (PDB ID: 6MIU) was downloaded from Protein Data Bank (https:// www.rcsb.org). Protein preparation was carried out in Cresset module Flare software. Hydrogen's and 3D protonation were carried out on the target protein and minimized for the active site residues. Docking experiments were performed by using Cresset Flare software in accurate mode and default settings.

Affinity-isolation assays of YT-6-2
Plasmids expressing SQSTM1 wild type and ZZ point mutant (D147K) constructs were transiently transfected into SQSTM1 −/− HeLa cells using Lipofectamine 2000. After 24 h, trypsinized cells were collected in growth medium and centrifuged. The cell pellets were resuspended in a hypotonic buffer (10 mM HCl, 1.5 mM MgCl 2 , 10 mM HEPES, pH 7.9) and incubated in ice for 30 min. The cell suspensions were subjected to ten freeze-thaw cycles, followed by centrifugation at 12,000 × g at 4°C for 15 min. For the affinity-isolation assay, biotinylated YT-6-2 was incubated with streptavidin agarose resin (272 µg of compound per ml settled resin; Thermo Fisher Scientific, 20359). The X-peptide beads were diluted in five volumes of PBS and incubated overnight at 4°C. The beads were centrifuged at 1000 × g for 3 min and washed three times in an equal volume of PBS. Soluble SQSTM1 −/− HeLa cell extracts containing 500 µg of total protein were diluted in 500 µl binding buffer (0.05% Tween 20, 10% glycerol, 0.2 M KCl, 20 mM HEPES, pH 7.9) and mixed with YT-6-2 conjugated beads (30 µl packed volume). The mixtures were incubated at 4°C for 1 h with gentle rotation. The beads were pelleted by centrifugation at 1000 × g for 3 min, washed five times with 500 µl of binding buffer at 4°C for 10 min, resuspended in 30 µl SDS sample buffer, and heated at 100°C for 5 min. Analysis was performed by SDS-PAGE and immunoblotting.

Transmission electron microscopy (TEM)
RAW264.7 cells were infected with S. Typhimurium for 30 min followed by incubation with 5 μM YTK-A76 for 4 h. BMDMs were infected with Mtb for 4 h and treated with 5 μM YTK-2205 for 18 h. The cells were scraped from the culture dish and pelleted by centrifugation at 3000 x g for 5 min. Pellets were resuspended in 2.5% glutaraldehyde in 0.1 M phosphate (pH 7.4) overnight at 4°C. Subsequently, 50 nm sections were cut and stained with uranyl acetate and lead citrate using the Reichert Ultracut S Ultramicrotome (Leica Microsystems) and FEI Vitrobot Mark IV (Thermo Scientific), respectively. Cell sections were examined using the JEOL JEM-1400 series 120 kV Transmission Electron Microscope at the Seoul National University Hospital Biomedical Research Institute.

Histology
For histopathology, liver and lung tissues were fixed in 10% formalin and embedded in paraffin wax. The tissues were sectioned with 4 μm thickness. Paraffin-embedded slides were freshly treated with Neo-Clear (Millipore, 65351) twice for 10 min each, followed by gradual rehydration in EtOH (100%, 90%, 80%, and 70%; 6 min each) and water for 20 min. The tissues were then stained with hematoxylin and eosin (H&E). H&E-stained sections were scanned with an Aperio digital pathology slide scanner (Leica) and imaged using an Aperio ScanScope® CS System.

Immunocytochemistry
Autoclaved 22-mm 2 coverslips were placed in 24-well plates. HeLa cells were seeded 2.5 × 10 5 cells/well and cultured overnight for further experiments. Cells were infected with S. Typhimurium S14028 with an MOI of 10. Cells were fixed in 4% PFA in PBS, for 15 min at room temperature. After washing twice with PBS, the cells were incubated for 1 h in blocking solution (5% bovine serum albumin and 0.3% Triton X-100 in PBS). The samples were incubated with primary antibodies overnight at 4°C and secondary antibodies were incubated at room temperature. After washing out secondary antibodies with PBS, the samples were mounted on a cover slide with the mounting solution. Confocal images were taken with a Zeiss LSM 700 laser scanning confocal microscope equipped with Zeiss C-Apochromat 60x (1.2 NA) and 40x (1.2 NA) water immersion lens and analyzed using ZEN (black edition) 2012 SP5 software (Zeiss). Using the ZEN software, z-stacks of images covering the entire cell thickness were acquired and projected maximally. Image processing and annotation were done with Adobe Photoshop, Adobe Illustrator and Fiji software.

Flow cytometry
BMDMs were analyzed by flow cytometry for phagocytosis using an ACEA NovoCyte flow cytometer as indicated by the manufacturer. BMDMs were infected with Mtb-ERFP for 2 h and treated with SQSTM1 agonist. Flow cytometry data were collected and analyzed using De Novo software (CA, Glendale, USA).

MTT assay
BMDMs were seeded in 96 well plate and treated with SQSTM1 agonists for 72 h treated with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Sigma-Aldrich, M5655) solution and incubated for 2 h. The medium was removed and cells were added DMSO to solubilize the crystals. The absorbance was measured with a reader spectrophotometer (Bio-Tek Synergy TH, Winooski, VT, USA) at 570 nm, using 690 nm as a reference.

Quantification and statistical analysis
All experiments were repeated at least three times and all data are presented as the mean ± SD or ± SED. In most data, twotailed unpaired t-test (for parametric data) or Mann-Whitney U-test (for nonparametric data) were used for comparisons of two groups and one-way analysis of variance (ANOVA; Dunnett's test) was performed for comparisons of multiple groups. For comparison of the magnitude of changes in different conditions, two-way ANOVA with Bonferroni posttests was used. Statistical significance was determined as values of p < 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001, ns: not significant). For each experiment, sample size (n) was determined as stated in the figure legends. All statistical analyses were performed with Prism 8.2 software (GraphPad).
Wonhyoung Seo for genomic DNA sequencing, Jin-Man Kim for analytical tools and analyzed the data.

Disclosure statement
No potential conflict of interest was reported by the author(s).