The loss of the PDIM/PGL virulence lipids causes differential secretion of ESX-1 substrates inMycobacterium marinum

14 The mycobacterial cell envelope is a major virulence determinant in pathogenic mycobacteria. 15 Specific outer lipids play roles in pathogenesis, modulating the immune system and promoting 16 the secretion of virulence factors. ESX-1 (ESAT-6 system-1) is a conserved protein secretion 17 system required for mycobacterial pathogenesis (1, 2). Previous studies revealed that mycobac-18 terial strains lacking the outer lipid PDIM have impaired ESX-1 function during laboratory growth 19 and infection (3-5). The mechanisms underlying changes in ESX-1 function are unkno wn. We 20 used a proteo-genetic approach to measure PDIM and PGL-dependent protein secretion in M. 21 marinum, a non-tubercular mycobacterial pathogen that causes tuberculosis-like disease in e c-22 tothermic animals (6, 7). Importantly, M. marinum is a well-established model for mycobacterial 23 pathogenesis (8, 9) . Our findings showed that M. marinum strains without PDIM and PGL 24 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint showed specific, significant reductions in protein secretion compared to the WT and compl e-25 mented strains. We recently established a hierarchy for the secretion of ESX-1 substrates in 26 four (I-IV) groups (10). Loss of PDIM differentially impacted secretion of Groups III and IV ESX-27 1 substrates, which are likely the effectors of pathogenesis. Our data suggests that the altered 28 secretion of specific ESX-1 substrates is responsible for the observed ESX-1-related effects in 29 PDIM-deficient strains. 30 31 Importance: Mycobacterium tuberculosis, the cause of human tuberculosis, killed an estimated 32 1.3 million people in 2022. Non-tubercular mycobacterial species are causing acute and chronic 33 human infections. Understanding how these bacteria cause disease is critical. Lipids in the cell 34 envelope are essential for mycobacteria to interact with the host and promote disease. Strains 35 lacking outer lipids are attenuated for infection , but the reasons are unclear . Our research aims 36 to identify a mechanism for attenuation of mycobacterial strains without the PDIM and PGL 37 outer lipids in M. marinum. These findings will enhance our understanding of the importance of 38 lipids in pathogenesis, and how these lipids contribute to other established virulence mech a-39 nisms. 40

41 Phthiocerol dimycocerosate (PDIM) and phenolic glycolipid (PGL) are outer lipid virulence fa c-42 tors in the mycolate outer membrane (MOM) (11-13) . Although the precise role of PDIM/PGL in 43 mycobacterial virulence remains elusive, PDIM/PGL support cell envelope structure, serve as a 44 permeability barrier, enhance antibiotic tolerance, and directly modulate the immune system 45 (14-18). These lipids recruit macrophages permissive for bacterial growth and survival while 46 promoting evasion of bactericidal macrophages (11, 13, 17, 19-22). 47 PDIM and PGL are synthesized from cytoplasmic fatty acids via a series of discrete 48 steps by the PpsA-E/FadD26, Mas/FadD28 and PapA5 enzymes (23-25). Their transport across 49 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint the cytoplasmic membrane ( CM) and mycobacterial arabinogalactan peptidoglycan complex 50 (mAGP) relies on the MmpL7 and DrrABC transporters, and the LppX lipoprotein (11, 13, 26-51 29), although the precise mechanisms remain unknown. 52 Mycobacteria secrete proteins across the MOM to the cell surface and into the surroun d-53 ing environment both during laboratory growth and infection (30-34) . Proteins are exported 54 across the CM by the Sec and SecA2 pathways, twin arginine transporter (TAT), and ESAT- 6 55 (ESX) systems (35, 36). Yet, the mechanisms used by proteins to cross the impermeable 56 mAGP and the MOM (37-39) are elusive. Recent studies suggest that PE/PPE and Esx 57 (WXG_100) proteins facilitate protein translocation across the MOM (10, 40, 41). 58 Early during infection, mycobacterial secreted proteins promote interaction with host 59 membranes and macrophage signaling pathways. The ESX-1 system secretes proteins that 60 contribute to phagosomal damage (42-44) , leading to pathogen release, along with its DNA, 61 RNA and secreted proteins, into the macrophage cytoplasm (22, 44-46) . Secreted proteins in 62 the cytoplasm hinder phagosomal repair , affect host trafficking and promote dissemination of 63 infected macrophages in the lung (44, 47-52) . Mycobacterial DNA in the cytoplasm triggers the 64 Type I IFN response (43, 45, 46, 53) , promoting macrophage cell death and bacterial spread 65 (43, 54). 66 In the absence of PDIM, ESX-1 function is reduced , affecting the Type I IFN response 67 and phagosomal lysis (3, 4). It is unclear whether PDIM is required for ESX-1 protein secretion 68 across the MOM (3) or for ESX-1 substrate function within the phagosome (4) . Our study uses a 69 proteogenetic approach (55) to quantify how PDIM and PGL influence protein secretion from 70 Mycobacterium marinum. Our studies provide an explanation for the ESX-1 phenotypes in the 71 absence of PDIM, suggesting a broader role of PDIM and PGL outer lipids in translocating m y-72 cobacterial proteins across the MOM. 73

74 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint Generation and characterization of PDIM deficient M. marinum strains. The PDIM/PGL b i-75 osynthetic pathway is conserved in both M. tuberculosis and M. marinum (Figure 1A). We previ-76 ously constructed M. marinum strains with disrupted PDIM production due to deletion of the mas 77 gene (56) or the second ER domain of the ppsC gene (57). We generated an M. marinum strain 78 with an unmarked deletion of the drrABC genes. PpsA-E/FadD26 synthesize phthiocerol; loss of 79 the PpsCER domain ablates activity and abrogates PDIM production (19, 57-59) . Mas/FadD28 80 synthesize mycocerosic acid; deletion of mas halts PDIM production (12, 23). The two PDIM 81 components are joined by PapA5 and transported by DrrABC, MmpL7 and LppX (12, 25-27) . 82 Deletion of drrABC abolishes PDIM production and transport (12) . Complementation strains 83 were generated by introducing integrating plasmids with the ppsCER domain, the drrABC operon 84 or the mas gene behind constitutive promoters into each deletion strain. We confirmed the re-85 sulting strains using PCR (Fig. S1) followed by targeted DNA sequencing, and qRT-PCR to 86 measure transcription in the deletion and complementation strains (Fig. 1B and Fig. S2A and B). 87 Deletion of the mas or drrABC genes abolished expression relative to the WT and complemen t-88 ed strains. Deletion of the ppsCER domain did not impact ppsC expression, as expected (57). 89 We validated PDIM and PGL loss in each deletion strain (Fig. 1C and Fig. S2C) by is o-90 lating total soluble lipids and visualiz ing PDIM and PGL using thin layer chromatography . All 91 three M. marinum deletion strains exhibited a lack of PDIM and PGL (Fig. 1C and Fig S2C). Re-92 introducing the mas gene restored PDIM in the Δmas strain (Fig. 1C) . PDIM and PGL were not 93 complemented by restoring gene expression in the ΔppsCER and ΔdrrABC strains (Fig. S2C). 94 95 The loss of PDIM/PGL has a general effect on mycobacterial protein secretion. 96 Past studies linked ESX-1 secretion and PDIM/PGL, proposing that PDIM is necessary 97 for ESX-1 protein secretion (3) . To clarify PDIM/PGL ’s role in mycobacterial protein secretion , 98 we isolated secreted and cell associated protein fractions and quantified the protein levels in the 99 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint mutant and complemented strains compared to the WT M. marinum strain using Label Free 100 Quantitative (LFQ) Proteomics (Dataset S1). We measured ESX-1-dependent protein secretion 101 as a control. 102 Deletion of the eccCb1 gene reduced ESX-1 substrate production and secretion (Fig. 2, 103 top). We reasoned that if PDIM was specifically required for ESX-1 secretion, then deletion of 104 the mas, ppsC ER or drrABC genes will abolish ESX-1 secretion, similar to the Δ eccCb1 strain. 105 Deletion of the mas gene (or deletion of the drrABC genes or the ER domain from the ppsC 106 gene) resulted in a significant decrease in the levels of most secreted proteins (Fig. 2, middle 107 and Fig. S3). Restoring mas expression in the Δmas strain complement ed the secretion defect 108 (Fig. 2, bottom). These data support that PDIM is important for overall protein secretion from M. 109 marinum during laboratory growth. 110 111 The loss of PDIM/PGL specifically impacts the secretion of a subset of ESX-1 substrates. 112 We previously demonstrated that ESX-1 substrates are hierarchically secreted from M. 113 marinum. We categorized them into at least four groups of ESX-1 substrates (Fig. 3A, (10)). 114 Group 1 includes EsxA, EsxB, PPE68 and MMAR_2894, which are essential for the secretion of 115 the other ESX-1 substrates (10, 60, 61) . Group 2 substrates include EspB, EspJ and EspK (10, 116 60, 62). These substrates are not required for the secretion of Group 1 substrates, but are e s-117 sential for the secretion of ESX-1 Group 2 substrates (10). ESX-1 Group 3 substrates include 118 EspE and EspF, and these substrates are dispensable for the secretion of the other ESX-1 su b-119 strate classes. Importantly, deletion of the Group 3 substrates is sufficient to abrogate the h e-120 molytic activity of M. marinum (63). 121 Figure 3B and 3C indicates that the loss of PDIM specifically impacts the secretion of 122 ESX-1 Group 3 substrates . In the ΔeccCb1 strain, ESX-1 substrate levels were significantly r e-123 duced (Figure S 4, Dataset S1 ) due to feedback control of substrate gene transcription in the 124 absence of the ESX -1 system (64). However, in the Δmas strain, the substrate levels were not 125 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint significantly reduced (Fig. S4), indicating no feedback control of substrate gene transcription 126 due to the loss of the PDIM/PGL lipids. While ESX-1 substrate secretion was significantly r e-127 duced from the ΔeccCb1 strain because EccCb 1 is directly required for substrate secretio n (1, 128 65, 66) , the levels of ESX-1 substrate s secreted from the Δmas strain differed, aligning with 129 previously established substrate groups (Fig. 3B and C , Dataset S1 ). The Group 3 substrates 130 (EspE, EspF) were the most affected by PDIM /PGL loss. The reduced secretion of Group 3 131 substrates were complemented by expression of mas in the Δmas strain (Fig. 3B and 3C) . To-132 gether, these data support that the loss of PDIM/ PGL differentially impacts the secretion of the 133 known ESX-1 substrates. 134 135 The absence of PDIM reduces ESX-1 function. 136 Prior research demonstrated that PDIM-deficient M. marinum strain exhibited lower 137 ESX-1-dependent hemolytic activity (5). EspE and EspF secretion were most impacted by 138 PDIM/PGL loss. Deleting espE or espF from M. marinum abrogated hemolytic activity and a t-139 tenuated M. marinum during macrophage infection (63). Because deleting espE or espF did not 140 significantly affect the secretion of the additional known ESX-1 substrates, EspE and EspF are 141 at the top of the hierarchy (10) . In Figure 4A, WT M. marinum lysed sheep red blood cells 142 (sRBCs) in a contact dependent, ESX-1 dependent manner. The ΔeccCb1 strain which lacks 143 ESX-1 secretion, had significantly less hemolytic activity than the WT strain ( P<.0001) (10, 61, 144 63). The Δmas strain exhibited intermediate hemolytic activity, between the WT ( P<.0001) and 145 ΔeccCb1 strains (P<.0001), similar to a prior report (5). Reintroducing the mas gene in the Δmas 146 strain significantly increased hemolysis ( P<.0001), but not to WT levels (WT vs Δ mas/pmas, 147 P<.0001). 148 We next used western blot analysis to examine ESX-1 substrate secretion . EsxB (CFP-149 10) was made (Fig. 4B, lane 1) and secreted (lane 4) from the WT strain. EsxB production was 150 reduced (lane 2) and it was not secreted (lane 6) from the ΔeccCb1 strain, as previously repor t-151 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint ed (57). EsxB production (lanes 3 and 4) and secretion (lanes 7 and 8) were not noticeably i m-152 pacted in the Δmas and Δmas complemented strains. EspE was produced (lane 1) and secre t-153 ed (lane 5) from the WT strain. EspE levels were reduced in (lane 2) and was not secreted from 154 the ΔeccCb1 strain (lane 6). Although EspE was made to WT levels in the Δmas strain (lane 3), 155 EspE secretion was reduced (lane 7). EspE secretion was restored in Δmas/pmas strain (lane 156 8). These data are consistent with the LFQ proteomics data in Figures 2 and 3. Interestingly, as 157 we replicated the western blot analysis, we sometimes observed no visible secretion phenotype, 158 as shown in Figure 4, right. The LFQ analysis did not reveal WT levels of EspE and EspF in any 159 of the replicates we tested. From these data, we conclude that western blot analysis , which is 160 not quantitative, may lack the sensitivity to detect intermediate but biologically relevant changes 161 in ESX-1 protein secretion, as supported by our prior work (10, 60). 162 163

164 This study aimed to understand the impact of the outer lipids PDIM and PGL on ESX-1 secre-165 tion. We measured a broad and significant reduction in M. marinum protein secretion in the ab-166 sence of PDIM and PGLs. Interestingly, the loss of PDIM and PGLs differentially impacted ESX-167 1 protein secretion, particularly affecting the secretion of the Group 3 substrates, EspE and 168 EspF. Our findings support that the ESX-1-associated phenotypes in strains lacking PDIM/ PGL 169 are due to the reduced secretion of specific ESX-1 substrates. 170 At least three publications previously linked PDIM to ESX -1 secretion or function (3-5). 171 Barczak et al. concluded PDIM is required for ESX-1 substrate secretion in M. tuberculosis. De-172 letion of drrC abrogated EsxA (ESAT -6) and EsxB (CFP -10) secretion , and the ESX -5 su b-173 strate, EspN, during laboratory growth. Other PDIM deficient strains in this study had WT-levels 174 of EsxB secretion. The loss of ESX-1 secretion led to reduced IFN induction during macrophage 175 infection with the ΔdrrC strain as compared to the WT and complemented strains. IFN induction 176 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint requires ESX-1 induced lysis of the phagosome membrane (44, 53). They concluded that the 177 PDIM lipid at the cell surface was important for ESX-mediated protein secretion (3). 178 In Barczak et al. an intermediate secretion signal may have been missed because the 179 amount of protein loaded in that study was based on the culture density as measured by 180 OD600(3). We normalized the protein levels based on protein concentration as measured by 181 BCA assay, because the OD 600 of M. marinum grown Sauton’s media with low Tween-80 is an 182 unreliable measure in our hands. Using this approach, we did not observe significant differences 183 in the secretion of the Group 1 substrates, which includes EsxB, by western blot analysis or u s-184 ing LFQ proteomics. Importantly, Barczak et al. observed variability in EsxB secretion by diffe r-185 ent PDIM deficient strains ( ΔdrrC vs ΔdrrA). Our study revealed that reduced EspE secretion 186 was variable when detected by western blot analysis, but not by proteomics. We think the i n-187 consistency in measuring these secretory phenotypes by western blot likely stems from their 188 intermediate nature. 189 Quigley et al. connected PDIM transport to ESX-1 dependent events during macrophage 190 infection, including phagosome lysis and the resulting macrophage cell death (4) . They showed 191 significantly reduced phagosomal lysis of an ΔmmpL7 M. tuberculosis strain compared to the 192 WT and complemented strains. Aligned with our studies, Quigley et al. loaded 10μg of protein, 193 found no defect in EsxA secretion by western blot analysis. 194 Osman et al. demonstrated that PDIM transport was important for phagosomal lysis , 195 finding reduced phagosomal lysis of an ΔmmpL7 M. marinum strain during macrophage infec-196 tion (5). Although t he ΔmmpL7 strain exhibited intermediate hemolytic activity between the 197 ΔRD1 and WT strains, their data did not meet statistical significance. They suggested that PDIM 198 enhanced the ESX-1 lytic activity, but offered that PDIM may contribute differently to hemolysis 199 and phagosomal lysis . Our results show that M. marinum strains lacking PDIM/PGL production 200 (Δmas or ΔppsC) or transport ( ΔdrrABC) have sign ificantly reduced hemolysis compared the 201 WT strain, and significantly higher hemolysis compared to the ΔeccCb1 M. marinum strain. We 202 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint assessed hemolysis 90 minutes after initiating contact between M. marinum and the sRBCs , 203 while Osman et al measured hemolysis after two hours (5). We previously found intermediate 204 hemolysis phenotypes are most apparent at earlier timepoints (64, 67). We suspect that the i n-205 termediate hemolytic activity of the ΔmmpL7 strain would have reached significance relative to 206 WT strain at earlier timepoints. 207 An underlying commonality in all of these studies is that in strains lacking PDIM or PGL, 208 ESX-1 activity is reduced, as reflected in hemolytic activity or phagosomal lysis during infection. 209 Notably, at the time of these manuscripts, the secretory relationship between ESX-1 substrates 210 was unclear. However, we now know that EspE/EspF secretion is essential for hemolysis, and 211 for pathogenesis in macrophages (63) . We have likewise defined a hierarchical secretory rel a-212 tionship between the ESX-1 substrates (10) . Here we showed that the loss of PDIM specifically 213 reduced EspE/EspF secretion from M. marinum . Therefore, we hypothesize that reduced Es-214 pE/EspF secretion caused the reduced hemolytic and phagolytic activity observed in the a b-215 sence of PDIM. It is tempting to speculate that the secretion of proteins that colocalize with 216 PDIM and PGL in the mycobacterial envelope is impacted by the loss of these lipids in M. mari-217 num. This idea requires further exploration beyond these initial studies. 218 Our unbiased approach to measuring protein secretion revealed that losing PDIM/PGL 219 broadly impacts protein secretion. Contrary to expectations based on PDIM ’s role in maintaining 220 envelope impermeability, making the cell envelope permeable did not increase protein secretion 221 by increasing cellular lysis. Instead, overall protein secretion in the culture supernatants d e-222 creased, despite evidence of cellular lysis. Our study has direct implications for mycobacterial 223 pathogenesis, as PDIM and PGL deficient Mycobacterium are attenuated in cellular infection 224 models (11, 13, 58) . Further study is needed to understand if the observed phenotypes due to 225 the loss of mycobacterial outer lipids stem from changes in the secreted proteome. 226 227

228 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint Growth of Bacterial Strains: Mycobacterium marinum strains were derived from Mycobacte-229 rium marinum M strain (ATCC BAA-535). M. marinum strains were maintained at 30°C in either 230 Middlebrook 7H9 Broth Base (Sigma Aldrich) supplemented with 0.5% glycerol and 0.1% 231 Tween-80 (Fisher Scientific) or Middlebrook 7H11 Agar (Sigma Aldrich) plates supplemented 232 with 0.5% glucose and 0.5% glycerol. Where appropriate, strains were supplemented with 50 233 μg/mL hygromycin B (Corning, Corning, NY), 20 μg/mL kanamycin (IBI Scientific, Dubuque, IA), 234 and 60 μg/mL X-Gal (RPI, Mount Prospect, IL). Escherichia coli DH5α strains were grown in Lu-235 ria-Bertani Broth (VWR) at 37°C and supplemented with 200 μg/mL hygromycin and 50 μg/mL 236 kanamycin, where appropriate. 237 238 Construction of M. marinum strains: The ΔdrrABC strain was generated using allelic e x-239 change using the p2NIL vector [Addgene plasmid number 20188; a gift from Tanya Parish (68)] 240 and the pGOAL19 vector [Addgene plasmid number 20190; a gift from Tanya Parish (68)], as 241 previously published (61, 63, 64, 67, 69) . Primers were purchased from Integrated DNA Tec h-242 nologies (IDT, Coralville, IA) and are listed in Table S2 . The complementation plasmids were 243 generated using FastCloning (70) using M. marinum M genomic DNA. Plasmids were intr o-244 duced into M. marinum using electroporation as described previously (61, 63, 64, 67, 69, 71) . 245 Strains and plasmids are listed in Supplementary Tables S1. Plasmids and genetic deletions 246 were confirmed by targeted DNA sequencing performed by the Notre Dame Genomics and Bi o-247 informatics Facility. 248 249 Thin-Layer Chromatography of PDIM and PGL Lipids: 5 mL cultures of M. marinum in 7H9 250 defined broth (Millipore Sigma), and grown for 2-3 days until turbid, diluted into 50mL of 7H9 251 defined broth to an OD 600 of 0.05 and grown for 72 hours. Mycobacterial cells were harvested via 252 centrifugation (4000 rpm for 10 minutes). The resulting cell pellets were washed three times with 253 phosphate-buffered saline (PBS). Lipids were extracted using a chloroform (Ricca Chemical 254 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint Company, Arlington, TX) :methanol (Millipore Sigma) mixture (2:1) overnight in a fume hood. 255 Overnight extraction mixtures were filtered through 6 mm Whatman filter paper (Millipore Sigma) 256 using a porcelain Buchner funnel before separating aqueous and organic phase layers. Extrac t-257 ed lipids were moved to a clean glass vial and dried overnight in a fume hood under a gentle 258 stream of air. Dried mycobacterial lipids were resuspended in a 100 μl mixture of chloroform: 259 methanol (2:1) prior to analysis. 260 9 μl of total mycobacterial lipids were spotted onto aluminum -backed silica TLC plates 261 (Millipore Sigma) and dried. To separate mycobacterial phthiocerol dimycocerosates (PDIM) 262 and triacylglycerols (TAG), spotted lipids were migrated three times sequentially in a running 263 solution of 50:1 petroleum ether (Sigma Aldrich) and ethyl acetate (Sigma Aldrich). The TLC 264 plate was sprayed with phosphomolybdic acid (Millipore Sigma) and charred via heat gun 265 (Westward) to visualize lipids. 266 Mycobacterial phenolic glycolipids (PGLs) were visualized using 9 μl of total mycobact e-267 rial lipids migrated once in a solution of 96:4 chloroform and methanol. The TLC plate was 268 sprayed with a solution of 1% α -naphthol (TCI Chemicals, Portland, OR) and charred via heat 269 gun to visualize lipids. 270 271 Hemolysis Assays: Hemolysis assays were performed exactly as described (10) except that 272 sheep red blood cell and bacteria were incubated at 30 °C for one and a half hours, rather than 273 two hours. 274 275 Secretion Assays: Cell-associated and secreted protein fractions were prepared from M. mari-276 num exactly as previously described (10). Briefly, M. marinum bacterial cultures were grown in 277 7H9 defined media, diluted into 50 mL of Sauton’s media at an OD 600 of 0.8 and grown for 48 278 hours before harvesting cellular and secreted fractions by centrifugation. Culture supernatants 279 were filtered through 0.2 μm Nalgene Rapid -Flow Bottle Top Filters with PES membranes 280 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint (Thermo Scientific) and concentrated using 15 mL, 3kDa Amicon Ultra Centrifugal Filter Units 281 (Millipore Sigma). Mycobacterial cell pellets were resuspended in 0.5 mL of PBS, lysed via 282 bead-beating with a BioSpec Mini-Beadbeater-24, and clarified by centrifugation. Protein con-283 centrations of the resulting fractions were measured Pierce Micro BCA Protein Assay Kit (The r-284 mo Scientific). 285 286 Western Blot Analysis: 10 μg of each protein sample was loaded into 4-20% TGX Gradient 287 Gels (Bio-Rad) for analysis. Antibodies were diluted in 5% non-fat dry milk (RPI) in PBS with 288 0.1% Tween-20 (VWR). Rpo β (anti-RNA polymerase beta mouse monoclonal antibody [clone: 289 8RB13]; VWR) was diluted 1 :20,000. The following reagents were obtained through BEI r e-290 sources, NIAID, NIH: polyclonal anti-Mycobacterium tuberculosis CFP-10 (gene Rv3874; antise-291 rum, rabbit; NR -13801) and polyclonal anti -Mycobacterium tuberculosis Mpt-32 (gene Rv1860; 292 antiserum, rab bit; NR -13807). CFP -10 was used at a 1:5,000 dilution. Mpt -32 was used at a 293 1:30,000 dilution. The EspE antibody (1:5,000 dilution) was a custom rabbit polyclonal antibody 294 against the CGQQATLVSDKKEDD peptide (Genscript). 295 296 Proteomics/LC-MS: LC-MS pure reagents (water, ethanol, acetonitrile, and methanol) were 297 purchased from J.T. Baker (Radnor, PA). Iodoacetamide (IAA) was purchased from MP Bi o-298 medicals (Solon. OH). All other reagents are from Sigma-Aldrich (St. Louis, MO) unless spec i-299 fied. S-Trap mini devices were from Protifi (Huntington, NY). Trypsin Gold was purchased from 300 Promega (Madison, WI). Hydrophilic –lipophilic balance (HLB) solid phase extraction (SPE) ca r-301 tridges (1 cc/10 mg) from Waters (Milford, MA) were used to desalt peptide samples prior t o 302 analysis on a timsTOF Pro from Bruker Scientific LLC (Billerica, MA). Protein and peptide ide n-303 tification including label-free quantitation was performed using the PEAKS Online X search e n-304 gine from Bioinformatics Solutions Inc (Waterloo, ON) (build 1.4.2020-10-21_171258). 305 Cell-associated and secreted cell lysate samples were prepared for LC-MS analysis as 306 described (72, 73). 50 µg of each sample was prepared in 100 mM triethylammonium bica r-307 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint bonate (TEAB), 5% sodium dodecyl sulfate (SDS), and 100 mM tris(2-carboxyethyl)phosphine 308 (TCEP). Samples were heated for 10 minutes at 95˚C, cooled, then IAA was added to 100 mM 309 IAA for 30 minutes in dark. 310 Samples were acidified with o-phosphoric acid to a final concentration of 1.2% in 50 µL, 311 then flocculated with 350 µL Binding Buffer containing 90% methanol and 10% 1M TEAB. Sa m-312 ples were passed through S-Trap Micro filters and followed by two 80 µL washes of Binding 313 Buffer and 80 µL of 1:1 methanol/chloroform solution. A new collection tube was added, and 1 314 µg of trypsin in 80 µL was added. Samples were wrapped in Parafilm to prevent evaporative 315 loss and incubated at 37˚C for 12 hours. Digested peptides were spun through the filte r, fol-316 lowed by two 80 µL elutions with 0.1% formic acid in water, and one 80 µL elution with 0.1% 317 formic acid in 50% acetonitrile. Eluted peptides were vacuum concentrated for 20 minutes (to 318 remove acetonitrile), and then desalted using 1cc/10 mg HLB SPE cartridges following man u-319 facturer’s specifications, and dried by vacuum concentrator prior to analysis. 320 Desalted peptides were resuspended in 0.1% formic acid and water, to 1 mg/mL co n-321 centration. 100 ng of each sample was injected in triplicate onto a n Evosep One and timsTOF 322 Pro LC-MS system. Each sample was prepared in biological triplicate and technical triplicate. 323 15spd (Sample-per-day) methods were used on a 150 µm x 150 mm PepSep column with C 18 324 ReproSil AQ stationary phase at 1.9mm particle size, 120Å pore size (Manufacturer’s protocol). 325 Samples were loaded using Evotips following manufacturer’s specifications. Nano-ESI was 326 used with a spray voltage of 1700V. MS was set to Parallel Accumulation, Serial Fragmentation 327 Data Dependent Mode (PASEF-DDA) with a mass range of 100 –1700 m/z, ion mobility range of 328 0.6–1.6 v*s/cm2, and ramp and accumulation times of 100 ms. Each precursor consisted of 10 329 PASEF ramps for a cycle time of 1.17 s. Precursors were filtered to contain only charges from 2 330 to 5. MS/MS collision energy settings were set to ramp from 20 eV at 0.6 ion mobility to 70 eV at 331 1.6 ion mobility. Instrument tune parameters were set to default for proteomic studies with the 332 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint following differences: quadrupole low mass set to 20 m/z, focus pre-TOF pre-pulse storage set 333 to 5 ms. 334 Protein and peptide identification and label-free quantitation were performed using the 335 PEAKS Online X search engine. Search database used was the M. marinum proteome from 336 Mycobrowser (v4 release). Search settings were set to manufacturer defaults unless specified. 337 3 missed cleavages were allowed, at a semi-specific search. Fixed modification included ca r-338 bamidomethylation of cysteine, while variable modifications included protein N-terminal acetyl a-339 tion, deamidation of asparagine and glutamine, pyroglutamic acid from glutamine and glutamic 340 acid, and oxidation of methionine. All peptide-spectrum matches were filtered to a 1% false di s-341 covery rate. Cell-associated samples were normalized by total ion current (TIC). RAW files are 342 deposited and available at MassIVE/Proteome Exchange 343 ftp://[email protected] (password PDIM2024#) MSV000093713. 344 LFQ search results were trimmed as follows. Technical replicate values for each sample 345 were averaged if at least two of the three replicates were nonzero values. The same averaging 346 logic was applied to biological replicates. Only nonzero values were averaged. All mutant sa m-347 ples were ratioed the corresponding WT values, to generate fold-change ratios. 348 RNA extraction: M. marinum strains were grown as described in the Secretion Assays met h-349 ods section. 15 mL of Sauton’s culture were pelleted by centrifugation. Bacterial pellets were 350 resuspended in Qiagen RLT buffer (Qiagen, Hilden, Germany) supplemented with 1% β -351 mercaptoethanol. Lysates were generated by bead beating with the Biospec Mini-BeadBeater-352 24 and clarified with centrifugation. Total RNA was extracted using the RNeasy Mini Kit (Q i-353 agen), according to manufacturer’s instructions. 354 qRT-PCR: RNA (500 ng) was treated with RQ1 DNase (Promega) according to the manufactu r-355 er’s instructions. 1 μl of DNase treat ed RNA was converted to cDNA using random hexamers 356 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint (IDT) and SuperScript II Reverse Transcriptase (Invitrogen) according to manufacturer instru c-357 tions. cDNA was quantified using a Nanodrop 2000 (Thermo Scientific, Waltham, MA). 10 μl 358 qRT-PCR reactions were prepped with 250 ng of cDNA, SYBR TM Green PCR Master Mix (A p-359 plied Biosystems, Carlsbad, CA) and 1 μM of each oligonucleotide. Reactions were run using 360 Applied Biosystems MicroAmp Fast 96-well plates (0.1 ml) (Life Technologies). Plates were run 361 on QuantStudio 3 Real-Time PCR system (Thermo Fisher). Cycle conditions were as follows: 362 50 °C for 2 min., 95 °C for 10 min. then 40 cycles at 95 °C for 15 sec and 60 °C for 1 min. Then, 363 a dissociation step of 95 °C for 15 sec., 60 °C for 1 min., 95 °C for 15 sec., and 60 °C for 15 364 sec. 365 qRT-PCR reactions were analyzed using ΔΔCt comparisons. qRT -PCR results were 366 normalized to WT transcript abundance using the following equations: 367 368 ΔCt = Ct(gene of interest) - Ct(housekeeping gene) 369 Then: 370 ΔΔCt = ΔCt(treated sample) – ΔCt(untreated sample) 371 Then: 372 2-ΔΔCt = fold change 373 374 R Code for Volcano Plots: 375 >WT_vs_dmas$threshold=as.factor(WT_vs_dmas$Significance g <- ggplot(data = WT_vs_dmas, aes(x=L2FC, y = Significance, color = threshold)) + ge-378 om_point(alpha = 0.4, size = 1.75) + xlim(c(-8, 8)) + xlab("log2 fold change") + (ylab("-log10 p-379 value")+ theme_bw() + theme(legend.position = "none")) 380 > g 381 > g +theme(text=element_text(size = 20))+ geom_hline(yintercept = -log10(0.05), linetype = 382 "dashed") + geom_vline(xintercept = c(log2(0.5), log2(2)), linetype = 383 "dashed")+scale_x_continuous(breaks = c(seq(-8,8,2))) 384 385 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint Acknowledgments: K.R.N and B.J. were supported by an Eck Institute for Global Health 386 Fellowship. K.R.N. was also supported by an Arthur J. Schmitt Fellowship. P.A.C. is 387 supported by the National Institutes of Health under award numbers AI156229, 388 AI106872, AI149147, and AI149235. M.M.C. is supported by the National Institutes of 389 Health under award numbers GM144372 and GM139277. We thank the Mass Spe c-390 trometry and Proteomics Facility at Notre Dame. The content of this article is solely our 391 responsibility and does not necessarily represent the official views of the National Inst i-392 tutes of Health. 393 394

395 396 1. S. A. Stanley, S. Raghavan, W. W. Hwang, J. S. Cox, Acute infection and macrophage 397 subversion by Mycobacterium tuberculosis require a specialized secretion system. Proc 398 Natl Acad Sci U S A 100, 13001-13006 (2003). 399 2. T. Hsu et al., The primary mechanism of attenuation of bacillus Calmette-Guerin is a loss 400 of secreted lytic function required for invasion of lung interstitial tissue. Proc Natl Acad 401 Sci U S A 100, 12420-12425 (2003). 402 3. A. K. Barczak et al., Systematic, multiparametric analysis of Mycobacterium tuberculosis 403 intracellular infection offers insight into coordinated virulence. PLoS Pathog 13, 404 e1006363 (2017). 405 4. J. Quigley et al., The Cell Wall Lipid PDIM Contributes to Phagosomal Escape and Host 406 Cell Exit of Mycobacterium tuberculosis. MBio 8 (2017). 407 5. M. M. Osman, A. J. Pagan, J. K. Shanahan, L. Ramakrishnan, Mycobacterium marinum 408 phthiocerol dimycocerosates enhance macrophage phagosomal permeabilization and 409 membrane damage. PLoS One 15, e0233252 (2020). 410 6. F. Linell, A. Norden, Mycobacterium balnei, a new acid-fast bacillus occurring in 411 swimming pools and capable of producing skin lesions in humans. Acta Tuberc Scand 412 Suppl 33, 1-84 (1954). 413 7. H. F. Clark, C. C. Shepard, Effect of Environmental Temperatures on Infection with 414 Mycobacterium Marinum (Balnei) of Mice and a Number of Poikilothermic Species. J 415 Bacteriol 86, 1057-1069 (1963). 416 8. T. C. Pozos, L. Ramakrishnan, New models for the study of Mycobacterium-host 417 interactions. Curr Opin Immunol 16, 499-505 (2004). 418 9. A. E. Chirakos, A. Balaram, W. Conrad, P. A. Champion, Modeling Tubercular ESX-1 419 Secretion Using Mycobacterium marinum. Microbiol Mol Biol Rev 84 (2020). 420 10. R. M. Cronin, M. J. Ferrell, C. W. Cahir, M. M. Champion, P. A. Champion, Proteo-421 genetic analysis reveals clear hierarchy of ESX-1 secretion in Mycobacterium marinum. 422 Proc Natl Acad Sci U S A 119, e2123100119 (2022). 423 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint 11. L. R. Camacho, D. Ensergueix, E. Perez, B. Gicquel, C. Guilhot, Identification of a 424 virulence gene cluster of Mycobacterium tuberculosis by signature-tagged transposon 425 mutagenesis. Mol Microbiol 34, 257-267 (1999). 426 12. L. R. Camacho et al. , Analysis of the phthiocerol dimycocerosate locus of 427 Mycobacterium tuberculosis. Evidence that this lipid is involved in the cell wall 428 permeability barrier. J Biol Chem 276, 19845-19854 (2001). 429 13. J. S. Cox, B. Chen, M. McNeil, W. R. Jacobs, Jr., Complex lipid determines tissue-430 specific replication of Mycobacterium tuberculosis in mice. Nature 402, 79-83 (1999). 431 14. A. M. Block et al. , Mycobacterium tuberculosis Requires the Outer Membrane Lipid 432 Phthiocerol Dimycocerosate for Starvation-Induced Antibiotic Tolerance. mSystems 8, 433 e0069922 (2023). 434 15. M. S. Siegrist, C. R. Bertozzi, Mycobacterial lipid logic. Cell Host Microbe 15, 1-2 (2014). 435 16. S. A. Stanley, J. S. Cox, Host-Pa thogen Interactions During Mycobacterium tuberculosis 436 infections. Curr Top Microbiol Immunol 374, 211-241 (2013). 437 17. C. J. Cambier et al., Mycobacteria manipulate macrophage recruitment through 438 coordinated use of membrane lipids. Nature 505, 218-222 (2014). 439 18. C. Passemar et al. , Multiple deletions in the polyketide synthase gene repertoire of 440 Mycobacterium tuberculosis reveal functional overlap of cell envelope lipids in host-441 pathogen interactions. Cell Microbiol 16, 195-213 (2014). 442 19. C. J. Cambier, S. M. O'Leary, M. P. O'Sullivan, J. Keane, L. Ramakrishnan, Phenolic 443 Glycolipid Facilitates Mycobacterial Escape from Microbicidal Tissue-Resident 444 Macrophages. Immunity 47, 552-565 e554 (2017). 445 20. M. A. Kirksey et al. , Spontaneous phthiocerol dimycocerosate-deficient variants of 446 Mycobacterium tuberculosis are susceptible to gamma interferon-mediated immunity. 447 Infect Immun 79, 2829-2838 (2011). 448 21. T. A. Day et al. , Mycobacterium tuberculosis strains lacking surface lipid phthiocerol 449 dimycocerosate are susceptible to killing by an early innate host response. Infect Immun 450 82, 5214-5222 (2014). 451 22. Y. Cheng, J. S. Schorey, Mycobacterium tuberculosis-induced IFN-beta production 452 requires cytosolic DNA and RNA sensing pathways. J Exp Med 215, 2919-2935 (2018). 453 23. A. K. Azad, T. D. Sirakova, L. M. Rogers, P. E. Kolattukudy, Targeted replacement of the 454 mycocerosic acid synthase gene in Mycobacterium bovis BCG produces a mutant that 455 lacks mycosides. Proc Natl Acad Sci U S A 93, 4787-4792 (1996). 456 24. A. K. Azad, T. D. Sirakova, N. D. Fernandes, P. E. Kolattukudy, Gene knockout reveals 457 a novel gene cluster for the synthesis of a class of cell wall lipids unique to pathogenic 458 mycobacteria. J Biol Chem 272, 16741-16745 (1997). 459 25. O. A. Trivedi et al. , Dissecting the mechanism and assembly of a complex virulence 460 mycobacterial lipid. Mol Cell 17, 631-643 (2005). 461 26. M. Jain, J. S. Cox, Interaction between polyketide synthase and transporter suggests 462 coupled synthesis and export of virulence lipid in M. tuberculosis. PLoS Pathog 1, e2 463 (2005). 464 27. G. Sulzenbacher et al., LppX is a lipoprotein required for the translocation of phthiocerol 465 dimycocerosates to the surface of Mycobacterium tuberculosis. EMBO J 25, 1436-1444 466 (2006). 467 28. M. H. Touchette, J. C. Seeliger, Transport of outer membrane lipids in mycobacteria. 468 Biochim Biophys Acta 10.1016/j.bbalip.2017.01.005 (2017). 469 29. M. H. Touchette et al. , A Screen for Protein-Protein Interactions in Live Mycobacteria 470 Reveals a Functional Link between the Virulence-Associated Lipid Transporter LprG and 471 the Mycolyltransferase Antigen 85A. ACS Infect Dis 3, 336-348 (2017). 472 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint 30. C. Abou-Zeid et al. , The secreted antigens of Mycobacterium tuberculosis and their 473 relationship to those recognized by the available antibodies. J Gen Microbiol 134, 531-474 538 (1988). 475 31. M. G. Sonnenberg, J. T. Belisle, Definition of Mycobacterium tuberculosis culture filtrat e 476 proteins by two-dimensional polyacrylamide gel electrophoresis, N-terminal amino acid 477 sequencing, and electrospray mass spectrometry. Infect Immun 65, 4515-4524 (1997). 478 32. Y. Ge et al. , Top down characterization of secreted proteins from Mycobacterium 479 tuberculosis by electron capture dissociation mass spectrometry. J Am Soc Mass 480 Spectrom 14, 253-261 (2003). 481 33. E. F. Perkowski et al., The EXIT Strategy: an Approach for Identifying Bacterial Proteins 482 Exported during Host Infection. MBio 8 (2017). 483 34. F. Sayes et al. , Multiplexed Quantitation of Intraphagocyte Mycobacterium tuberculosis 484 Secreted Protein Effectors. Cell Rep 23, 1072-1084 (2018). 485 35. L. S. Ligon, J. D. Hayden, M. Braunstein, The ins and outs of Mycobacterium 486 tuberculosis protein export. Tuberculosis (Edinb) 92, 121-132 (2012). 487 36. C. Chagnot, M. A. Zorgani, T. Astruc, M. Desvaux, Proteinaceous determinants of 488 surface colonization in bacteria: bacterial adhesion and biofilm formation from a protein 489 secretion perspective. Front Microbiol 4, 303 (2013). 490 37. P. A. Lambert, Cellular impermeability and uptake of biocides and antibiotics in Gram-491 positive bacteria and mycobacteria. J Appl Microbiol 92 Suppl, 46S-54S (2002). 492 38. A. H. Delcour, Outer membrane permeability and antibiotic resistance. Biochim Biophys 493 Acta 1794, 808-816 (2009). 494 39. P. S. Manzanillo et al. , The ubiquitin ligase parkin mediates resistance to intracellular 495 pathogens. Nature 501, 512-516 (2013). 496 40. L. S. Ates et al. , Essential Role of the ESX-5 Secretion System in Outer Membrane 497 Permeability of Pathogenic Mycobacteria. PLoS Genet 11, e1005190 (2015). 498 41. Q. Wang et al., PE/PPE proteins mediate nutrient transport across the outer membrane 499 of Mycobacterium tuberculosis. Science 367, 1147-1151 (2020). 500 42. N. van der Wel et al., M. tuberculosis and M. leprae translocate from the phagolysosome 501 to the cytosol in myeloid cells. Cell 129, 1287-1298 (2007). 502 43. R. Simeone et al., Phagosomal rupture by Mycobacterium tuberculosis results in toxicity 503 and host cell death. PLoS Pathog 8, e1002507 (2012). 504 44. D. Houben et al. , ESX-1-mediated translocation to the cytosol controls virulence of 505 mycobacteria. Cell Microbiol 14, 1287-1298 (2012). 506 45. P. S. Manzanillo, M. U. Shiloh, D. A. Portnoy, J. S. Cox, Mycobacterium Tuberculosis 507 Activates the DNA-Dependent Cytosolic Surveillance Pathway within Macrophages. Cell 508 Host Microbe 11, 469-480 (2012). 509 46. R. O. Watson, P. S. Manzanillo, J. S. Cox, Extracellular M. tuberculosis DNA Targets 510 Bacteria for Autophagy by Activating the Host DNA-Sensing Pathway. Cell 150, 803-815 511 (2012). 512 47. A. Coros, B. Callahan, E. Battaglioli, K. M. Derbyshire, The specialized secretory 513 apparatus ESX-1 is essential for DNA transfer in Mycobacterium smegmatis. Mol 514 Microbiol 69, 794-808 (2008). 515 48. M. S. Siegrist et al. , Mycobacterial Esx-3 is required for mycobactin-mediated iron 516 acquisition. Proc Natl Acad Sci U S A 106, 18792-18797 (2009). 517 49. S. R. Elliott, A. D. Tischler, Phosphate responsive regulation provides insights for ESX- 5 518 function in Mycobacterium tuberculosis. Curr Genet 62, 759-763 (2016). 519 50. T. A. Gray et al. , Intercellular communication and conjugation are mediated by ESX 520 secretion systems in mycobacteria. Science 354, 347-350 (2016). 521 51. C. Portal-Celhay et al. , Mycobacterium tuberculosis EsxH inhibits ESC RT-dependent 522 CD4(+) T-cell activation. Nat Microbiol 2, 16232 (2016). 523 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint 52. E. Mittal et al. , Mycobacterium tuberculosis Type VII Secretion System Effectors 524 Differentially Impact the ESCRT Endomembrane Damage Response. MBio 9, e01765-525 01718 (2018). 526 53. S. A. Stanley, J. E. Johndrow, P. Manzanillo, J. S. Cox, The Type I IFN response to 527 infection with Mycobacterium tuberculosis requires ESX-1-mediated secretion and 528 contributes to pathogenesis. J Immunol 178, 3143-3152 (2007). 529 54. D. G. Russell, The ins and outs of the Mycobacterium tuberculosis-containing vacuole. 530 Cell Microbiol 18, 1065-1069 (2016). 531 55. K. R. Nicholson, C. B. Mousseau, M. M. Champion, P. A. Champion, The genetic 532 proteome: Using genetics to inform the proteome of mycobacterial pathogens. PLoS 533 Pathog 17, e1009124 (2021). 534 56. O. A. Collars et al. , An N-acetyltransferase required for ESAT-6 N-terminal acetylation 535 and virulence in Mycobacterium marinum. mBio 14, e0098723 (2023). 536 57. E. A. Williams et al. , A Nonsense Mutation in Mycobacterium marinum That Is 537 Suppressible by a Novel Mechanism. Infect Immun 85 (2017). 538 58. J. Yu et al. , Both phthiocerol dimycocerosates and phenolic glycolipids are required for 539 virulence of Mycobacterium marinum. Infect Immun 80, 1381-1389 (2012). 540 59. O. Vergnolle et al. , Biosynthesis of cell envelope-associated phenolic glycolipids in 541 Mycobacterium marinum. J Bacteriol 197, 1040-1050 (2015). 542 60. M. M. Champion, E. A. Williams, R. S. Pinapati, P. A. Champion, Correlation of 543 Phenotypic Profiles Using Targeted Proteomics Identifies Mycobacterial Esx-1 544 Substrates. J Proteome Res 3, 5151-5164 (2014). 545 61. R. E. Bosserman, K. R. Nicholson, M. M. Champion, P. A. Champion, A New ESX- 1 546 Substrate in Mycobacterium marinum That Is Required for Hemolysis but Not Host Cell 547 Lysis. J Bacteriol 201 (2019). 548 62. B. McLaughlin et al. , A mycobacterium ESX-1-secreted virulence factor with unique 549 requirements for export. PLoS Pathog 3, e105 (2007). 550 63. A. E. Chirakos, K. R. Nicholson, A. Huffman, P. A. Champion, Conserved ESX- 1 551 substrates EspE and EspF are virulence factors that regulate gene expression. Infect 552 Immun 10.1128/IAI.00289-20 (2020). 553 64. R. E. Bosserman et al. , WhiB6 regulation of ESX-1 gene expression is controlled by a 554 negative feedback loop in Mycobacterium marinum. Proc Natl Acad Sci U S A 555 10.1073/pnas.1710167114 (2017). 556 65. K. M. Guinn et al., Individual RD1-region genes are required for export of ESAT-6/CFP-557 10 and for virulence of Mycobacterium tuberculosis. Mol Microbiol 51, 359-370 (2004). 558 66. L. Y. Gao et al. , A mycobacterial virulence gene cluster extending RD1 is required for 559 cytolysis, bacterial spreading and ESAT-6 secretion. Mol Microbiol 53, 1677-1693 560 (2004). 561 67. K. G. Sanchez et al. , EspM Is a Conserved Transcription Factor That Regulates Gene 562 Expression in Response to the ESX-1 System. mBio 11 (2020). 563 68. T. Parish, N. G. Stoker, Use of a flexible cassette method to generate a double 564 unmarked Mycobacterium tuberculosis tlyA plcABC mutant by gene replacement. 565 Microbiology 146 ( Pt 8), 1969-1975 (2000). 566 69. R. E. Bosserman, C. R. Thompson, K. R. Nicholson, P. A. Champion, Esx Paralogs Are 567 Functionally Equivalent to ESX-1 Proteins but Are Dispensable for Virulence in 568 Mycobacterium marinum. J Bacteriol 200, e00726-00717 (2018). 569 70. C. Li et al. , FastCloning: a highly simplified, purification-free, sequence- and ligation-570 independent PCR cloning method. BMC Biotechnol 11, 92 (2011). 571 71. T. Parish, Electroporation of Mycobacteria. Methods Mol Biol 2314, 273-284 (2021). 572 72. J. W. Saelens et al. , An ancestral mycobacterial effector promotes dissemination of 573 infection. Cell https://doi.org/10.1016/j.cell.2022.10.019 (2022). 574 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint 73. O. A. Collars et al. , An N-acetyltransferase required for EsxA N-terminal protein 575 acetylation and virulence in Mycobacterium marinum. bioRxiv 576 10.1101/2023.03.14.532585, 2023.2003.2014.532585 (2023). 577 578 579 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint Figures and Legends: 580 581 582 Figure 1. Characterization of PDIM/PGL deficient M. marinum strains. A. Schematic of the 583 genetic loci required for PDIM production and transport. B. qRT-PCR of mas transcript relative 584 to the sigA transcript compared to the WT M. marinum strain. Each datapoint is an independent 585 biological replicate, and an average of three technical replicates. Statistical analysis was per-586 formed using an ordinary one-way ANOVA ( P=.0003), followed by a Dunnett’s multiple compa r-587 ison test, *** P=.0002, ** P=.0093. C. TLC of total lipids isolated from the WT, Δmas and com-588 plemented M. marinum strains. 9 μl of total lipids was analyzed. This TLC is representative of at 589 least three independent biological replicates. 590 591 592 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint 593 Figure 2. Loss of PDIM/ PGL results in widespread changes to protein secretion from M. 594 marinum. Volcano plots of the secreted protein levels from c ell associated (left) and secreted 595 protein fractions from M. marinum. Values are the average of four biological replicates. Signif i-596 cance was plotted against the log2 fold change compared to the WT strain. blue dots indicate P 597 ≤ 0.05; red dots indicate P > 0.05. Vertical, black dashed lines signify a log2 foldchange = +/- 1. 598 599 600 601 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint 602 Figure 3. The loss of PDIM/PGL impacts ESX-1 secretion differentially. A. Schematic of the 603 ESX-1 secretion hierarchy. ESX-1 represents the ESX-1 membrane components. Group 1: 604 EsxB, EsxA, PPE68 and MMAR_2894, Group 2: EspB, EspK, EspJ, Group 3: EspE and EspF, 605 from (10) . B. Log2 Fold-Change of ESX-1 substrate levels in the secreted protein fractions 606 compared to the WT strain. Data in Dataset 1C. C. Heat map of the ESX-1 substrate levels in 607 the secreted protein fractions normalized to the WT strains (Data in Dataset 1D, from Dataset 608 S1B) 609 610 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint 611 Figure 4. The loss of PDIM/PGL impacts hemolysis and the secretion of the EspE sub-612 strate. A. Hemolytic activity of M. marinum. Each data point is a single technical replicate, ma k-613 ing up four biological replicates. Significance was determined using a one-way ANOVA, 614 P<.0001, followed by a Tukey’s multiple comparison test **** P<.0001. B. Western blot analysis 615 of M. marinum cell associated and secreted protein fractions. 10 μg of protein were loaded in 616 each lane. RpoB is a control for lysis, MPT32, a protein secreted by the Sec system, is a loa d-617 ing control. The two blots shown were representative of at least four biological replicates. 618 (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. The copyright holder for this preprintthis version posted January 10, 2024. ; https://doi.org/10.1101/2024.01.09.574891doi: bioRxiv preprint