bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1 SapM mutation to improve the BCG vaccine: genomic, transcriptomic and preclinical safety 2 characterization 3 Nele Festjensa,b#, Kristof Vandewallea,b,1 , Erica Houthuysa,b, 2, Evelyn Pletsa,b, Dieter Vanderschaeghea,b,3, 4 Katlyn Borgersa,b, Annelies Van Heckea,b, Petra Tielsa,b, Nico Callewaerta,b# 5 a 6 b 7 Belgium VIB-UGent Center for Medical Biotechnology; Technologiepark 927, 9052 Zwijnaarde – Belgium. Department of Biochemistry and Microbiology, Ghent University; Technologiepark 927, 9052 Zwijnaarde - 8 9 # Joint corresponding authors 10 Technologiepark 927 11 9052 Zwijnaarde 12 Nele.Festjens@vib-ugent.be 13 Nico.Callewaert@vib-ugent.be 14 Phone: +32 (0) 9 33 13 630 Current address: 1Inbiose, Technologiepark 3, bus 41, 9052 Zwijnaarde; 2iTeos Therapeutics, Rue des Frères Wright 29, 6041 Gosselies; 3NLO, AA Tower, Technologiepark 19, 9052 Zwijnaarde bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 15 Abstract 16 The Mycobacterium bovis Bacille Calmette Guérin (BCG) vaccine shows variable efficacy in protection against 17 adult tuberculosis (TB). Earlier, we have described a BCG mutant vaccine with a transposon insertion in the 18 gene coding for the secreted acid phosphatase SapM, which led to enhanced long-term survival of vaccinated 19 mice challenged with TB infection. To facilitate development of this mutation as part of a future improved 20 live attenuated TB vaccine, we have now characterized the genome and transcriptome of this sapM::Tn 21 mutant versus parental BCG Pasteur. Furthermore, we show that the sapM::Tn mutant had an equal low 22 pathogenicity as WT BCG upon intravenous administration to immunocompromised SCID mice, passing this 23 important safety test. Subsequently, we investigated the clearance of this improved vaccine strain following 24 vaccination and found a more effective innate immune control over the sapM::Tn vaccine bacteria as 25 compared to WT BCG. This leads to a fast contraction of IFNγ producing Th1 and Tc1 cells after sapM::Tn BCG 26 vaccination. These findings corroborate that a live attenuated vaccine that affords improved long-term 27 survival upon TB infection can be obtained by a mutation that further attenuates BCG. These findings suggest 28 that an analysis of the effectiveness of innate immune control of the vaccine bacteria could be instructive 29 also for other live attenuated TB vaccines that are currently under development, and encourage further 30 studies of SapM mutation as a strategy in developing a more protective live attenuated TB vaccine. 2 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 31 Keywords 32 Tuberculosis, vaccine, SapM, immunomodulation 3 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 33 1. Introduction 34 Today, TB remains one of the major causes of infectious disease and death throughout the world [1]. About 35 10 million people developed TB in 2017, with an estimated 1.3 million deaths. 36 Although new and promising anti-tubercular drugs are in or approaching the clinic, full control over this 37 devastating disease will necessarily require a good, preventive anti-TB vaccine [2]. M. bovis Bacille-Calmette- 38 Guérin (BCG) is the only licensed vaccine today, and already in use for almost 100 years. Although BCG is 39 protective against TB meningitis and disseminated TB in children, its efficacy against pulmonary TB is highly 40 variable and not sufficient for disease control [3–5]. 41 Currently, 13 vaccine candidates are under active development and in the process of clinical testing. Besides 42 two live recombinant candidates (VPM1002 and MTBVAC) which are being tested to replace BCG as a priming 43 vaccine, the candidates being mainly variations of booster vaccines with recurrent key mycobacterial 44 antigens [2,6,7]. Several of these vaccine candidates yield enhanced interferon γ (IFNɣ)-producing cellular 45 immunity and enhanced control over bacterial replication in the first weeks after TB challenge in animal 46 models. However, reports that show extended long-term survival of vaccinated animals upon TB challenge 47 are much rarer. In the category of priming vaccines, it has been shown that rBCG30 (BCG Tice overexpressing 48 the M. tb 30-kDa major secretory protein antigen 85B)-immunized mice survive longer than BCG-immunized 49 mice following TB challenge [8]. SO2 (phoP inactivated in M. tb strain MT103), the MTBVAC (containing a 50 double deletion in phoP and fadD26) prototype, conferred greater efficacy than BCG in a high-dose challenge, 51 long-term survival experiment in guinea pigs, while no difference to WT BCG was observed after low-dose 52 challenge [9]. The recombinant BCG strain AFRO-1 (BCG1331 ΔureC::ΩpfoAG137Q, Ag85A, Ag85B and TB10.4) 53 affords better protection than BCG in mice following aerosol challenge with Mycobacterium tuberculosis 54 (M.tb) [10]. Additionally, some prime-boost vaccination strategies demonstrated improved long-term 55 survival of M.tb-challenged guinea pigs versus BCG alone: BCG priming followed by the M72 recombinant 56 protein booster vaccine (fusion of Mtb32 and Mtb39 antigens of Mtb) [11] and ID93 (combines four antigens 4 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 57 belonging to families of M.tb proteins associated with virulence (Rv2608, Rv3619, Rv3620) or latency 58 (Rv1813)/GLA-SE)) [12]. 59 Earlier, we have demonstrated how a sapM transposon mutant of M. bovis BCG, hereafter called sapM::Tn 60 BCG, resulted in a TB vaccine candidate that provides improved long-term survival in mice upon both systemic 61 and intratracheal challenge with M. tb as compared to the parental BCG Pasteur-derived strain (hereafter 62 called WT BCG) [13]. We suggested that the secreted acid phosphatase SapM may have evolved as an 63 important immunomodulatory protein of Mycobacteria. The importance of SapM for mycobacterial 64 immunomodulation was since confirmed by independent laboratories [14–17]. In the present study, we 65 report on the in-depth genomic and transcriptomic analysis of the sapM::Tn BCG mutant, to assess which 66 alterations are induced by the transposon insertion in the sapM locus. 67 One of the major impediments to TB vaccine development is the incomplete understanding of the 68 mechanisms of protective immunity against M.tb, so far obstructing rational vaccine development [18]. 69 Research on animal models and ongoing clinical trials may reveal markers that correlate with a vaccine’s 70 protective potential. However, as long as we lack these correlates, vaccine design will be challenging. One 71 general idea that dominated immunology of infectious diseases over the last decades, is the T helper type 72 1/type2 (Th1/Th2) paradigm. Following this model, Th1 cells would protect the host from intracellular 73 pathogens, like M.tb. Indeed, the early appearance of Th1 type CD4+ T cells secreting IFNγ is necessary for 74 the orchestration of protective immunity in the infected lung [19]. However, controversy has arisen about 75 whether the quantitative extent of IFNɣ production in recall experiments has value as a correlate of vaccine- 76 induced protection against TB. Indeed, several reports show that other anti-tuberculosis CD4 T-cell effector 77 functions are at play [20–24]. 78 In this evolving context, we also report on the safety in immunocompromised mice as well as rapid innate 79 control over the sapM::Tn vaccine bacteria, with intriguing impact on the induced adaptive immunity. Our 80 findings suggest that a live attenuated TB vaccine that behaves more like an acute, rapidly immune-controlled 5 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 81 infection, rather than the protracted chronic infection caused by the current BCG vaccine, may yield an 82 immune status that affords more prolonged control of a subsequent TB infection. 6 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 83 2. Results 84 2.1 In depth characterization of the sapM::Tn BCG versus WT BCG strain 85 2.1.1 86 The M. bovis BCG Pasteur strain 1721, used to create the sapM::Tn BCG mutant strain [13,25], differs from 87 the 1173P2 Pasteur vaccine strain by a K43R point mutation in the rpsL gene, conferring streptomycine 88 resistance [26], which is useful to avoid contaminations during the lengthy and involved procedures in 89 ordered transposon mutant library production. To investigate whether there are any additional variants 90 between the WT BCG Pasteur and the sapM::Tn BCG mutant, we performed a whole-genome resequencing 91 analysis on both strains (paired end 2 x 150 bp), mapping to the M. bovis BCG Pasteur 1173P reference 92 genome, resulting in ± 80x average coverage of mapped reads (Suppl. Fig. 1). We could reconfirm the location 93 of the inserted Himar1 transposon, at the TA dinucleotide 8 bp before the sapM start codon (Fig. 1). This was 94 also demonstrated by de novo assembly of the sequencing reads from the sapM::Tn BCG mutant (data not 95 shown). In addition, we performed a probabilistic variant analysis on both WT BCG and sapM::Tn BCG 96 mappings and detected only minor modifications (single amino acid change) with unknown impact. These 97 variants are listed in Table 1 and were verified by Sanger sequencing. All variants versus the Pasteur reference 98 are present in both the parental as well as the sapM::Tn strains, with only one exception, i.e. the frameshift 99 mutation in the sugI gene, coding for sugar-transport integral membrane protein. This mutation probably 100 arose after the library preparation (in a later passage of the strain), since this frameshift mutation is not 101 present in sapM::Tn. The exact function of sugI is unknown, however, it shows distant sequence similarity to 102 glucose permease GlcP of S. coelicolor and the galactose (GalP) and arabinose (AraE) transporters of E. coli. 103 Thus, the system is likely to transport a monosaccharide [27]. It is described to be non-essential in the H37Rv 104 strain [28], which explains the absence of phenotype in WT BCG due to the frameshift mutation. The SNV in 105 sapM::Tn BCG SugI also likely arose after library preparation since it is not present in WT BCG. Hence, sugI 106 appears to be a highly variable gene in the BCG lineage during in vitro cultivation, as the three strains each 107 have a different sequence variant. Interestingly, both the parental as well as the sapM::Tn strain contain a Whole-genome resequencing 7 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 108 frame-shift mutation in the gene coding for FadD26, a key enzyme in the biosynthesis pathway of the 109 phthiocerol dimycocerosate (PDIM) class of virulence lipids. This fadD26 mutation is interesting from a 110 vaccine engineering point-of-view, as it further attenuates virulence [29,30]. 111 2.1.2 112 Transcriptome analysis, by both RNAseq analysis and RT-PCR, demonstrated that the sapM transcript is ~20- 113 50-fold reduced compared to the WT BCG, while the transcription levels of other genes remain largely 114 unchanged (< 2-fold change), except for a ~10-fold upregulation of upp (encoding uracil 115 phosphoribosyltransferase) which is immediately upstream of sapM in the genome, and a modest 2.4-fold 116 downregulation of its downstream gene (BCG_3376) (Fig. 2A-C, Suppl. Fig. 2A-B, Suppl. Table 1). A few (~6) 117 other genes show a differential expression pattern (~2-fold change) between WT BCG and sapM::Tn, 118 however, this effect is marginal compared to the differences in expression of sapM and upp. The transposon 119 insertion thus only majorly affects one nearby gene, and the changes in sapM and upp transcription have 120 almost no secondary effects on overall transcriptional regulation. According to the Operon Correlation 121 Browser on the TB Database [31], the gene coding for SapM in M. tb (Rv3310) and the downstream gene 122 Rv3311 potentially form an operon. However, in M. bovis BCG we see that transposon disruption of the sapM 123 upstream region in sapM::Tn BCG very strongly reduces the levels of sapM transcript, while having a much 124 more modest effect on the levels of BCG_3376 (Rv3311 ortholog) transcription. These data indicate that both 125 genes are likely differentially regulated, questioning the idea of an operonic structure. 126 2.1.3 127 As the SapM protein contains a 43 AA N-terminal signal sequence and thus gets secreted in the culture 128 medium, we collected culture supernatant samples at various time-points. 129 First, we performed an ELISA on these supernatant samples with a newly generated antibody directed against 130 the recombinant SapM protein produced in E. coli (Fig. 2D). While the amount of SapM protein steadily 131 increases over time of culture in the WT BCG, the signal is nearly absent in the sapM::Tn BCG mutant samples Transposon-induced transcriptional effects SapM protein is undetectable and secreted total phosphatase activity is strongly reduced 8 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 132 for all time-points. We further confirmed the absence of full-length SapM protein in the sapM::Tn BCG strain 133 by SDS-PAGE and western blotting in an independent experiment (Fig. 2E). We detected a band of ± 28 kDa 134 in the WT BCG, but not in the sapM::Tn BCG mutant sample. A smear is also detected above the 28 kDa band, 135 which likely represents aspecific binding (as the pattern appears to be identical for all samples). To check for 136 any remaining phosphatase activity in the culture medium, we performed an in vitro phosphatase assay with 137 p-Nitrophenyl phosphate (pNPP) as a substrate [32] (Fig. 2F). The measured in vitro activity is significantly 138 reduced for the sapM::Tn BCG mutant sample compared to WT BCG at early time-points. At later time-points 139 (day 8), we still observe a considerable signal, which is expected, as the assay would also pick up activity of 140 other phosphatases such as the protein tyrosine phosphatases (PtpA and PtpB) [33]. 141 Consequently, based on all of the data on this vaccine strain, we concluded at this point that its improved 142 vaccine efficiency must indeed be due to a loss of function of SapM or the upregulation of upp transcription, 143 or both. 144 2.2 In vivo safety of sapM::Tn BCG versus WT BCG 145 A requirement of live attenuated vaccines is that they should be safe, even in immunocompromised hosts. 146 Our previous analysis of bacterial replication in immunocompetent Balb/c mice, infected intravenously, did 147 not show any difference between WT BCG and sapM::Tn BCG [13]. The number of granulomas formed in 148 livers and lungs post-infection was unaffected [13]. To unambiguously demonstrate safety of the sapM::Tn 149 BCG mutant compared to WT BCG in vivo, we investigated bacterial virulence in the absence of adaptive 150 immunity. Immunocompromised SCID mice, lacking both T and B cells, were infected intravenously (i.v.) with 151 a high (3x107 CFU) or lower dose (3x106 CFU) of both strains and survival was monitored (Fig. 3). After both 152 low and high doses, SCID mice infected with sapM::Tn BCG mutant showed similar survival to mice infected 153 with WT BCG. Thus, in the SCID model, virulence of sapM::Tn BCG is comparable to WT BCG. 154 2.3 At the lymph node draining the vaccination site, the sapM::Tn BCG mutant is controlled faster than 155 the WT BCG strain 9 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 156 Our previously reported observations indicated that vaccination with the sapM::Tn BCG mutant triggered 157 more effective recruitment of iDCs into the draining lymph nodes (LNs) in Balb/c mice [13]. This lead us to 158 study the influence of this more effective induction of the innate immune response at early time points after 159 sapM::Tn BCG vaccination on the fate of the vaccine inoculum. To enhance the validity, experiments were 160 performed in F1 mice (C57BL/6J x Balb/c) to assess robustness of the observed effects to a more 161 heterogeneous immunogenetic background [34]. After subcutaneous (s.c) vaccination of F1 mice, less 162 bacteria (~50%) were recovered from LNs of sapM::Tn BCG infected mice as compared to WT BCG-infected 163 mice at all time-points analyzed (i.e. days 8, 14 and 28 post-infection) (Fig. 4A). It is clear that whereas WT 164 BCG could amplify from the inoculum, this was barely the case for the sapM::Tn BCG mutant. Since 165 differences are clear already 8 days post-vaccination, this result is most consistent with improved control by 166 innate immune mechanisms. Indeed, an analysis of intracellular survival of WT BCG and sapM::Tn BCG upon 167 infection of bone marrow derived macrophages (BM-DMs), showed that WT BCG and sapM::Tn BCG are 168 taken up similarly (as shown before [13]) but that growth of the sapM::Tn BCG mutant is better controlled 169 compared to the WT BCG strain upon infection (Fig. 4B). In line with these findings, such improved clearance 170 of SapM-mutated M.tb by BM-DMs has recently also been described [16], an effect that was attributed to 171 improved phagosomal maturation. 172 To further evaluate the impact of early innate growth control of the bacteria, we have analyzed survival of 173 sapM::Tn BCG and WT BCG after intravenous infection of F1 mice. Twenty-four hours post-infection, the 174 bacterial load in the lungs and spleens from mice that were infected with WT BCG was higher than the loads 175 of those infected with the sapM::Tn BCG mutant (~50%) (Fig. 4C). Three weeks post-infection, the difference 176 was still observed in the lungs, but not in the spleens. Similar results were obtained in the spleen of C57BL/6J 177 mice, but not in the lungs (Suppl. Fig. 3A). Cytokine levels were analyzed in the serum of i.v. exposed mice at 178 different time-points (6h, 24h, 48h post-vaccination), and significantly higher levels of IL1-β and IL17 could 179 be measured 24h post-sapM::Tn BCG vaccination compared to WT BCG vaccination (Suppl. Fig. 3B). In view 180 of this early time point post-vaccination, both IL1-β and IL17 are produced by innate immune cells. These 10 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 181 particular cytokines are known to be important innate cytokines for bacterial killing, particularly in 182 mycobacterial infection [35,36], which is consistent with the reduced sapM::Tn BCG bacterial load. 183 2.4 Complementation of the sapM::Tn mutation reverses the phenotypes. 184 As the Tn insertion in the sapM locus caused both a loss of function of sapM and a strong upregulation of the 185 upstream upp gene, we assessed whether the improved early innate growth control of the sapM::Tn BCG 186 bacteria was due to the loss of function of SapM or the upregulation of upp transcription. For this purpose, 187 we have complemented the sapM::Tn BCG mutant strain with the sapM gene under control of its own 188 promotor. The sapM expression is reverted to wild type levels in this sapM::Tn:compl BCG mutant, but upp 189 levels are still increased similarly as in the sapM::Tn BCG mutant (Fig. 5). Analysis of bacterial load in the 190 lungs and spleens from mice that were infected with either sapM::Tn:compl BCG mutant, sapM::Tn BCG or 191 WT BCG showed that complementation of sapM expression abolished the improved growth control of the 192 sapM::Tn BCG mutant compared to the WT BCG strain upon infection (Fig. 6). These results confirm that the 193 observed phenotypes are due to the reduced expression of sapM and not because of upregulation of upp 194 (Fig. 2C). 195 2.5 Faster kinetics of iDC recruitment to lymphoid organs and decrease in IFNγ-producing CD4 + and CD8+ T 196 recall response in sapM::Tn BCG vaccinated mice compared to WT BCG vaccinated mice 197 We then further analyzed how improved innate control over vaccine bacteria changes the immune response 198 to the vaccine. Hereto, we have vaccinated C57BL/6J mice s.c. with WT BCG or sapM::Tn BCG mutant. At 199 different time points, mice were sacrificed and LNs and spleens were isolated. DC phenotyping was 200 performed. Significant differences were observed in total leukocyte numbers and iDC recruitment kinetics 201 after sapM::Tn BCG versus WT BCG vaccination. In general, we observe a faster induction of the innate 202 immune response, reflected by a faster recruitment of innate inflammatory cells to the secondary lymphoid 203 organs (dLN-spleen), and by a faster contraction of the innate immune response, seen by lower leukocyte 204 cell numbers at day 14 post-vaccination with BCG sapM::Tn BCG (Fig. 7). Next, T cell immunity was assessed 205 by restimulation of splenocytes or LN cells with either the mycobacterial antigen Ag85A, the M.tb extract 11 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 206 PPD or by anti-CD3/28, a stimulus that activates all T cells, as a control. Vaccination with sapM::Tn BCG leads 207 to a lower number of lymph node and spleen IFNγ-producing CD4+ (Th1) and CD8+ (Tc1) T cells 2 weeks post- 208 vaccination compared to WT BCG (Fig. 8A). After vaccination with a lower dose, this observation is only seen 209 for Tc1 cells, following polyclonal stimulation in the LNs and spleen (Suppl. Fig. 4). IFNγ measurements in the 210 supernatant following stimulation with the different antigens also demonstrated lower levels of IFNγ in the 211 sapM::Tn BCG condition (Fig. 8B). These lower numbers of IFNγ-producing CD4+ and CD8+ T cells were seen 212 at early time points (2 weeks post-infection), but the difference became non-significant at later time points 213 (Suppl. Fig. 5), consistent with the hypothesis that a fast innate control of bacterial load results in a swift 214 contraction of the TB-specific Th1 and Tc1 cellular response after sapM::Tn BCG vaccination. The sapM::Tn 215 BCG vaccine thus behaves like a rapidly (though not completely) controlled bacterial infection, contrary to 216 WT BCG, which expands and causes a more protracted infection at the vaccination site. 217 3. Discussion 218 Our former research demonstrated that an M. bovis BCG sapM::Tn mutant led to enhanced long-term 219 survival of vaccinated mice challenged with TB [13]. Further development of this TB vaccine candidate 220 requires further characterization of the mutant and its safety, which was the main purpose of the extensive 221 work presented in this study. 222 We have performed a whole genome resequencing-based variant analysis of this sapM::Tn BCG mutant, as 223 well as of its parent BCG strain [26], using a shotgun Illumina sequencing approach. Very few polymorphisms 224 could be detected compared to the M. bovis BCG Pasteur reference genome, most of which were already 225 present in the streptomycin-resistant Pasteur derivative in which we designed our transposon mutant. A 226 frame-shift mutation had inactivated the gene coding for FadD26, a key enzyme in the biosynthesis pathway 227 of the PDIM class of virulence lipids that are involved in hiding M. tb’s own PAMPs from the host’s innate 228 immune system [29]. Knocking out this gene in M. tb severely impairs the pathogen’s ability to survive in vivo 229 [29,30]. For this reason, the fadD26 gene is deleted in MTBVAC, the first live-attenuated M. tb vaccine [37]. 230 Chen et al. compared PDIM/PGL production in 12 different M. bovis BCG substrains and found that while 12 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 231 most were PDIM-positive (including BCG Pasteur), three of these strains, BCG Japan (or Tokyo), Moreau and 232 Glaxo, do not produce this class of lipids [38]. The authors further reported that there was a correlation 233 between the virulence that was associated with the BCG substrains and their lipid profile: the on average 234 more virulent, PDIMs/PGLs-producers and the on average less virulent, PDIMs/PGLs non-producers. In a later 235 study it was shown that the PDIM-defect in BCG Moreau was due to a deletion in the FadD26 and the directly 236 downstream ppsA gene in the same operon, which is also involved in the PDIM/PGL lipid biosynthesis 237 pathway [39]. The authors reported that this locus was intact in BCG Japan and Glaxo. Naka et al later showed 238 that BCG Tokyo 172 is actually divided into two subpopulations, in which type II, but not type I, has a 239 frameshift mutation in the ppsA gene and thus does not produce PDIM lipids [40]. It is unclear whether this 240 strain is the same as the Japan strain that was earlier reported to lack PDIMs [38]. As the Japan/Tokyo, 241 Moreau and Glaxo substrains are not derived from one another [41], and as we have here detected a fadD26 242 frameshift in a substrain derived from BCG Pasteur, these independent mutations in genes coding for key 243 enzymes in the PDIM/PGL lipid biosynthesis pathway indicate that there may be a selective force in favor of 244 losing this class of lipids during in vitro BCG cultivation. We also conclude that, as our strain has this 245 background fadD26 mutation, the sapM::Tn vaccine strain in our studies is in fact a fadD26/sapM::Tn double 246 mutant, which is important knowledge for further development, as it shows that the sapM mutation 247 improves vaccine efficacy even in this fadD26 background, as is currently being used for safety enhancement 248 of the MTBVAC lead clinical TB vaccine candidate. In addition to our sapM::Tn BCG mutant, this FadD26 249 mutation will almost certainly also be present in the Zmp1 BCG mutant candidate vaccine that was described 250 in [42], since it is also derived from the M. bovis BCG Pasteur 1721 strain. 251 By comparative transcriptome profiling between WT BCG and sapM::Tn BCG, we observed an almost 252 complete absence of sapM transcript in the mutant strain, together with a significant upregulation of sapM’s 253 upstream gene upp. By RNAseq analysis, we further showed that the global transcriptional landscape in the 254 sapM::Tn BCG mutant strain is not significantly altered compared to the WT BCG. 255 The most promising strategy for improved TB vaccines currently is the development of an improved 13 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 256 attenuated live vaccine, either used alone or in combination with subunit booster vaccines [6]. Furthermore, 257 re-vaccination with BCG in adolescence has recently shown promise as well [43]. An improved version of the 258 BCG vaccine could thus be valuable in this context too. However, working with live vaccines encompasses 259 potential safety issues, especially since a high incidence of HIV is found in areas where TB is endemic. A novel 260 vaccine candidate should therefore be at least as safe as BCG. Safety of the sapM::Tn BCG strain was tested 261 in immunocompromised SCID mice, in comparison with the WT BCG. Survival time upon infection of the mice 262 was similar for both strains, at both high and low infectious doses, demonstrating comparable safety of the 263 sapM::Tn BCG strain and WT BCG. 264 Due to the lack of natural infection-induced protection, the type of immune response that is crucial in 265 preventing M.tb infection and TB, and which therefore should be induced by vaccination, remains largely 266 unknown. At present, it is believed that a particular balance between differentially polarized adaptive 267 immune responses is crucial to overcome a TB infection. M. tb as well as M. bovis, from which BCG derives, 268 have evolved immunomodulatory mechanisms to perturb this balance to its own benefit [44,45]. We and 269 others propose that an improved, live attenuated vaccine will need to be engineered to take down these 270 immunomodulatory virulence factors from the pathogenic parental Mycobacteria from which these vaccines 271 (incl. BCG) have been derived, and we propose that the SapM secreted phosphatase is such a factor. Our 272 results demonstrate a better innate control of sapM::Tn BCG vaccine bacteria as compared to WT BCG upon 273 vaccination, correlating with a faster influx of iDCs in lymphoid organs draining the vaccination site, and more 274 modest primary expansion of TB-antigen-specific IFNγ-producing CD4 + and CD8+ T cells. Both WT BCG and 275 sapM::Tn BCG have equivalent growth characteristics during in vitro cultivation (Suppl. Fig. 6), excluding that 276 mere growth rate differences could have caused any difference in abovementioned phenotypes. Since 277 complementation of sapM expression abolished the improved growth control of the sapM::Tn BCG mutant 278 upon infection, the observed phenotypes are specifically due to the reduced expression of sapM. To 279 conclude, all of this shows that the sapM mutated BCG vaccine behaves like a more immediately effectively 280 controlled bacterial infection as compared to WT BCG. Importantly, in the context of viral infections, it has 281 been well established that a protracted fight of the immune system to a chronic infection yields poorly 14 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 282 efficacious memory due to poor generation of central memory T cells [46]. Possibly, the same holds true for 283 chronic bacterial intracellular infections, of which BCG vaccination is an example. Together with the 284 comparable safety of the SapM-mutated BCG and WT BCG in immunodeficient mice, and the fact that the 285 sapM::Tn BCG vaccine is presently one of the few priming vaccines for which enhanced long-term survival of 286 TB-infected animals has been demonstrated [13], these results warrant inclusion of the SapM inactivation 287 strategy in future generations of live attenuated TB vaccines. 288 4. Materials and Methods 289 4.1 Mycobacterial strains and media 290 The streptomycin resistant M. bovis BCG Pasteur strain 1721 [26] (RpsL, K43R; a gift of Dr. P. Sander, Institute 291 for Medical Microbiology, Zürich) and its sapM transposon insertion mutant [13] were grown in shaking 292 culture flasks in Middlebrook 7H9 broth (Difco) supplemented with 0.05% Tween80 and Middlebrook OADC 293 (Becton Dickinson) when grown in liquid culture. Difco Middlebrook 7H10 agar was used for growth on solid 294 culture. For the SDS-PAGE and western blotting experiment, M. bovis BCG cultures were grown in 7H9 295 medium without OADC, but supplemented with 0.2% glucose, 0.2% glycerol and 0.05% of Tween80, because 296 the large quantity of albumin in the OADC supplement interferes with protein electrophoresis. All strains 297 were frozen as 1 mL cultures at OD600 1 in 20% glycerol (final concentration). For all in vitro and in vivo 298 infections described in this paper, cells were started from a glycerol stock, grown in shaking culture flasks 299 (OD600 always ≤1) as described above, subcultured once and further grown until OD600~0.8 after which cells 300 were collected and prepared for infection. For each experiment, we confirmed that we immunized with the 301 same number of viable CFU by CFU plating of the inoculum. 302 4.2 Whole genome Illumina resequencing and data analysis 303 Genomic DNA of the M. bovis BCG strain 1721 and its sapM::Tn BCG mutant (Tn-7432) was prepared [47] 304 and used for an Illumina library prep (Nextera XT DNA Library Preparation kit). The libraries were sequenced 305 on an Illumina MiSeq instrument (2 x 150 bp reads), with an average 80x coverage per genome. The library 15 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 306 prep and sequencing was performed by the VIB Nucleomics Core facility (www.nucleomics.be). The data was 307 analyzed by the VIB Nucleomics Core facility and the VIB Bioinformatics Core facility (www.bits.vib.be) using 308 CLC Genomics Workbench (CLC-GWB [48], http://www.clcbio.com/products/clc-genomics-workbench/). All 309 reads were processed using the standard CLC-GWB settings. The reads from the sapM::Tn BCG strain were 310 used for de-novo assembly in order to locate the inserted transposon. To this end, the phiMycoMarT7 311 transposon sequence (GenBank AF411123.1) was blasted against all de-novo contigs (n = 201). Using this 312 information, a modified reference sequence was built from the M. bovis BCG Pasteur str. 1173P2 reference 313 (NCBI NC_008769.1) by introducing the phiMycoMar T7 transposon sequence. All reads from both strains 314 were then mapped to their corresponding reference sequences using the CLC-GWB default command 315 settings (mismatch cost 2, insertion cost 3, deletion cost 3, length fraction 0.5, similarity fraction 0.8). A 316 probabilistic variant analysis as implemented in the CLC-Genomics Workbench package [48] was then 317 performed (min coverage 10, 90% variant probability) to obtain the variants listed in Table 1. 318 To confirm these variants, PCR primers were designed flanking the mutations, after which the regions were 319 amplified (Phusion polymerase), purified (AMPure XP beads) and Sanger sequenced (VIB Genetics Service 320 Facility (http://www.vibgeneticservicefacility.be) with nested primers (see table below). PCR conditions were 321 as follows: 98°C for 3 minutes; 25 cycles of denaturation (98°C for 20 seconds), annealing (69°C for 20 322 seconds) and extension (72°C for 1 minute); 72°C for 5 minutes. The primer details are given in Suppl. Table 323 2. 324 4.3 RT-qPCR analysis 325 10 mL of M. bovis BCG cultures (grown in standard 7H9 medium until an OD600 of 0.8 – 1.0) were centrifuged 326 and the pellets were washed once with 0.5% Tween80 solution in PBS. The pellet was then resuspended in 327 500 µl of RLT buffer (RNeasy Mini Kit, Qiagen; supplemented with β-ME). The cells were disrupted with glass 328 beads in Retsch MM2000 bead beater at 4°C in screw-cap tubes (pre-baked at 150°C). After centrifugation (2 329 min, 13,000 rpm, 4°C), the supernatant was transferred to a fresh eppendorf tube. To recover the lysate 330 trapped in between the beads, 800 µl of chloroform was added to the beads and centrifuged, after which the 16 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 331 upper phase was transferred to the same eppendorf tube as before. Then, 1 volume of Acid 332 Phenol/Chloroform (Ambion) was added, incubated for 2 minutes and centrifuged (5 min, 13,000 rpm, 4°C). 333 The upper aqueous phase was transferred to a fresh eppendorf tube and this last step was repeated once. 334 An equal volume of 70% ethanol was added and the sample was transferred to an RNeasy spin column 335 (RNeasy Mini Kit, Qiagen). The kit manufacturer’s instructions were followed to purify the RNA. After elution 336 in RNase-free water (30 µl), an extra DNase digestion was performed with DNaseI (10 U of enzyme, 50 µl 337 reaction volume). Then an extra clean-up step was performed with the RNeasy Mini Kit (Qiagen). Finally, the 338 RNA concentration was determined on a Nanodrop instrument. Typically, RNA yields of 2-3 µg were obtained 339 with this method. 340 cDNA was prepared from 1 μg of DNase-treated RNA using the iScript Synthesis Kit (BioRad) and a control 341 reaction lacking reverse transcriptase was included for each sample. The RT-PCR program was as follows: 10 342 minutes at 25°C, 30 minutes at 42°C, 5 minutes at 85°C, and cooling down to 12°C. 343 Real time quantitative PCR was done on a LightCycler 480 (Roche Diagnostics) using the SensiFast SYBR- 344 NoRox kit (BioLine), in triplicate for each cDNA sample, on a 384-multiwell plate, with 1 ng of cDNA in a total 345 volume of 10 μL. Primers were used at a final concentration of 10 μM. All primer pairs were generated using 346 Primer3Plus (http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi) as described [49]. All gene 347 expression values were normalized using the geometric mean of the GroEL gene and 16S rRNA. 348 Determination of amplification efficiencies and conversion of raw Cq values to normalized relative quantities 349 (NRQ) was performed using the qbasePLUS software (Biogazelle). Statistical analysis of the NRQs was done 350 with the Prism 6 software package using a two-tailed t-test. Primer details are given in Suppl. Table 2. 351 4.4 RNA-Seq analysis 352 M. bovis BCG cultures (sapM::Tn BCG mutant and WT BCG 1721) were grown in standard 7H9 medium in 353 triplicates until an OD600 of 0.8 – 1.0 and RNA was prepared as described for the RT-qPCR analysis. This RNA 354 was then depleted of ribosomal RNA (Ribo-Zero rRNA removal kit for gram-positive bacteria) and the 355 remaining RNA was prepared for Illumina sequencing (TruSeq stranded total RNA preparation kit). This library 17 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 356 preparation takes into account the strandedness of the RNA and is able to distinguish between sense and 357 antisense transcripts. The libraries were sequenced on an Illumina NextSeq500 (75 bp single-end). The library 358 preparation and sequencing was performed by the VIB Nucleomics Core facility (www.nucleomics.be). The 359 data was analyzed using the standard RNA-seq analysis workflow in CLC Genomics Workbench v7 (CLC-GWB 360 [48], http://www.clcbio.com/products/clc-genomics-workbench/). In short, raw sequencing reads were 361 quality trimmed and mapped to the reference genome (M. bovis BCG Pasteur str. 1173P2), taking into 362 account the read directionality. The total number of unique reads per gene (averaged over the triplicate 363 samples) of both the WT BCG and the sapM::Tn BCG strain were then compared using the EdgeR statistical 364 test [50,51]. 365 4.5 Recombinant SapM production and anti-SapM antibody preparation 366 To express SapM in E. coli, we constructed an expression vector pLSPelB-SapM mature-His, being pLSAH36 367 carrying a PelB signal sequence, mature SapM without signal sequence and a C-terminal His tag. This was 368 transformed to the E. Coli BL21+pICA2 expression strain. E. coli was grown in LB medium supplemented with 369 ampicillin and kanamycin at 28°C. Expression was induced with 1 mM IPTG overnight (ON) at 18°C. Bacteria 370 were centrifuged and the pellet was sonicated in sonication buffer (50 mM Tris, 100 mM NaCl, 1 mM PMSF, 371 pH8). After centrifugation at 18000 rpm for 1h, supernatant was removed and inclusion bodies were 372 extracted in extraction buffer (20 mM NaH2PO4, 0,5 M NaCl, 5 mM β-mercaptoethanol, 6 M guanidine 373 hydrochloride, 5 mM imidazole, pH 7.5) ON at 4°C. Denatured protein was purified using nickel sepharose 374 and eluted in elution buffer (20 mM NaH2PO4, 50 mM NaCl, 5 mM β-mercaptoethanol, 8 M guanidine 375 hydrochloride, 400 mM imidazole, pH 7.5) and a final desalting was performed on a sephadex G25. The 376 purification was performed by the VIB Protein Core facility. The rabbit polyclonal anti-SapM antibody was 377 generated by immunization of 2 rabbits (CER groupe, Belgium) with purified SapM protein (625µg/mL, in the 378 text named anti-SapM; denatured, full-length protein produced in E. coli as described above). Immunizations 379 and bleedings were done following the standard CER protocol. In brief, rabbits were subcutaneously injected 380 on days 0, 14, 28 and 56 with 500µL antigen/500µL adjuvant. Bleedings were performed on day 0 18 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 381 (preimmune 2mL), day 38 (small test sample 2mL), day 66 (large bleed 20+2mL) and day 80 (final bleeding). 382 Sample from the final bleeding of rabbit 2 is used in the experiments described here. The rabbit polyclonal 383 anti-SapM-C-term antibody (Anti-PepC) was raised against a synthetic C-terminal peptide (peptide synthesis 384 (CYATNAPPITDIWGD, terminal modification: amidation (C-terminal) and conjugation to KLH) and 385 immunizations performed by GenScript). Immunizations and bleedings were done following the standard 386 GenScript protocol. In brief, 2 rabbits were subcutaneously injected on days 0 and day 14. First bleedings 387 were performed on day 0 (pre-immune, 2mL) and 21 (1mL). A third immunization was performed on day 35. 388 A second bleeding was done on day 42 (1mL). A fourth immunization was done on day 56. Final bleeding was 389 done on day 63. Only pre-immune serum and final anti-sera was provided by Genscript. 390 4.6 SDS-PAGE and western blotting 391 M. bovis BCG cultures were grown in 7H9 medium without OADC (as the albumin in the OADC supplement 392 would otherwise interfere on SDS-PAGE), but supplemented with 0.2% glucose, 0.2% glycerol and 0.05% of 393 Tween80. The supernatant samples were precipitated by first adding sodium deoxycholate (DOC, 0.1% final 394 concentration), incubating 15 min on ice and then adding trichloroacetic acid (TCA, 7.7% final concentration) 395 followed by 30 min of incubation on ice. The samples were centrifuged and the pellet was washed twice with 396 acetone and dissolved again in PBS with Laemmli loading dye. Precipitated protein samples corresponding to 397 200 µl of culture medium were analyzed by SDS-PAGE and subsequent western blotting. The membranes 398 were blocked overnight (4°C) with PBST (PBS + 0.05% Tween-20) containing 5% milk powder. To visualize the 399 SapM protein, the blots were incubated for 1h at room temperature with a rabbit anti-SapM polyclonal 400 antibody (1:5,000 dilution in PBST), washed 3 times with PBST and then incubated for 1h with a secondary 401 goat anti-rabbit DyLight 800 conjugated antibody (30 ng/ml in PBST). After the final washing steps, the 402 membranes were scanned with a LI-COR Odyssey system. 403 4.7 ELISA 404 M. bovis BCG culture supernatant samples (grown in 7H9; 50 µl) or control samples (0.25 µg/ml of purified 405 SapM protein) were diluted 1:2 in 100 mM of carbonate buffer (3.03 g of Na2CO3 and 6 g of NaHCO3 in 1 L of 19 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 406 water; pH 9.6) and coated on 96-well maxisorp plates (Nunc) (overnight at 4°C). The plate was then washed 407 three times with PBST (PBS + 0.05% Tween-20). The plate was blocked with 1% of BSA in PBS (2 hours at room 408 temperature (RT)). After plate washing, the rabbit polyclonal anti-SapM or anti-PepC antibody was added 409 (diluted 1:5000 in PBS with 0.1% Tween-20 and 0.1% goat serum) (1 hour at RT). After washing, the plate was 410 incubated with the secondary HRP-labeled donkey anti-rabbit antibody (GE healthcare) (diluted 1:5000 in 411 PBS with 1% BSA) (1 hour at RT). After washing, 100 µl of the TMB substrate was added to the plate (Becton 412 Dickinson), incubated at RT for 30 minutes, after which 50 µl of 2N H2SO4 was added to stop the reaction. 413 The optical density at 450 nm was then determined with a bio-rad plate reader (wavelength correction at 414 655 nm). 415 4.8 Phosphatase assay 416 A 150 µl reaction mixture, containing 50 µl of M. bovis BCG culture supernatant samples, 85 µl of sodium 417 acetate (0.2 M; pH 7.0) and 15 µl of para-nitrophenylphosphate (pNPP, 50 mM), was incubated for 6 hours 418 or overnight at 37°C. Finally, 50 µl of stop buffer (1 M of Na2CO3) was added and the optical density at 415 419 nm was determined with a bio-rad plate reader. 420 4.9 Laboratory animals 421 Female C57BL/6J mice (Janvier), C57BL/6J x Balb/c mice (F1, Harlan) and CB-17 SCID mice (Charles River) 422 were housed under specific pathogen-free conditions in micro-isolator units. At the beginning of the 423 experiment, the mice were 7–8 weeks old. All experiments were approved by and performed according to 424 the guidelines of the ethical committee of Ghent University, Belgium. 425 4.10 426 Eight-week-old female CB-17 SCID mice (Charles River Laboratories, 7 or 8 mice/group as indicated) were 427 immunized intravenously via tail vein injection with 100 μL (3x10 6 versus 3x107 cfu) of WT BCG Pasteur or 428 sapM::Tn BCG mutant (diluted in PBS). The control group received PBS (pH 7.5). The animals were observed 429 and weighed initially on a ~2-weekly basis and from day 82 post-infection on, ~every 2 days (until day 201 SCID safety study 20 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 430 post-infection). The differences of time to survival between the different groups were compared using log- 431 rank test. Weight loss of 20% of initial body weight was set as the ethical endpoint. 432 4.11 433 Bone marrow cells from C57BL/6J mice were harvested and differentiated to macrophages in DMEM 434 containing 20% of L929-cell supernatant, 10% of heat-inactivated FCS, 0.4 mM sodiumpyruvate, 50 µM β- 435 mercaptoethanol and 0.1 mM MEM non-essential amino acids for 7 days. Cells were cultured in antibiotic- 436 free medium at all times. To analyze the survival/replication of M. bovis BCG following infection of 437 macrophages, bone marrow-derived macrophages (BM-DMs, differentiated as described in [52]) were either 438 left untreated or infected with WT BCG versus sapM::Tn BCG (MOI 10:1) in infection medium (DMEM + 439 supplements described earlier) for 4 hours. Cells were then washed 2x with pre-warmed PBS and 1x with pre- 440 warmed medium to remove extracellular bacteria. Fresh medium was added and cells were further incubated 441 for the indicated time points. BM-DMs were lysed in cell lysis buffer (0.05% v/v SDS, 200mM NaCl, 10mM 442 Tris-HCl pH7.5, 5mM EDTA, 10% glycerol) at the indicated times and serial dilutions were plated onto 443 Middlebrook 7H10 agar plates followed by incubation at 37°C. 444 4.12 445 Three experiments were performed. At the age of 8 weeks, female C57BL/6J x Balb/c (F1, Harlan) mice were 446 infected s.c. (at the base of the tail) with 2x106 CFU of WT BCG or sapM::Tn BCG mutant in 100 µl of PBS. At 447 day 8, 14 and 28 post-infection, 10 mice/group were sacrificed by cervical dislocation; inguinal and brachial 448 LNs were removed aseptically and homogenized in PBS-0.5% Tween 20 + EDTA-free complete protease 449 inhibitor (Roche). Neat, 1/10, 1/100 and 1/1000 dilutions were made of which 50µL was plated in duplicate 450 on 7H10-OADC agar and the colonies were counted after 21 days. The CFU counts were calculated from the 451 highest dilution showing distinct colonies. 452 In another experiment, at the age of 8 weeks, female C57BL/6J x Balb/c (F1, Harlan) mice were infected 453 intravenously (i.v.) with 2x106 CFU WT BCG or sapM::Tn BCG mutant in 200 µl of PBS. At 24h and 3 weeks Cell culture and in vitro infection Vaccination with M. bovis BCG and analysis of in vivo BCG CFU dynamics 21 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 454 post-infection, 8 mice/group were sacrificed; spleens and lungs were removed aseptically, homogenized in 455 PBS-0.5% Tween 20 + EDTA-free complete protease inhibitor (Roche) and analyzed for CFU counts as above. 456 Again, neat, 1/10, 1/100 and 1/1000 (lung) and neat, 1/10, 1/100, 1/1000, 1/10 000, 1/100 000 (spleen) 457 dilutions were made of which 50µL was plated in duplicate on 7H10-OADC agar. Additionally, 6h, 24h and 458 48h post-infection, blood was taken retro-orbitally and collected in BD microtainer SST tubes for serum 459 preparation, which was frozen at -20°C until assay. Serum cytokines were determined by bioplex (BioPlex Pro 460 Cytokine Express Assay, Biorad). 461 In at third experiment, at the age of 8 weeks, female C57BL/6J (Janvier) mice were infected s.c. (at the base 462 of the tail) with 2x106 CFU WT BCG or sapM::Tn BCG mutant in 100 µl PBS. At day7, 14 and 28 post-infection, 463 6 mice/group were sacrificed; inguinal and brachial LNs were removed aseptically and analyzed for CFU 464 counts as above. 465 4.13 466 To complement the sapM::Tn BCG mutant strain with sapM gene, we constructed an integration plasmid 467 carrying sapM as follows: the SapM operon was amplified by PCR using primers SapMPo-XbaI 468 (CACTCTAGACAGTGGGTGGTCAACGACA) and Rv3311HindIII (GTTAAGCTTCCGGCTGACGGTGCCTATTC) from 469 gDNA of BCG Pasteur. The PCR fragment was cloned in pMV261Hyg (kindly provided by Prof. E. Rubin) via 470 XbaI and HindIII to form pMV261HygSapMoperon. The vector pMV261HygSapMoperon was digested with 471 XbaI and ApaLI and the integrative vector pMV306Kan (kindly provided by Prof. W.R. Jacobs jr.) was digested 472 with NheI and ApaLI, after which the appropriate fragments were purified from gel and ligated together to 473 form pMV306HygSapMoperon, an integrating vector with the SapM operon (Plasmid map in Suppl. Fig. 7). 474 The plasmid pMV306HygSapMoperon was electroporated into the sapM::Tn BCG mutant strain [53]. 475 Hygromycin resistant colonies were selected and surviving clones were tested on a phosphatase assay. 476 4.14 477 C57BL/6J mice were immunized s.c. at the base of the tail with WT BCG or sapM::Tn BCG mutant, either with 478 a high dose (2x106 cfu, 5-10 mice/group) or a low dose (2x105 cfu, 10 mice/group). Seven and 14 days later, 22 Complementation sapM::Tn BCG mutant Vaccination with M. bovis BCG and analysis of DC phenotype bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 479 the mice were sacrificed, and inguinal LNs, brachial LNs and spleens were isolated. Cells were prepared by 480 crushing the organs on a 70µm cell strainer and labelled with an optimized DC-analysis antibody panel and 481 analyzed by flow cytometry. Fluorochrome-conjugated mAbs specific for mouse CD11c (clone N418, 482 ebioscience), CD8 (clone 53-6.7, BD), CD40 (clone 1C10, ebioscience), CD80 (clone 16-10A1, Biolegend), Ly- 483 6C (clone AL-21, BD), MHCII (clone M5/114.15.2, ebioscience), CD103 (clone 2E7, ebiosciene), F4/80 (clone 484 BM8, ebioscience), CD11b (clone M1/70.15, Life Technologies), Ly-6G (clone RB6-8C5, ebioscience) were 485 used. Dilutions of the antibodies are given in Suppl. Table 3 and gating strategy is given in Suppl. Fig. 8. 486 4.15 487 C57BL/6J mice were vaccinated s.c. at the base of the tail with either a high dose (2x10 6 cfu) or a low dose 488 (2x105 cfu, in 100µL of PBS) WT BCG or sapM::Tn BCG mutant. 14 days after infection, mice were killed by 489 cervical dislocation and inguinal + brachial LNs and spleens were removed aseptically. Cells were prepared 490 by crushing the organs on a 70 µm cell strainer and adjusted to a concentration of 1x106 (LN) or 2x106 (spleen) 491 cells/ mL in either flat-bottom 96-well microwell plates (Bioplex/ELISA) or round-bottom plates (flow 492 cytometry) (Nunc), in RPMI 1640 medium, supplemented with 10% fetal calf serum, L-glutamine (0.03%), 0,4 493 mM sodiumpyruvate, 0,1 mM non-essential amino acids and 50 μM β-mercaptoethanol. Cells were 494 restimulated with purified native Ag85A (Rv3804c) from M. tb, strain H37Rv (15 µg/mL in PBS, obtained 495 through BEI Resources, NIAID, NIH), purified protein derivative (PPD, 10 μg/mL, produced at Scientific 496 Institute for Public Health, Brussels, Belgium) and anti-CD3/28 (10 µg/mL, BD) in a volume of 100 μL added 497 to 100 μL of cell suspension. For surface and intracellular staining, cells were incubated at 37°C in a humidified 498 CO2 incubator, and after 1h, brefeldin was added (10 µg/mL-Sigma Aldrich) for another 5 hours. LNs and 499 splenocytes were subsequently labelled with an optimized antibody panel for T cell phenotyping and 500 analyzed by flow cytometry. Fluorochrome-conjugated mAbs specific for mouse CD3 (clone 17A2, BD), NK1.1 501 (clone PK136, BD), CD4 (clone RM4-5, Invitrogen), CD8 (clone 53-6.7, BD), CD335 (clone 29A1.4, BD), TCRγδ 502 (clone GL-3, ebioscience), IL17 (clone TC11-18H10, BD), IFNγ (clone XMG1.2, BD) and TNF (clone MP6-XT22, 503 ebioscience) were used. Fixable viability dye eFluor 780 (eBioscience) was used as a live/dead marker. Vaccination with M. bovis BCG for T cell phenotyping and antigen-specific cytokine determination 23 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 504 Dilutions of the antibodies and gating strategy are given in Suppl. Materials and Methods. For determination 505 of cytokine levels in the supernatant, the latter was harvested after 48h post-stimulation. Supernatants were 506 frozen at -20°C until assay. Cytokines were determined by bioplex (BioPlex Pro Cytokine Express Assay, 507 Biorad) and ELISA (mouse IFNγ duoset Elisa, R&D systems) as indicated in the text. 508 4.16 509 Cells were first blocked by adding Fc block (purified rat anti-mouse CD16/CD32 monoclonal Ab, clone 2.4G2, 510 BD, 0.1µg/sample) and incubating for 15 min at 4°C. Staining is performed for 20 min at 4 °C in the dark. For 511 intracellular staining, cells were incubated for 5 h with 10 µg/mL brefeldin A (10 µg/mL- Sigma Aldrich) before 512 fixation and permeabilization (Fixation-Permeabilization concentrate, Permeabilization buffer, ebioscience). 513 In all analyses, following doublet exclusion, live cells were identified using a fixable viability dye (Molecular 514 Probe, Life Technologies). Data were acquired on an LSR II equipped with three lasers (488, 635, 405nm) 515 (APC panel) or Fortessa with 4 lasers (405, 488, 561, 633 nm) (T cell panel) (BD Biosciences) and analyzed 516 using FlowJo software (Tree Star, Ashland, OR). 517 4.17 518 Results are presented as means ± standard error of the mean (SEM) unless otherwise stated and groups 519 were compared using statistical tests as mentioned in the text, using Prism Software (GraphPad Software, 520 San Diego, CA). 521 5. Acknowledgements 522 This work was supported by an ERC Consolidator grant to N.C. (GlycoTarget; 616966) and a GOA grant. N.F. 523 and D.V. were postdoctoral fellows and K.V. and K.B. were predoctoral fellows of FWO (Fonds 524 Wetenschappelijk Onderzoek-Vlaanderen). We thank Dr. Peter Sander (Institute for Medical Microbiology, 525 Faculty of Medicine, University of Zurich) for providing us with the M. bovis BCG strain (strain 1721, RpsL, 526 K43R), Dr. S. Alonso (Dept, Microbiology, National University of Singapore) for providing us the optimal 527 protocol for electroporation of M. bovis BCG and Dr. K. Huygen (WIV, Brussels) for providing us with the PPD. 24 Flow cytometry analysis Statistical analysis bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 528 We thank Prof. Dr. E. Rubin (Dept, Immunology and Infectious diseases, Harvard School of Public Health, 529 Boston) for providing us the pMV261Hyg vector and Prof. W.R. Jacobs Jr (Albert Einstein College of Medicine, 530 New York) for providing us the pMV306Kan vector. We thank the VIB Nucleomics core (www.nucleomics.be) 531 and Genomics Core (UZ – K.U. Leuven, gc.uzleuven.be) facilities for the Illumina MiSeq and HiSeq sequencing 532 services. We thank Jannick Leoen from the VIB Protein Core facility (https://corefacilities.vib.be/psf) for 533 purification of SapM protein. The following reagent was obtained through BEI Resources, NIAID, NIH: Ag85 534 Complex, Purified Native Protein from Mycobacterium tuberculosis, Strain H37Rv, NR-14855. 535 N.C and N.F. are named inventors on a patent covering the use of SapM mutation as a vaccine improvement 536 strategy. 537 6. Author contributions 538 NF coordinated the project, designed and performed experiments, analyzed and interpreted data and wrote 539 the manuscript. KV analyzed genome sequencing and RNAseq experiments, performed RT-qPCR experiments 540 and co-wrote the manuscript. EH designed and performed in vivo experiments, analyzed and interpreted 541 data. EP performed all in vitro experiments and helped in performing in vivo experiments. DV and KB helped 542 in performing in vivo experiments. PT and AV made the sapM::Tn:complementation mutant. NC coordinated 543 the project, interpreted data and co-wrote the manuscript. 544 25 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 545 7. References 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] WHO | Global tuberculosis report 2018. WHO n.d. http://www.who.int/tb/publications/global_report/en/. Nieuwenhuizen NE, Kaufmann SHE. Next-Generation Vaccines Based on Bacille Calmette–Guérin. Front Immunol 2018;9. doi:10.3389/fimmu.2018.00121. Rodrigues LC, Diwan VK, Wheeler JG. Protective effect of BCG against tuberculous meningitis and miliary tuberculosis: a meta-analysis. Int J Epidemiol 1993;22:1154–8. Trunz BB, Fine P, Dye C. Effect of BCG vaccination on childhood tuberculous meningitis and miliary tuberculosis worldwide: a meta-analysis and assessment of cost-effectiveness. Lancet Lond Engl 2006;367:1173–80. doi:10.1016/S0140-6736(06)68507-3. Colditz GA, Brewer TF, Berkey CS, Wilson ME, Burdick E, Fineberg HV, et al. Efficacy of BCG vaccine in the prevention of tuberculosis. Meta-analysis of the published literature. JAMA 1994;271:698–702. Hatherill M, Tait D, McShane H. Clinical Testing of Tuberculosis Vaccine Candidates. Microbiol Spectr 2016;4. doi:10.1128/microbiolspec.TBTB2-0015-2016. Triccas JA, Counoupas C. Novel vaccination approaches to prevent tuberculosis in children. Pneumonia 2016;8. doi:10.1186/s41479-016-0020-z. Horwitz MA, Harth G. A New Vaccine against Tuberculosis Affords Greater Survival after Challenge than the Current Vaccine in the Guinea Pig Model of Pulmonary Tuberculosis. Infect Immun 2003;71:1672–9. doi:10.1128/IAI.71.4.1672-1679.2003. Martin C, Williams A, Hernandez-Pando R, Cardona PJ, Gormley E, Bordat Y, et al. The live Mycobacterium tuberculosis phoP mutant strain is more attenuated than BCG and confers protective immunity against tuberculosis in mice and guinea pigs. Vaccine 2006;24:3408–19. doi:10.1016/j.vaccine.2006.03.017. Sun R, Skeiky YAW, Izzo A, Dheenadhayalan V, Imam Z, Penn E, et al. Novel recombinant BCG expressing perfringolysin O and the over-expression of key immunodominant antigens; pre-clinical characterization, safety and protection against challenge with Mycobacterium tuberculosis. Vaccine 2009;27:4412–23. doi:10.1016/j.vaccine.2009.05.048. Brandt L, Skeiky YAW, Alderson MR, Lobet Y, Dalemans W, Turner OC, et al. The Protective Effect of the Mycobacterium bovis BCG Vaccine Is Increased by Coadministration with the Mycobacterium tuberculosis 72-Kilodalton Fusion Polyprotein Mtb72F in M. tuberculosis-Infected Guinea Pigs. Infect Immun 2004;72:6622–32. doi:10.1128/IAI.72.11.6622-6632.2004. Bertholet S, Ireton GC, Ordway DJ, Windish HP, Pine SO, Kahn M, et al. A Defined Tuberculosis Vaccine Candidate Boosts BCG and Protects Against Multidrug Resistant Mycobacterium tuberculosis. Sci Transl Med 2010;2:53ra74. doi:10.1126/scitranslmed.3001094. Festjens N, Bogaert P, Batni A, Houthuys E, Plets E, Vanderschaeghe D, et al. Disruption of the SapM locus in Mycobacterium bovis BCG improves its protective efficacy as a vaccine against M. tuberculosis. EMBO Mol Med 2011;3:222–34. doi:10.1002/emmm.201000125. Puri RV, Reddy PV, Tyagi AK. Secreted acid phosphatase (SapM) of Mycobacterium tuberculosis is indispensable for arresting phagosomal maturation and growth of the pathogen in guinea pig tissues. PloS One 2013;8:e70514. doi:10.1371/journal.pone.0070514. Bahal RK, Mathur S, Chauhan P, Tyagi AK. An attenuated quadruple gene mutant of Mycobacterium tuberculosis imparts protection against tuberculosis in guinea pigs. Biol Open 2017;7. doi:10.1242/bio.029546. Zulauf KE, Sullivan JT, Braunstein M. The SecA2 pathway of Mycobacterium tuberculosis exports effectors that work in concert to arrest phagosome and autophagosome maturation. PLoS Pathog 2018;14. doi:10.1371/journal.ppat.1007011. Saikolappan S, Estrella J, Sasindran SJ, Khan A, Armitige LY, Jagannath C, et al. The fbpA/sapM Double Knock Out Strain of Mycobacterium tuberculosis Is Highly Attenuated and Immunogenic in Macrophages. PLoS ONE 2012;7. doi:10.1371/journal.pone.0036198. 26 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 [18] Nunes-Alves C, Booty MG, Carpenter SM, Jayaraman P, Rothchild AC, Behar SM. In search of a new paradigm for protective immunity to TB. Nat Rev Microbiol 2014;12:289–99. doi:10.1038/nrmicro3230. [19] Chin KL, Anis FZ, Sarmiento ME, Norazmi MN, Acosta A. Role of Interferons in the Development of Diagnostics, Vaccines, and Therapy for Tuberculosis. J Immunol Res 2017;2017. doi:10.1155/2017/5212910. [20] Cowley SC, Elkins KL. CD4+ T Cells Mediate IFN-γ-Independent Control of Mycobacterium tuberculosis Infection Both In Vitro and In Vivo. J Immunol 2003;171:4689–99. doi:10.4049/jimmunol.171.9.4689. [21] Gallegos AM, van Heijst JWJ, Samstein M, Su X, Pamer EG, Glickman MS. A Gamma Interferon Independent Mechanism of CD4 T Cell Mediated Control of M. tuberculosis Infection in vivo. PLoS Pathog 2011;7. doi:10.1371/journal.ppat.1002052. [22] Orr MT, Windish HP, Beebe EA, Argilla D, Huang P-WD, Reese VA, et al. Interferon γ and Tumor Necrosis Factor Are Not Essential Parameters of CD4+ T-Cell Responses for Vaccine Control of Tuberculosis. J Infect Dis 2015;212:495–504. doi:10.1093/infdis/jiv055. [23] Orr MT, Beebe EA, Hudson T, Argilla D, Huang P-WD, Reese VA, et al. Mucosal delivery switches the response to an adjuvanted tuberculosis vaccine from systemic TH1 to tissue-resident TH17 responses without impacting the protective efficacy. Vaccine 2015;33:6570–8. doi:10.1016/j.vaccine.2015.10.115. [24] Sallin MA, Kauffman KD, Riou C, Bruyn ED, Foreman TW, Sakai S, et al. Host resistance to pulmonary Mycobacterium tuberculosis infection requires CD153 expression. Nat Microbiol 2018:1. doi:10.1038/s41564-018-0231-6. [25] Vandewalle K, Festjens N, Plets E, Vuylsteke M, Saeys Y, Callewaert N. Characterization of genomewide ordered sequence-tagged Mycobacterium mutant libraries by Cartesian Pooling-Coordinate Sequencing. Nat Commun 2015;6. doi:10.1038/ncomms8106. [26] Master SS, Davis AS, Rampini SK, Keller C, Ehlers S, Springer B, et al. Mycobacterium tuberculosis Prevents Inflammasome Activation. Cell Host Microbe 2008;3:224–32. doi:10.1016/j.chom.2008.03.003. [27] Titgemeyer F, Amon J, Parche S, Mahfoud M, Bail J, Schlicht M, et al. A Genomic View of Sugar Transport in Mycobacterium smegmatis and Mycobacterium tuberculosis. J Bacteriol 2007;189:5903– 15. doi:10.1128/JB.00257-07. [28] Sassetti CM, Boyd DH, Rubin EJ. Genes required for mycobacterial growth defined by high density mutagenesis. Mol Microbiol 2003;48:77–84. doi:10.1046/j.1365-2958.2003.03425.x. [29] Infante E, Aguilar LD, Gicquel B, Pando RH. Immunogenicity and protective efficacy of the Mycobacterium tuberculosis fadD26 mutant. Clin Exp Immunol 2005;141:21–8. doi:10.1111/j.13652249.2005.02832.x. [30] Hotter GS, Wards BJ, Mouat P, Besra GS, Gomes J, Singh M, et al. Transposon mutagenesis of Mb0100 at the ppe1-nrp locus in Mycobacterium bovis disrupts phthiocerol dimycocerosate (PDIM) and glycosylphenol-PDIM biosynthesis, producing an avirulent strain with vaccine properties at least equal to those of M. bovis BCG. J Bacteriol 2005;187:2267–77. doi:10.1128/JB.187.7.2267-2277.2005. [31] Reddy TBK, Riley R, Wymore F, Montgomery P, DeCaprio D, Engels R, et al. TB database: an integrated platform for tuberculosis research. Nucleic Acids Res 2009;37:D499–508. doi:10.1093/nar/gkn652. [32] Saleh MT, Belisle JT. Secretion of an Acid Phosphatase (SapM) by Mycobacterium tuberculosis That Is Similar to Eukaryotic Acid Phosphatases. J Bacteriol 2000;182:6850–3. doi:10.1128/JB.182.23.68506853.2000. [33] Grundner C, Ng H-L, Alber T. Mycobacterium tuberculosis Protein Tyrosine Phosphatase PtpB Structure Reveals a Diverged Fold and a Buried Active Site. Structure 2005;13:1625–34. doi:10.1016/j.str.2005.07.017. [34] Sellers RS, Clifford CB, Treuting PM, Brayton C. Immunological Variation Between Inbred Laboratory Mouse Strains Points to Consider in Phenotyping Genetically Immunomodified Mice. Vet Pathol Online 2012;49:32–43. doi:10.1177/0300985811429314. 27 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 [35] Doisne J-M, Soulard V, Bécourt C, Amniai L, Henrot P, Havenar-Daughton C, et al. Cutting Edge: Crucial Role of IL-1 and IL-23 in the Innate IL-17 Response of Peripheral Lymph Node NK1.1− Invariant NKT Cells to Bacteria. J Immunol 2011;186:662–6. doi:10.4049/jimmunol.1002725. [36] Cooper AM, Mayer-Barber KD, Sher A. Role of innate cytokines in mycobacterial infection. Mucosal Immunol 2011;4:252–60. doi:10.1038/mi.2011.13. [37] Arbues A, Aguilo JI, Gonzalo-Asensio J, Marinova D, Uranga S, Puentes E, et al. Construction, characterization and preclinical evaluation of MTBVAC, the first live-attenuated M. tuberculosis-based vaccine to enter clinical trials. Vaccine 2013;31:4867–73. doi:10.1016/j.vaccine.2013.07.051. [38] Chen JM, Islam ST, Ren H, Liu J. Differential productions of lipid virulence factors among BCG vaccine strains and implications on BCG safety. Vaccine 2007;25:8114–22. doi:10.1016/j.vaccine.2007.09.041. [39] Leung AS, Tran V, Wu Z, Yu X, Alexander DC, Gao GF, et al. Novel genome polymorphisms in BCG vaccine strains and impact on efficacy. BMC Genomics 2008;9:413. doi:10.1186/1471-2164-9-413. [40] Naka T, Maeda S, Niki M, Ohara N, Yamamoto S, Yano I, et al. Lipid phenotype of two distinct subpopulations of Mycobacterium bovis Bacillus Calmette-Guerin Tokyo 172 substrain. J Biol Chem 2011;286:44153–61. doi:10.1074/jbc.M111.310037. [41] Brosch R, Gordon SV, Garnier T, Eiglmeier K, Frigui W, Valenti P, et al. Genome plasticity of BCG and impact on vaccine efficacy. Proc Natl Acad Sci U S A 2007;104:5596–601. doi:10.1073/pnas.0700869104. [42] Sander P, Clark S, Petrera A, Vilaplana C, Meuli M, Selchow P, et al. Deletion of zmp1 improves Mycobacterium bovis BCG-mediated protection in a guinea pig model of tuberculosis. Vaccine 2015;33:1353–9. doi:10.1016/j.vaccine.2015.01.058. [43] Results from Innovative Phase 2 Tuberculosis Vaccine Trial Offer Potential for New BCG Revaccination. Aeras n.d. http://www.aeras.org//. [44] Ernst JD. Mechanisms of M. tuberculosis Immune Evasion as Challenges to TB Vaccine Design. Cell Host Microbe 2018;24:34–42. doi:10.1016/j.chom.2018.06.004. [45] Brighenti S, Lerm M. How Mycobacterium tuberculosis Manipulates Innate and Adaptive Immunity – New Views of an Old Topic. Underst Tuberc - Anal Orig Mycobacterium Tuberc Pathog 2012. doi:10.5772/29651. [46] Kalia V, Sarkar S, Ahmed R. CD8 T-cell memory differentiation during acute and chronic viral infections. Adv Exp Med Biol 2010;684:79–95. [47] Belisle JT, Sonnenberg MG. Isolation of genomic DNA from mycobacteria. Methods Mol Biol Clifton NJ 1998;101:31–44. doi:10.1385/0-89603-471-2:31. [48] CLC bio. CLC bio Genomics Workbench - Manual n.d. [49] D’haene B, Vandesompele J, Hellemans J. Accurate and objective copy number profiling using realtime quantitative PCR. Methods San Diego Calif 2010;50:262–70. doi:10.1016/j.ymeth.2009.12.007. [50] Robinson MD, McCarthy DJ, Smyth GK. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 2010;26:139–40. doi:10.1093/bioinformatics/btp616. [51] McCarthy DJ, Chen Y, Smyth GK. Differential expression analysis of multifactor RNA-Seq experiments with respect to biological variation. Nucleic Acids Res 2012;40:4288–97. doi:10.1093/nar/gks042. [52] Weischenfeldt J, Porse B. Bone Marrow-Derived Macrophages (BMM): Isolation and Applications. Cold Spring Harb Protoc 2008;2008:pdb.prot5080. doi:10.1101/pdb.prot5080. [53] Goude R, Roberts DM, Parish T. Electroporation of Mycobacteria. Mycobact. Protoc., Humana Press, New York, NY; 2015, p. 117–30. doi:10.1007/978-1-4939-2450-9_7. 28 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 688 8. Figures 689 Figure 1| Whole genome resequencing of the WT BCG Pasteur 1721 strain and of the sapM::Tn BCG mutant. 690 Detailed representation of the read mappings at the sapM locus for both strains. The red square marks the 691 TA dinucleotide site where the transposon is inserted in the mutant strain. 692 Figure 2| (A) Volcano plot comparing the sapM::Tn BCG mutant and the WT BCG. Genes with a > 2-fold 693 change and a p-value < 0.001 are depicted in red. (B) Mapped (sense) reads at the SapM locus in the WT 694 BCG vs sapM::Tn BCG condition. Only one replicate of both WT BCG and sapM::Tn BCG samples is depicted. 695 BCG_3374c = upp, BCG_3375 = sapM. (C) RT-PCR analysis of the sapM locus. RNA was prepared of cultures 696 of biological triplicates of the sapM::Tn BCG disruption mutant, as well as of the WT BCG. An RT-PCR on the 697 cDNA using primer sets directed against the sapM gene (blue bar) and the directly up- and downstream genes 698 (upp and BCG_3376, orange and green bars, respectively). The data presented here are averages (± SD) of 699 the three biological replicates. For each mutant, the targets are individually normalized to the transcription 700 levels in the WT BCG strain (grey dotted line set at an average relative quantity of 1). The * indicates 701 significance (p < 0.05) by a two tailed t-test comparing the transcription levels of each target in the mutant 702 strain to the same target in the WT BCG. (D-F) SapM protein and activity analysis. The sapM::Tn BCG mutant 703 and WT BCG 1721 were grown in 7H9 medium and supernatant samples were collected on various time 704 points (D6 (Day 6) – D15). (D) ELISA using an anti-SapM polyclonal antibody. (E) SDS-PAGE and western 705 blotting of WT BCG and sapM::Tn BCG mutant supernatant samples collected on day 15 and grown in 7H9 706 medium without OADC and supplemented with glucose only. The blots were developed with the anti-SapM 707 antibody. The processed SapM protein (= removal of the N-terminal signal sequence) is ± 28 kDa in size and 708 is only present in the WT BCG strain (black arrow). (F) In vitro phosphatase assay using p-Nitrophenyl 709 phosphate (pNPP) as a substrate to check the activity of the SapM enzyme. The plotted data for both ELISA 710 and phosphatase assay are averaged over 2 technical replicates and corrected for the background signal 711 induced by sterile 7H9 medium supplemented with 10% OADC. 29 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 712 Figure 3 | SapM::Tn BCG safety study in SCID mice. Immunocompromised SCID mice were infected 713 intravenously with 3x106 versus 3x107 cfu of the WT BCG or SapM::Tn BCG mutant and survival was 714 monitored. (Log-rank test; WT BCG vs. sapM::Tn; P=not significant). The control groups represent animals 715 vaccinated with PBS. 716 Figure 4 | In vivo and in vitro replication analysis: sapM::Tn BCG versus WT BCG 717 (A) F1 mice (Balb/c x C57BL/6J) were vaccinated s.c. with the WT BCG or sapM::Tn BCG (2x106 cfu, 10 718 mice/group). At day 8, 14 and 28 post-infection, mice were sacrificed and the number of bacteria in the 719 draining LNs was determined by cfu plating (Mann-Whitney test; ** P<0.01). (B) BM-DMs were infected with 720 WT BCG or sapM::Tn BCG (MOI 10:1) for 4 hours. Cells were washed and bacterial uptake was determined 721 4hr post-infection. Bacterial replication was analyzed by cfu plating 24, 48 and 72hr, 6 days post-infection. 722 (C) F1 mice were vaccinated i.v. with the WT BCG or sapM::Tn BCG (2x106 cfu, 8 mice/group). At 24h and 3 723 weeks post-infection, mice were sacrificed and the number of bacteria in the spleens and lungs was 724 determined by cfu plating (Mann-Whitney test; ** P<0.01). 725 Figure 5 | RT-PCR analysis of the sapM locus and SapM protein and activity analysis in the 726 sapM::Tn:complementation mutant versus WT BCG 727 (A) RNA was prepared of cultures of biological triplicates of the sapM::Tn BCG disruption mutant (red), the 728 WT BCG (black) and the sapM::Tn:complementation BCG mutant (green). An RT-PCR on the cDNA used 729 primer sets directed against the sapM gene (sapM b primers - left panel) and the directly upstream gene upp 730 (upp2 primers - right panel). The data presented here are averages (± SD) of three technical replicates. For 731 each mutant, the targets are individually normalized to the transcription levels in the WT BCG strain (grey 732 dotted line set at an average relative quantity of 1). The * indicates significance (p < 0.05) by a t-test 733 comparing the transcription levels of each target in the mutant strain to the same target in the WT BCG. (B- 734 D) The sapM::Tn BCG mutant, WT BCG 1721, sapM::Tn:complementation BCG mutant and sapM::Tn:empty 735 vector were grown in 7H9 medium and supernatant samples were collected on various time points (D1 (Day 736 1) – D15). (B) ELISA using an anti-SapM polyclonal antibody. The plotted data for both ELISA and phosphatase 30 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 737 assay is averaged over 2 technical replicates and corrected for the background signal induced by sterile 7H9 738 medium supplemented with 10% OADC. (C) SDS-PAGE and western blotting of indicated supernatant samples 739 collected on day 10 and grown in 7H9 medium without OADC and supplemented with glucose only. The blots 740 were developed with the anti-SapM antibody. The processed SapM protein (= removal of the N-terminal 741 signal sequence) is ± 28 kDa in size and is only present in the WT BCG strain and sapM::Tn:complementation 742 mutant. (D) In vitro phosphatase assay using p-Nitrophenyl phosphate (pNPP) as a substrate to check the 743 activity of the SapM enzyme. 744 Figure 6| In vivo replication analysis: sapM::Tn BCG versus sapM::Tn:complementation BCG versus WT BCG 745 (A) C57BL/6J mice were vaccinated s.c. with the WT BCG, sapM::Tn BCG or sapM::Tn:complementation BCG 746 (2x106 cfu, 6 mice/group). At day 8, 14 and 28 post-infection, mice were sacrificed and the number of bacteria 747 in the draining LNs was determined by cfu plating (Mann-Whitney test; ** P<0.01). (B) C57BL/6J mice were 748 vaccinated i.v. with the WT BCG, sapM::Tn BCG or sapM::Tn:complementation BCG (2x106 cfu, 7-8 749 mice/group). At 24h and 3 weeks post-infection, mice were sacrificed and the number of bacteria in the 750 spleens and lungs was determined by cfu plating (Mann-Whitney test; ** P<0.01). 751 Figure 7 | iDC recruitment to lymphoid organs starts earlier when mice are vaccinated with the BCG 752 sapM::Tn BCG strain 753 C57BL/6J mice were immunized s.c. at the base of the tail (2x106 cfu, 5-10 mice/group (A-B); 2.105 cfu, 10 754 mice/group (C-D) ) with WT BCG or sapM::Tn BCG mutant. Seven and 14 days later, the mice were sacrificed, 755 and inguinal LNs, brachial LNs and spleens were isolated. Cells were prepared, labelled with different 756 antibodies staining DCs and analyzed by flow cytometry. Total cell numbers (A, C) and the kinetics of iDC 757 (CD103+MHCIIhi CD11b+ Ly6C+) recruitment (B, D) differ upon BCG sapM::Tn BCG and WT BCG vaccination. 758 (Mann-Whitney; *P<0.05, **P<0.01). 759 Figure 8 | Vaccination with sapM::Tn BCG induces reduced frequencies of IFNγ-producing CD4+ and CD8+ T 760 cells compared to WT BCG 31 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 761 C57BL/6J mice were immunized s.c. at the base of the tail with a high dose of WT BCG (black bars) or sapM::Tn 762 BCG mutant (red bars) (2x106 cfu, 5 mice/group). 14 days later, the mice were sacrificed, and inguinal LNs, 763 brachial LNs and spleens were isolated. Cells were prepared and the T cell response was analyzed by 764 intracellular cytokine staining followed by flow cytometry (A). Additionally, supernatants of stimulated 765 spleen cells and controls were harvested 48hr post-stimulation and IFNγ concentration was determined by 766 bioplex (B). (Mann-Whitney or 2way ANOVA; *P<0.05, **P<0.01). 767 Table 1 | Variants detected in WT BCG and the sapM::Tn BCG disruption mutant. 768 A probabilistic variant analysis was performed in CLC genome workbench on the reads mappings of both the 769 WT BCG and sapM::Tn BCG mutant strain. We observed only very few variations compared to the M. bovis 770 BCG Pasteur str. 1173P2 reference genome. (SNV/MNV = single/multiple nucleotide variants; Indel = 771 insertions or deletions). The Tn insertion was picked up using de novo assembly of the sapM::Tn reads. 32 Fig. 1 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Fig. 2 A. bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. B. C. E. D. F. Fig. 3 bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Fig. 4 A. B. C. bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Fig. 5 50 kDa 37 kDa 25 kDa 20 kDa SapM::Tn WT C. SapM::Tn:empty vector bioRxiv preprint doi: https://doi.org/10.1101/486993. this versionB. posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. SapM::Tn:Compl A. D. Fig. 6 A. B. bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Fig. 7 A. C. B. bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. D. Fig. 8 A. B. bioRxiv preprint doi: https://doi.org/10.1101/486993. this version posted December 4, 2018. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Vaccination-site draining lymph node; 2 weeks post-vaccination Spleen; 2 weeks post-vaccination Note: in all figures: Variant 6a AA change K43R Synonymous L85fs A433InsA G261Deletion 3,854,971 InDel cut3 Insertion GR 3,685,759 Tn insertion SapM promoter 3,708,425 InDel sugI S13fs Variant 6b 3,708,517 Variant 1 Variant 2 Variant 3 Variant 4 Variant 5 Position 813,096 1,344,671 2,764,157 3,197,940 3,833,491 Mutation Type SNV MNV SNV* InDel InDel** SNV Gene rpsL BCG_2507c fadD26 BCG_3499c sugI L44M Illumina resequencing 1173P2 WT 1721 (reference) SapM ::Tn 1721 BCG WT A G G CG GC GC T A A A A CGC CGC 1173P2 WT 1721 SapM ::Tn A GC 1721 BCG WT G GC - A A G GC - TCG A TCG tn - TCG - TCG TCG A - C C A C C A * Not verified by Sanger sequencing, because the mutation is synonymous ** Sequencing did not work, due to the presence of repeats in the region of the mutation Table 1 Sanger analysis