Diverse pathogens activate the host RIDD pathway to subvert BLOS1-directed immune defense

The phagocytosis and destruction of pathogens in lysosomes constitute central elements of innate immune defense. Here, we show that Brucella, the causative agent of brucellosis, the most prevalent bacterial zoonosis globally, subverts this immune defense pathway by activating regulated IRE1α-dependent decay (RIDD) of mRNAs encoding BLOS1, a protein that promotes endosome-lysosome fusion. RIDD-deficient cells and mice harboring a RIDD-incompetent variant of IRE1α were resistant to infection. Non-functional Blos1 struggled to assemble the BLOC-1-related complex (BORC), resulting in differential recruitment of BORC-related lysosome trafficking components, perinuclear trafficking of Brucella-containing vacuoles (BCVs), and enhanced susceptibility to infection. The RIDD-resistant Blos1 variant maintains the integrity of BORC and a higher-level association of BORC-related components that promote centrifugal lysosome trafficking, resulting in enhanced BCV peripheral trafficking and lysosomal-destruction, and resistance to infection. These findings demonstrate that host RIDD activity on BLOS1 regulates Brucella intracellular parasitism by disrupting BORC-directed lysosomal trafficking. Notably, coronavirus MHV also subverted the RIDD-BLOS1 axis to promote intracellular replication. Our work therefore establishes BLOS1 as a novel immune defense factor whose activity is hijacked by diverse pathogens.

Brucella is an intracellular vacuolar pathogen that invades many cell and tissue types, including non-2 professional and professional phagocytes (de Figueiredo et al, 2015). Brucellosis has eluded 3 systematic attempts at eradication for more than a century (Godfroid et al, 2002), and even in most 4 developed countries, no approved human vaccine is available (Ficht & Adams, 2009). The 5 intracellular lifestyle limits exposure to host innate and adaptive immune responses and sequesters 6 the organism from the effects of some antibiotics. Brucella evades intracellular destruction by 7 limiting interactions of the Brucella-containing vacuole (BCV) with the lysosomal compartment 8 (Criscitiello et al, 2013;Pizarro-Cerda et al, 1998). BCVs harboring internalized Brucella traffic from 9 endocytic compartments (eBCVs) to a replicative niche within vacuoles (rBCVs) that are decorated 10 with markers of the endoplasmic reticulum (ER) (Pizarro-Cerda et al., 1998;Starr et al, 2012). BCVs 11 also accumulate autophagic membranes (aBCVs), which constitute a distinctive aspect of the 12 intracellular lifestyle of the pathogen (Pandey et al, 2018;Starr et al., 2012). The VirB type IV 13 secretion system (T4SS) is a significant virulence factor that regulates Brucella intracellular 14 trafficking (Marchesini et al, 2011;Paredes-Cervantes et al, 2011;Sa et al, 2012;Smith et al, 2012). Brucella effectors secreted by the T4SS promote bacterial intracellular trafficking and growth via 16 modulation of host functions (de Barsy et al, 2011;De Jong et al, 2008;Dohmer Pisani et al, 2014;17 Miller et al, 2017;Myeni et al, 2013) and organisms that lack this system fail to establish productive 18

infections. 19
The Unfolded Protein Response (UPR) is an evolutionarily conserved signaling pathway that allows 20 the ER to recover from the accumulation of misfolded proteins (Gardner et al, 2013;Walter & Ron, 21 2011) during ER stress. The UPR signals through the stress sensors IRE1α, ATF6, and PERK located 22 in the ER membrane. When the luminal domains of these proteins sense unfolded proteins, they 23 transduce signals to their cytoplasmic domains, which initiate signaling that ultimately results in UPR 24 (Lee et al, 2008). IRE1α plays a central role in triggering UPR through an endonuclease/RNase 25 activity in its cytoplasmic tail that catalyzes the splicing of Xbp1 mRNA, which is then translated to 26 generate the XBP1 transcription factor (Lee et al., 2008;Ron & Walter, 2007). IRE1α RNase activity 27 can also cleave a wide variety of cellular mRNAs that leads to their degradation in a process termed 28 regulated IRE1-dependent mRNA decay (RIDD) (Hollien & Weissman, 2006). The RIDD pathway 29 displays selectivity. For example, the pathway cleaves a specific subset of mRNAs encoding 30 showed that these animals had similar B cell (B220 + ), T cell (CD4 + or CD8 + ), and CD11b + profiles 23 (Figure 1-figure supplement 1C). We also showed that the IL-1b, IL-6 and TNF-α responses of 24 BMDMs to LPS stimulation were reduced in mutant mice ( Figure 1A-C), consistent with previous 25 findings that these LPS-mediated responses are controlled, in part, by IRE1α activity (Martinon et al, 26 2010). 27 To determine whether IRE1α activity in macrophages contributed to pathogen burden, dissemination, 28 and disease progression, we infected wt-IRE1α control and m-IRE1α mice with Bm16M via the 29 intraperitoneal route, humanely sacrificed the mice at various times post-infection, and then 30 determined the bacterial burden in assorted tissues by quantifying the number of recovered colony 1 forming units (CFUs). We found that tissue-specific mutation of IRE1α resulted in enhanced 2 resistance to bacterial infection with significant reductions in bacterial load in the spleen and liver 3 compared to wt-IRE1α controls at 7 or 14 days post-infection (dpi), respectively (Figure 1D-E). 4 However, both infected wt-and m-IRE1α mice displayed similar spleen weights (Figure 1-figure  5 supplement 1D-E) and spleen or liver inflammation ( Figure 1F-H), revealing that the lower numbers 6 of CFU recovered from m-IRE1α animals were not accompanied by corresponding decreases in 7 inflammation. To test the hypothesis that the differential bacterial burden in macrophage cells, the 8 predominant cell type in which the pathogen resides and replicates in vivo, accounted for this 9 reduction, we compared the bacterial load in CD11b + cells from control and m-IRE1α mice that had 10 been infected with Bm16M for seven days. We found that indeed CD11b + cells from the spleens of 11 m-IRE1α mice displayed striking reductions in bacterial load ( Figure 1I), thereby suggesting that the 12 resistance of these cells to intracellular parasitism contributed to the resistance phenotype observed 13 at the organismal level. Divergent Brucella species display distinct host preferences; however, their 14 interactions with host cells share common features (de Figueiredo et al., 2015). To test the hypothesis 15 that m-IRE1α mice also displayed resistance to infection by other Brucella species, we infected these 16 mice with B. abortus strain S2308 (BaS2308), a strain that displays tropism for cattle. We then 17 assessed tissue burden in spleen and liver at 7 dpi. We found that m-IRE1α mice also exhibited 18 resistance to BaS2308 infection ( Figure 1J), thereby indicating that the resistance phenotype of the 19 mutant mice was not pathogen species-specific. 20

IRE1α RNase activity confers susceptibility to Brucella infection 21
Xbp1 splicing was dramatically diminished in BMDMs from m-IRE1α mice (Figure 1-figure  22 supplement 1B), indicating that BMDMs from the m-IRE1α mice carried the expected functional 23 defects in IRE1α RNase activity. We thus tested the hypothesis that IRE1α RNase activity confers 24  recovered from BMDMs or RAW264.7 macrophages treated with 4µ8C, a compound that specifically 30 IRE1α (Figure 3-figure supplement 1B). At 4 or 24 h.p.i., we harvested host mRNA for RNA-seq 23 analysis. Differential expression analysis was then performed to identify genes that were down-24 regulated following Bm16M infection of wt-IRE1α cells but were unchanged or up-regulated in either 25 infected, drug-treated cells, or infected, m-IRE1α cells. Genes that displayed reduced expression (p 26 < 0.05) in response to infection at 4 and/or 24 h.p.i., and also whose infection-dependent reductions 27 in expression were reversed upon treatment with 4µ8C, or in m-IRE1α cells, were defined as 28 candidate "RIDD genes". This analysis resolved 847 candidate RIDD genes (  . Fourth, we compared our list of candidate RIDD genes to genes previously 16 reported to be subject to RIDD control (Bright et al., 2015;Han et al, 2009;So 17 et al, 2012). This comparison identified 40 genes that were previously shown to be substrates of 18 IRE1a RNase activity and/or displayed expression patterns consistent with RIDD targeting ( Figure  19 3A-B). Finally, we found that the expression of the key RIDD gene Blos1 was also reduced in ∆Xbp1 20 BMDMs infected with Bm16M, or when ER stress was induced in these cells ( Figure 3H). These 21 data suggested that the observed changes in host gene expression patterns were not a consequence of 22 alterations in XBP1 transcription factor activity. Taken together, these data supported the hypotheses 23 that (1) Bm16M infection induces RIDD activity in host cells, and (2) RIDD activity confers 24 enhanced susceptibility to intracellular parasitism by Brucella. However, these findings left open the 25 question of the molecular mechanism by which RIDD activity controlled Bm16M replication. 26

RIDD activity on Blos1 controls Brucella intracellular parasitism 28
Our RNA-seq analysis identified Blos1 as a Brucella-induced RIDD gene. However, the mechanisms 29 by which Blos1 regulates microbial infection are largely unknown. This fact encouraged us to test the 30 hypothesis that Blos1 plays a central role in regulating Bm16M intracellular parasitism. First, we 1 generated a cell line carrying a non-functional Blos1 mutant allele (mBlos1). Mammalian BLOS1 2 contains three conserved XAT hexapeptide-repeat motifs that are essential for acetyltransferase 3 activity and may also be a necessary structure-defining feature for acetyl-CoA contact (Scott et al, 4 2018;Wu et al, 2021a). Using CRISPR/cas9-mediated gene editing, we mutated the first XAT 5 hexapeptide-repeat motif, which in the wild-type encodes "EALDVH," and in the mutant encodes 6 "EVVDH or EVDH" (Figure 4-figure supplement 1A, Table S1). A cell line containing gene 7 encoding Cas9 and a non-specific gRNA was used as a control of the mBlos1 mutant line. Second, 8 we generated a RIDD-resistant Blos1 cell line (henceforth Rr-Blos1). In this line, a mutation (from 9 "G" to "U") was introduced into Blos1 mRNA stem-loop structure (i.e., the target of IRE1α RNase  Table S1). 13

14
We characterized the developed cell lines in several ways. First, we noted that a-tubulin acetylation 15 levels had been reported to be controlled, in part, by BLOS1 activity levels (Wu et al, 2018). 16 Therefore, we monitored a-tubulin acetylation to assess whether our developed cell lines did, in fact, 17 display alterations in BLOS1 activity. We found that mBlos1 and Rr-Blos1 cells displayed reduced 18 levels ( Figure 4A-B) and maintained relatively higher levels ( Figure 4C-D), respectively, of 19 acetylated α-tubulin, compared to their corresponding controls. These data supported the hypothesis 20 that these cells had the expected levels of BLOS1 activity. Second, we tested the replication of the 21 pathogen in different Blos1 cell lines. We found that mBlos1 cells exhibited increased susceptibility course of infection. We found that BLOS1 expression was reduced at 16 h.p.i., and continuously 26 decreased during Bm16M infection in wt-IRE1α control cells; however, in m-IRE1α BMDMs, 27 BLOS1 expression was relatively stable or increased (at 48 h.p.i.) (Figure 4G-H). Similar results 28 were observed in 4µ8C treated or untreated mBlos1, Rr-Blos1, and control cells infected with 29 BaS2308 (Figure 4-figure supplement 1D-E). These data demonstrate that low or high BLOS1 30 expression levels promote or impair Brucella infection, respectively. 31

BLOS1 regulates Brucella intracellular trafficking 1
The mechanism by which BLOS1 regulates Brucella infection was unknown. However, the observed 2 subcellular trafficking defect of the pathogen in host cells harboring mutant or deficient variants of 3 IREα (Figure 2; Figure 2-figure supplement 1) suggested that BLOS1 may control the 4 intracellular parasitism of the pathogen by regulating its subcellular trafficking. To illuminate this 5 aspect, we treated mBlos1, Rr-Blos1, and the corresponding control cell lines with tunicamycin (Tm, 6 an UPR inducer) or 4µ8C, or infected them with Bm16M. We then assessed the trafficking of the 7 pathogen in these cells using CIM. We found that low levels of BLOS1 or non-functional BLOS1 in 8 uninfected or infected cells were associated with the accumulation of late endosome/lysosome 9 (LE/Lys) membranes in the vicinity of nuclei, reduced colocalization of latex beads with cathepsin 10 D, and increased perinuclear LC3b index or autophagic activity near nuclei, in both control and Tm-11 observed at 24 and 48 h.p.i., the mBlos1 cells supported enhanced BCV trafficking to ER 20 compartments during bacterial infection, compared to that in the wild-type control cells, or 4µ8C-21 treated mBlos1 cells (Figure 4I-K). In contrast, Rr-Blos1 cells displayed reduced BCV trafficking to 22 ER compartments, but instead promoted BCVs trafficking to lysosomes during Bm16M infection, 23 compared to controls (Figure 4L-N). 24 To test the hypothesis that Bm16M infection alters the dynamics of associations between BLOS1 and 25 BCVs, we used CIM approaches to localize these elements during a time course of infection after 26 confirmation of the specificities of antibodies used in the work (see below) (Figure 4-figure  27 supplement 3A-D). We found higher levels of BLOS1 colocalization with BCVs in 4µ8C-treated 28 control or mBlos1 cells than their corresponding untreated cells ( Figure 4O); moreover, at 24 h.p.i., 29 lower levels of BLOS1 + BCVs were observed in mBlos1 cells compared to wild-type controls ( Figure  30 (Pu et al., 2015). We hypothesized that degradation of Blos1 mRNA by IRE1a 30 during Brucella infection interferes with BORC assembly, resulting in the alteration of recruitment 1 or disassociation of BORC-related trafficking components, and increased LAMP1 + -BCV perinuclear 2 trafficking and fusion with the ER and/or macroautophagosome membranes. To test this hypothesis, 3 we performed protein co-immunoprecipitation (Co-IP) assays to measure the association of BORC 4 components with each other in Brucella infected or uninfected host cells. We found that in uninfected 5 cells, BLOS1 interacted with protein components of BLOC-1 (PALLIDIN), BORC (KXD1), and 6 both BLOC-1 and BORC (BLOS2, SNAPIN) ( Figure 7A     the hypothesis that MHV infection of host cells activates RIDD activity. To test this hypothesis, 7 mBlos1 or control host cells were untreated or treated with 4µ8C. Next, these cells were infected with 8 MHV for 24 hr. Virus plaque-forming units (PFU) and host Blos1 expression were then measured. 9 We found that viral PFUs were reduced in 4µ8C-treated cells. However, significantly increased PFU 10 in mBlos1 cells at 24 h.p.i. compared to controls was observed ( Figure 7H). Expression levels of 11 Blos1 mRNA were dramatically reduced during infection ( Figure 7I). Collectively, these findings 12 suggested that coronavirus MHV, like Brucella, subverts the host RIDD pathway to promote 13 intracellular infection. 14 15 Discussion 16 RIDD, a fundamental component of UPR in eukaryotic cells, cleaves a cohort of mRNAs encoding 17 polypeptides that influence ER stress, thereby supporting the maintenance of ER homeostasis. In this 18 report, we found that Brucella infection subverts UPR, in general (Pandey et al., 2018;Qin et al., 19 2008;Smith et al., 2013;Taguchi et al., 2015), and RIDD activity on Blos1, in particular, to promote 20 intracellular parasitism. BLOS1, encoded by RIDD gene Blos1, is a shared subunit of both BLOC-1 21 and BORC complexes (Pu et al., 2015). Mutation or a reduction in BLOS1 expression affects both of BCVs decorated with ER proteins increases due to the fusion of BCVs with ER membranes and/or 6 with noncanonical macrophagosomes (Pandey et al., 2018;Starr et al., 2012;Taguchi et al., 2015). 7 These final events support the intracellular replication and cell-to-cell movement of the pathogen. (i.e., ARL8b, KIF1b, and mTORC1) with BCVs or lysosomes in cells expressing Rr-Blos1 variants 5 were maintained at a relatively higher level (Figures 5-6). These findings demonstrate that blocking 6 BORC function via the disassembly of the BORC complex through depletion of Blos1 by Brucella 7 infection drives BCVs towards the perinuclear region and ER accumulation, which likely facilitates 8 the fusion of BCVs with the ER, thereby supporting intracellular parasitism. 9 Finally, RIDD-mediated Blos1 degradation may promote BCV fusion with autophagosomes. RIDD pathway promotes BCV perinuclear or ER-region clustering, and may also avoid the peripheral 27 movement of BCVs away from the ER region as a consequence of reduced a-tubulin acetylation. 28 These processes may facilitate BCV fusion with ER membranes or (macro)phagosomes, promote the 29 enlargement of aBCVs and further bacterial replication, and ultimately relieve Brucella induced ER 1 stress (Pandey et al., 2018;Qin et al., 2008;Starr et al., 2012;Taguchi et al., 2015). 2 In addition to Brucella, the betacoronavirus MHV also subverts the RIDD-BLOS1 axis to promote 3 intracellular replication (Figure 7H-I), thereby indicating that RIDD control of BLOS1 activity is 4 not pathogen-specific. How the host RIDD-Blos1 axis regulates interactions between host cells and 5 coronaviruses merits further investigation. However, additional possibilities for regulatory control 6 can be envisioned. First, coronaviruses utilize many proteins such as nsp1 to inhibit host protein 7 synthesis in the first 6 hr of infection (Nakagawa & Makino, 2021). Second, BLOS1 contains a 8 potential coronavirus 3C-like protease cleavage site, LQ^SAPS, near its C-terminus, thereby 9 rendering it potentially susceptible to direct subversion by coronaviral pathogens. Finally, 10 coronaviruses have evolved to subvert host interferon defenses (Thoms et al, 2020), which may 11 contribute to immune evasion. Future work will be directed toward examining these possibilities and 12 the roles and mechanisms by which the RIDD-Blos1 axis controls these and other host-pathogen 13 interactions. 14 15

Bacterial strains, cell culture, Brucella infection and antibiotic protection assays 18
Brucella melitensis strain 16M (WT), and B. abortus strain 2308 (WT), and B. abortus vaccine strain 19 S19 and other bacterial strains were used in this work. Bacteria were grown in tryptic soy broth (TSB) 20 or on tryptic soy agar (TSA, Difco™) plates, supplemented with either kanamycin (Km, 50 μg/ml) 21 or chloramphenicol (Cm, 25 μg/ml) when required. For infection, 4 ml of TSB was inoculated with 22 a loop of bacteria taken from a single colony grown on a freshly streaked TSA plate. Cultures were 23 then grown with shaking at 37°C overnight, or until OD 600 ≈3.0. 24

Mammalian host cells including murine macrophages RAW264.7 and its derived non-functional and 25
Rr-Blos1 variants and corresponding control cells, BMDMs, J774.A1 cells, MEFs and were routinely 26 cultured at 37°C in a 5% CO 2 atmosphere in Dulbecco's Modified Eagle's Medium (DMEM) 27 supplemented with 10% fetal bovine serum (FBS). Murine osteoblasts MC3T3-E1 and its derived 28 Rr-Blos1 variant and corresponding control cells (generously provided by the Hollien Lab) were 29 routinely cultured at 37˚C in a 5% CO 2 atmosphere in alpha minimum essential media (MEMα) with 30 nucleosides, L-glutamine, and no ascorbic acids, supplemented with 10% FBS. Murine fibroblasts 1 L2 cells were routinely cultured at 37˚C in a 5% CO 2 atmosphere in F12 medium supplemented with 2 10% fetal calf serum (FCS). For BMDMs, the above-mentioned DMEM with 20% L929 cell 3 supernatant, 10% FBS, and antibiotics was used. Cells were seeded in 24-well or 96-well plates and 4 cultured overnight before infection. For antibiotic protection assays, 1.25×10 5 (BMDMs) or 5 2.5×10 5 (RAW264.7) host cells were seeded in each well; for fluorescence microscopy assays, 1x10 4 6 or 5×10 4 cells were seeded in 96-well plates or on 12-mm glass coverslips (Fisherbrand) placed on 7 the bottom of 24-well microtiter plates respectively; for host RNA analysis, 1x10 5 host cells were 8 seeded in each well of 24-well plates before infection. Host cells were infected with Brucella at an 9 MOI of 100, unless otherwise indicated. Infected cells were then centrifugated for 5 min (200 × g) 10 and incubated at 37°C. Thirty minutes to 1 hr post-infection, culture media was removed, and the 11 cells were rinsed with 1 × phosphate buffered saline (PBS, pH 7.4). Fresh media supplemented with 12 50 μg/ml gentamicin was then added for 1 hr to kill extracellular bacteria. Infected cells were 13 continuously incubated in the antibiotic. At the indicated time points post infection, viable bacteria 14 in infected cells were analyzed using the antibiotic protection assay or the immunofluorescence 15 microscopy assay as previously described (Pandey et al., 2018;Qin et al., 2008). were infected with MHV-A59 in triplicate at a MOI of 1. Infected cells were incubated at room 20 temperature with gentle rocking for 1 hour. Afterwards, culture media was removed, and the cells 21 were rinsed with 1´ PBS (pH 7.4). Fresh media supplemented with 2% FBS was added. Infected were washed away, and the retained cells were propagated in fresh L929-cell conditioned media for 12 another 4 days. BMDMs were split in 24-well plates (2.5×10 5 cells/well) in L929-cell conditioned 13 media and cultured at 37°C with 5% CO2 overnight before use. 14 15

Whole animal infections with Brucella and tissue analysis 16
Mice from CKO and littermate control groups were intranasally infected with B. melitensis and B. 17 abortus (Bm16M and BaS2038, respectively) with a dose of 1×10 6 CFU. At 7 and 14 dpi, infected 18 mice were euthanized, and the bacterial burden was assessed in spleen and liver. A portion of the 19 tissue was fixed, and paraffin embedded for histopathological examination following H&E staining. 20 To assess Bm16M tissue burden, spleen or liver tissues were homogenized and subjected to a serial 21 dilution. Finally, the diluted tissue homogenates (200 µl) were plated on TSA solid plates and CFUs 22 were determined at 48 to 60 hr post incubation at 37°C in 5% CO 2 . 23 24

Latex bead phagocytosis assays 25
Phagocytosis assays for testing the phagocytic uptake and route of a substrate in the non-functional 26 and RIDD resistant Blos1 variants in RAW264.7 murine macrophages were performed using the Complementary DNA was amplified from mRNA using the High-Capacity cDNA Reverse 2 Transcription Kit (4368813 Applied Biosystems) per manufacturer's guidelines. For qRT-PCR of 3 the macrophage infections of MHV strain A59, BaS19, Bm16M, or Ba2308 1/5 dilution of each 4 cDNA was added into nuclease free water in a respective well in a 96-well plate. SYBR green (50 -5 90 μl) with primers (5 -9 μl) were put into triplicate wells of each respective primer in the same 6 respective 96-well plate for all experiments. Primers for Blos1 and GAPDH were used (Table S2). 7 The cDNA and master mix were transferred to a 384-well plate using E1 ClipTip pipettor (4672040 8 Thermo Scientific). The qPR-PCR was run on a CFX384 TM Real-Time System (Bio-Rad). 9 10

RNA-seq analysis 11
All RNA-seq reads were mapped to Mus musculus reference genome GRCm38.p4, release 84, which 12 is provided by Ensembl.org, by using STAR-2.5.2a (Dobin et al, 2012). The aligned reads were then 13 Non-functional Blos1 variants in RAW264.7 murine macrophages were generated using a protocol 6 previously described (Hoffpauir et al, 2020). One clone containing either an amino acid deletion 7 substitution or deletion in one of the XAT regions of murine BLOS1 (Figure S4E) was selected. For 8 generation of Rr-Blos1 variant and its control (wt-Blos1) in RAW264.7 murine macrophages, RNA 9 was first extracted using RNeasy Plus mini-kit. Blos1 cDNA was generated from mRNA and cloned 10 into pCR™ 2.1-TOPO vector. Site-directed mutagenesis was utilized to generate a g449t mutation 11 into one of the Blos1 plasmid clones ( Figure S4F). Both the wild-type (WT) and mutated Blos1 12 segments were removed from the pCR™ 2.1-TOPO vector and cloned into pE2n vectors. Gateway

Drug treatments 21
Host cells were coincubated in 24-well plates with tunicamycin or 4µ8C at the indicated 22 concentrations. Cells were treated with drugs 1 hr before, and during, infection with the 23 indicated Brucella strains and incubated at 37°C with 5% CO 2 . At the indicated time points post 24 infection, the treated cells were fixed with 4% formaldehyde and stained for immunofluorescence 25 analysis or lysed to perform CFU assay or for RNA extraction assays as described above. To 26 investigate whether the drugs inhibit Brucella growth, the drugs were individually added 27 to Brucella TSB cultures at 37°C and incubated for 1 and 72 hr. CFU plating was used to assess 28 bacterial growth in the presence of drugs, and thereby to evaluate the potential inhibitory effects. Host 29 cells in which drug treatment or Brucella infection induced no significant differences in viability and 30 membrane permeability as well as drugs that have no adversary effect on Brucella growth were used 1 in the experiments reported in this work.  Table. Samples were stained with Alexa Fluor 488-conjugated, 17 Alexa Fluor 594-conjugated, and/or Alexa Fluor 647 secondary antibody (Invitrogen/Molecular 18 Probes, 1:1,000). Acquisition of confocal images, and image processing and analyses were performed 19 as previously described (Pandey et al., 2017;Pandey et al., 2018;Qin et al., 2011;Qin et al., 2008). 20 The BioTek Cytation 5 and Gen 5 software (version 3.05) were used to calculate perinuclear Lamp-21 1 index and autophagic flux. Specifically, for perinuclear Lamp-1 index, the area of the nucleus and 22 average intensity for Lamp1 was measured, noted as AN and IntN, respectively. The area of the whole 23 cell and the average intensity for Lamp1 was measured, noted as AWC and IntWC, respectively. Next 24 the area of the cell 3 micrometers off from the nucleus and the average intensity for Lamp1 was 25 measured, noted as AC and IntC, respectively. Then, IntWCN 26 (Average intensity of Lamp1 in the whole cell minus nucleus) and IntP (Average intensity of Lamp1 27 in the perinuclear region) were calculated as the following formula: If IntP/IntWCN ≤ 1, which means that there is no or very little perinuclear colocalization; If 1 IntP/IntWCN > 1, which means that there is perinuclear colocalization. 2 For autophagic activity, using BioTek Cytation 5 and Gen 5 software (version 3.05), the mean 3 intensity of LC3b in the nucleus (M1) and in the cytoplasm (M2) were measured using a primary and 4 secondary mask in every individual cell. In the Gen 5 software, a subpopulation analysis was carried 5 out to identify cells that had a ratio of M2/M1 <1. From this, the perinuclear LC3b index or 6 autophagic activity was calculated as the following formula: Perinuclear LC3b index (or autophagic 7 activity) = (Total number of identified cells with M2/M1 <1)/Total number of the analyzed cells. 8 9 For calculation of the BCV-BLOS1 + index, the corrected total cell fluorescence was measured by 10 taking the integrated density (area of cells ´ mean fluorescence), then the number of Brucella in each 11 cell and colocalization of BLOS1 and Bm16M (BCV-BLOS1 + , %) were counted. Cells that did not 12 contain bacteria were removed from the calculation. The integrated density was divided by the 13 number of bacteria in the cell to obtain the "Total fluorescence per bacteria". The "BCV-BLOS1 + 14 index" was then calculated by multiplying the "Total fluorescence per bacteria" with the 15 colocalization percentage. Since BLOS1 protein is more stable in control cells treated with 4µ8C, the 16 value of BCV-BLOS1 + index was normalized as 100%. 17 18

Protein pull-down assays and immunoblotting analysis 19
Pull-down assays for testing physical interaction of proteins were perform using the Pierce Co-20 Immunoprecipitation (Co-IP) Kit (Thermo Scientific, USA) according to the manufacturer's 21 instructions. Preparation of protein samples and immunoblotting blot analysis were performed as 22 described previously (Ding et al, 2021;Pandey et al., 2017;Pandey et al., 2018;Qin et al., 2011). 23 Densitometry of blots was performed using the ImageJ (http://rsbweb.nih.gov/ij/) software package. 24 All Westerns were performed in triplicate and representative findings are shown. 25 26

Statistical analysis 27
All the quantitative data represent the mean ± standard error of mean (SEM) from at least three 28 biologically independent experiments, unless otherwise indicated. The data from controls were 29 normalized as 1 or 100% to easily compare results from different independent experiments. The               immunoblotting assays or fixed and subjected to confocal immunofluo-rescence assays. Blots/images 20 are representative of three independent experiments. Statistical data represent the mean ± SEM from 21 three independent experiment. *, p < 0.05; **, p < 0.01; ***, p < 0.001.        infected with or without Bm16M, and at the indicated h.p.i., the cells were fixed and subjected to 10 confocal immunofluorescence assays. Images are representative of three independent experiments. 11 Statistical data represent means ± SEM from three independent experiments. *, p < 0.05; **, p < 0.01; 12 ***, p < 0.001.