All manipulations of wildtype Bp were carried out in CDC approved and registered BSL-3 facilities at the University of Hawaiʻi at Mānoa (UHM) John A. Burns School of Medicine (JABSOM) and experiments with the select agent were performed in accordance with the recommended BSL-3 practices (Richmond and McKinney, 2007). Select-agent excluded Bp 1026b ∆asd strain and Bp82 strain (Propst et al., 2010) were manipulated in the BSL-2 laboratory at the UHM. The derivative of Escherichia coli EPmax10B (Bio-Rad) containing lacI^q and pir [Lab ID: E1869 (Norris et al., 2009)] were routinely used as a cloning strain. The E. coli conjugal and suicidal strain HPS1-pir-Δasd-mob-Km-Δpro (Lab ID: E463 (Kang et al., 2009)) was used for the mobilization of plasmid containing oriT-origin into B. pseudomallei 1026b ∆asd [B0011 (Norris et al., 2011)] strain.

Growth of E. coli Δasd strains was carried out as previously described with 100 μg/ml of diaminopimelic acid (DAP; Sigma) supplied (Norris et al., 2009). Luria-Bertani (LB) medium (Difco) was generally used to culture all the bacteria. Wildtype Bp strain 1026b was cultured in LB while the B0011 strain (Norris et al., 2009) was cultured in LB with 200 μg/ml diaminopimelic acid (LB + DAP200) or in minimal glucose (MG) media (1 × M9 minimal medium plus 20 mM glucose) containing 1 mM lysine, 1 mM methionine, 1 mM threonine, and 200 μg/ml DAP (3AA + DAP200). Diaminopimelic acid was prepared in 1 M NaOH as a 100 mg/ml stock and was used when necessary as described previously (Barrett et al., 2008). Bp82 (Propst et al., 2010) is a ΔpurM derivative of Bp 1026b, and is an adenine and thiamine auxotroph. To culture Bp82, LB was supplemented with 0.6 mM adenine (LB + Ade) and M9-glucose medium was supplemented with both 0.6 mM adenine and 0.0005% thiamine (Propst et al., 2010). Merodiploids arising from chromosomal integration of the suicidal plasmid are resolved by utilization of the counter selectable marker pheS and sacB. 0.2% (w/v) p-chlorophenylalanine (cPhe; dl-4-chlorophenylalanine; Acros Organics) and 15% sucrose were used in Bp, respectively. LB + 1,000 μg/ml kanamycin was used for purifying the 1026b::T24 insertion mutants. All bacterial growth was carried out at 37°C and shaken cultures were maintained at 225 rpm.

Murine macrophage cell RAW264.7 was cultured in a humidified CO₂ incubator at 37°C with 5% CO_2. Dulbecco’s modified Eagle’s medium (DMEM) containing 4,500 mg/l glucose with 4 mM L-glutamine was used for culturing cell lines. All culture media were supplemented with 10% (v/v) heat-inactivated fetal bovine serum (FBS; HyClone). Gibco 100× antibiotic/antimycotic (AntiAnti) was added at a 1 × working concentration (containing 100 U/ml of penicillin, 100 μg/ml of streptomycin, and 250 ng/ml of amphotericin B) to prevent contamination by various bacteria and fungi. RAW264.7 cells were maintained at 60–80% confluency and were passaged by scrapping cells from the Corning^™ T-75 flask. The passage numbers of cell lines were carefully recorded. CellBIND multi-well plates were used for cell infection assays.

Oligonucleotides were synthesized through Integrated DNA Technology. Unless otherwise indicated, all restriction enzymes, deoxynucleoside triphosphates, T4 DNA polymerase, T4 polynucleotide kinase, and T4 DNA ligase were purchased from New England Biolabs and used according to the manufacturer’s suggestion. Plasmids and DNA gel bands were isolated using the Zyppy plasmid miniprep kit and Zymoclean gel DNA recovery kit, respectively (Zymo Research Corporation). Escherichia coli competent cells were prepared as previously described (Kang et al., 2007). DNA ligations, restriction endonuclease digestions, and agarose gel electrophoresis were performed according to standard techniques, and all other molecular techniques were carried out according to methods described previously by Sambrook and Russell (2001).

Prokaryotes promoter predictor from Berkeley Drosophila Genome Project (Reese, 2001)¹ was used to predict sap1 transcription start site. The transcript start sites of Sap1 was then experimentally mapped using a previously described method by direct sequencing of SMART^™ RACE (rapid amplification of cDNA end) products (Tabansky and Nurminsky, 2003). All reagents required for promoter mapping were supplied in the SMART RACE cDNA Amplification kit (Clontech, BD Biosciences). Briefly, Bp wildtype 1026b strain was cultured in 3 ml LB broth and harvested in exponential phase for RNA extraction. Total RNA was prepared using the TRIzol^® reagent (Invitrogen, Carlsbad, CA, United States), and isolated with an on-column DNase I treatment, using the Qiagen RNeasy plus mini kit. One μg of total RNA was used to generate RACE cDNA using the SMART RACE cDNA Amplification kit (Clontech, BD Biosciences) and following the manufacturer’s instructions. cDNA ends were amplified with SMART primer (5′-AAGCAGTGGTATCAACGCAGAGTACGCGGG-3′) and gene-specific primer (5′-GTCTTTCGACAGACATCC-3′) at a final concentration of 2 μM. The diluted cDNA was used as a template in a PCR with a second nested gene-specific primer (5′-CAGTTGAATCTCCTGCTGC-3′) and the SMART RACE primer (5′-AAGCAGTGGTATCAACGCAGAGT-3′). PCR products were purified from 2% agarose gel slices and sequenced with a third nested gene-specific primer (5′-AGCGATGCATCGCCGTTC-3′) at the University of Hawaii genomic core facility.

The engineering of P_sap1-lacZ reporter strain was performed as described below. About 200 bp from up and down stream flanking the mapped transcriptional start sit was amplified from Bp genomic DNA with primers 5′-GAATGCCGGGTACCGAATGCGCCTG-3′ and 5′-ATGTCGTGTCTCGAATTCGCTCGGGTC-3′. A 400 bp PCR product was digested with EcoRI and KpnI and cloned into mini-Tn7-Km-lacZ vector digested with the same restriction enzymes. A reporter strain was constructed with a single-copy integration of a P_sap1-lacZ fusion at a neutral glmS site in the chromosome of B0011 strain using the well-established mini-Tn7 system (Koch et al., 2001; Choi et al., 2005; Kang et al., 2009) with the aid of a nonreplicating helper plasmid pTNS3 encoding the transposase(Kang et al., 2009). Briefly, plasmids mini-Tn7-Km-P_sap1-lacZ and pTNS3 was isolated using Zyppy Plasmid Kit. These two plasmids were mixed at an equal molar ratio and transformed into B0011 strain via electroporation. To make electro-competent bacterial cells, B0011 was grown to exponential phase in LB + DAP200 broth and 1 ml bacterial culture was harvested by centrifugation and washed four times with sterile water to remove salts. After electroporation, B0011 cells were immediately resuspended in 2 ml fresh LB + DAP200 and recovered in a 37°C incubator shaking at 225 RPM for 2 h. Cells were plated on LB + DAP200 + Km500 to select for the successful integration of P_sap1-lacZ fusion. The B. pseudomallei chromosomes contain three insertion sites downstream of three glmS genes (Choi et al., 2008). P_sap1-lacZ fusion was PCR confirmed to insert once at the glmS2 site with primers 5′-ATTAGCTTACGACGCTACACCC-3′ and 5′-ACACGACGCAAGAGCGGAATC-3′. The insertion of mini-Tn7 in bacterial chromosome is quite stable and the selective maintenance is not required (Norris et al., 2010). The reporter strain was purified three times before saving away in the -80°C freezer stocks due to the intrinsic stickiness of Bp cells.

Random mutagenesis was performed in the B0011 P_sap1-lacZ fusion strain using the pBT20-gat plasmid containing the Himar1 mariner transposon and its transposase-encoding tnp gene (Kulasekara et al., 2005). Escherichia coli strain E463 (HPS1-pir-Δasd-mob-Km-Δpro) was used as the conjugal donor to introduce transposon plasmid pBT20-gat into B0011 P_sap1-lacZ fusion strain. Briefly, the donor and recipient were grown up to mid-exponential phase and the cells were harvested separately by centrifugation at 9000 × g for 1 min at room temperature and washed twice with 1 ml of fresh LB. The pellets were combined at approximately equal cell densities and gently resuspended in 50 ul of LB. The mixture was spotted onto a LB + DAP200 plate for conjugation at 37°C. After 4 h of conjugation, the bacteria were harvested and washed twice with 1 ×M9 buffer, then plated on 10×150 mm agar plates containing MG + DAP 200 μg/ml + 3AA (1 mM of methionine, lysine, and threonine) + 60 μg/ml X-gal + 0.4% glyphosate. Without adding amino acid proline, the donor strain is unable to grow. The recipient strain cannot grow with the presence of glyphosate unless it has pBT20-gat transposon inserted into the chromosomes. The plates were incubated at room temperature for at least 24 h to allow the color to fully develop after colonies were visible following incubation at 37^o C (Supplementary Figure S1). The white and dark blue colonies were picked and patched on the same agar to confirm β-galactosidase activity. A total of 230,000 transposon mutant colonies, approximately 40 × coverage of genome, were screened for the β-galactosidase activity. 205 colonies with higher (dark blue) and 220 colonies with lower (white) β-galactosidase activity were pooled separately for transposon junctions identification through Illumina^® deep-sequencing (Tn-seq).

Genomic DNA (gDNA) of dark blue and white colony pools was purified with phenol-chloroform extraction and quantified via Nanodrop. Shearing buffer (40% glycerol + 10 mM Tris pH 8.0 + 1 mM EDTA) was used to resuspend 10 μg of gDNA sample to a total volume of 800 μl followed by nebulization. Briefly, the samples were each nebulized for 5 min at 35 p.s.i. of N₂ and followed by adding 30 mg/ml of dextran sulfate dissolved in water to each sample to a final concentration of 0.2 μg/ml. One tenth the volume of 3 M sodium acetate was added and vortexed to mix well. To precipitate DNA, an equal volume of ice cold isopropanol was added to the sample and freeze at –80^o C for at least 30 min. The samples were centrifuged at 4°C at full speed for 20 min and washed with old 70% ethanol two times. Excessive ethanol was further removed by heated speed vacuum and DNA pellets were resuspended in 16 μl Qiagen EB buffer for Nanodrop quantification. The sheared DNA was end repaired with the NEB Next End Repair Kit according to the manufacturer’s protocol and purified with Qiagen QIAquick PCR Purification Kit following the protocol. Each sample was further divided into four samples and 3′ A-tailing was carried out with Taq polymerase at 70°C for 20 min. The adapters were designed following previously described method with modification for sequence specific to our transposon (Gallagher et al., 2011). Purified DNA samples were digested with XbaI overnight at 37°C to separate the two genome-transposon junctions. Two % TAE agarose gel was used to visualize the DNA samples and the DNA smears between 200 and 400 bp were excised and purified via QIAquick Gel Extraction Kit and eluted with 30 μl of EB buffer. Circularization of adaptor ligated gDNA and exonuclease digestion were performed as previously described using Ampligase heat stable ligase (Gallagher et al., 2011). The transposon ends were amplified with PCR and Tn-seq libraries were built with the NEBNext DNA library prep kit for Ilumina^® sequencing based on manufacturer’s protocols. Library samples were cleaned up with MinElute PCR purification Kit and eluted with EB buffer.

The sample was sequenced at the Genetics Core and Confocal Microscopy Facility at University of Hawaiʻi at Mānoa. The dark blue and the white colony sample pools received 3.2 and 4.5 million reads, respectively. Reads were mapped onto the Bp genome with Bowtie (Langmead et al., 2009) using the Galaxy suite (Blankenberg et al., 2001; Giardine et al., 2005; Goecks et al., 2010) to identify the transposon junctions.

The well-established mini-Tn7 integration system was used to construct the 1026b sapR::T24 complement strain (Norris et al., 2010) in BSL-3 laboratory. Briefly, sapR and its predicted native promoter was cloned into mini-Tn7-gat plasmid, yielding mini-Tn7-gat-sapR. The helper plasmid pTNS3 encoding transposase was electroporated with mini-Tn7-gat-sapR into the 1026b sapR::T24 mutant. After recovering in LB for 2 h, cells were harvested by centrifugation in a bio-safety cabinet and washed twice with 1x M9 minimal salts buffer to remove trace nutrient from LB. Mutants contains the mini-Tn7 insertion were selected for on MG + 0.4% GS plate. Colonies were purified on the same medium three times and insertion at one of the three glmS sites in the chromosome was verified by PCR as previously described (Kang et al., 2009). This extensive purification procedure is very important and highly recommended, because GS is bacteriostatic, rather than bactericidal, at this effective concentration. 1026b sapR::T24 mutant, sapR::T24 complement, and the wildtype strain were used in the RT-PCR validation study and RNA-seq analysis.

We used our previously published primers and probe for sap1 in this real-time PCR experiment (Heacock-Kang et al., 2021). The primers and probe were designed using Integrated DNA Technologies Primer Quest software² following the criteria (Son et al., 2007) that were mentioned below. Briefly, the amplicon size should range from 60 to 100 bp and the primer melting temperatures were designed for 62°C, when more one primer pair were designed a melting temperature difference of less than 4°C for each pair would be recommended. The probe melting temperatures were designed to be 5–10°C higher than those for the corresponding primer pairs. To eliminate the possibility of nonspecific binding, primer and probe sequences were also subjected to BLASTN analysis against the B. pseudomallei wildtype 1026b genome, and the primers were also ensured to had no significant complementarity at the 3′ end and the probe to had no significant complementarity at the 5′ end to nonspecific locations on Bp 1026b genome.

The same preparation of RNA as described above for high throughput RNA-Seq experiment was used for real-time RT-PCR. Three μg of RNA was used for cDNA synthesis with random hexamer and SuperScript III (Invitrogen) per manufacturer’s protocol. To ensure no genomic DNA contamination was carried over from the RNA extraction, a control lacking reverse transcriptase was included for cDNA synthesis. The final cDNA product was dissolved in 1,200 μl RNase and DNase free water. In order to yield more accurate normalization analysis (Vandesompele et al., 2002) and control for variations between runs, three housekeeping genes (Heacock-Kang et al., 2021) (BPSL0602, BPSL2502, and BPSS2061 express consistently across all conditions) were chosen to be amplified at the same time with sap1 on one 96-well plate. The primers set to amplify sap1 are forward_5′-CATCGTTGCATTTTCGG-3′, reverse_ 5′-GGAATCCAGATTGTCTTTCG-3′, and probe_/56-FAM/TTCGTCAGTTGAATCTCCTGC/3BHQ_1/. Briefly, master mixtures were made and aliquoted into sub mixtures for each gene assayed. The final volume of each real-time PCR mixture is 25 μl containing 10 μl of cDNA, 12.5 μl of iQ Supermix (Bio- Rad), 120 nM of each primer, and 12 nM probe. Real-time PCR was performed with an iCycler iQ (Bio-Rad) using the following protocol: denaturation at 95°C for 10 min and then 55 cycles of amplification and quantification at 95°C for 20 s and at 65°C for 45 s. For each gene, reactions with amplification efficiency variant within 5% are subjected to data analysis. Real-time PCR fold changes were calculated using the amplification plot method and the available macro functions for data analyses of real-time PCR (DART-PCR; Peirson et al., 2003).

Deletion mutation of sapR gene in Bp82 was made by allelic-replacement method based on the concept of the double homologous recombination as previously described (Barrett et al., 2008) with minor modification. Bp82 was grown in LB + Ade to exponential phase and harvested by centrifugation. Electro-competent Bp82 cells were made as described above (Barrett et al., 2008). Briefly, cells were washed four times with sterile ddH₂O for electroporation with the allelic-replacement and suicidal vector pEX18Km-pheS-ΔsapR-gat-FRT2. After 2–3 h recovery in fresh LB + Ade, the bacterial cells were plated onto MG + 0.4% GS Ade/Thi agar plate. Merodiploids arising from chromosomal integration of the suicide plasmid are resolved by utilization of the counter selectable marker pheS. Colonies were purified on MG + 0.4% GS + 0.2% cPhe + Ade/Thi agar plates to resolve the merodiploid state. Plates were incubated at 37°C for 48 h, colonies were patched with toothpicks onto both MG + 0.4% GS + 0.2% cPhe + Ade/Thi and LB + Ade + Km 500 μg/ml plates. Mutants unable to grow in the presence of kanamycin were purified on MG + 0.4% GS Ade/Thi. Single colonies were purified on MG + 0.4% GS Ade/Thi media. The same purification procedure was repeated three times ensuring no wildtype background carrying over. Further screening and confirmation of mutants were performed by PCR using oligonucleotides pairs 5′-TGTCGATGCCGTCGAAG-3′/5′-CACATCCTGATGTACAAGCGGATCA-3′ (upstream junction) and 5′-GGCATCGAGAGTGTCATAC-3′/5′-CGTTGATGGGCTTGACCTCGATC-3′ (downstream junction) to amplify the chromosome and vector junctions. Primers 5′-CTGTCGATGCCGTCGAAG-3′ and 5′-GGCATCGAGAGTGTCATAC-3′ annealed to the upstream region of the chromosome outside of the region cloned for allelic replacement and the downstream region, respectively.

An unmarked strain was required to be constructed before engineering the complement strain. The gat-FRT2 marker excision was performed using a previously described protocol (Choi et al., 2008) with minor modifications. Briefly, ΔsapR mutant was electroporated with 200 ng of pFlpe4, a curable Flp recombinase-expressing plasmid. Hundred μl aliquot of 1 ml recovery culture was plated on LB + Ade + Km500 + 0.2% l-rhamnose plate and incubated at 30°C for 48 h or until colonies were grown up. Single colonies were picked and patched on the same media to incubate at 30°C to ensure full excision. Fully-grown patches were streaked out on LB + Ade plate for single colony isolation and pFlpe4 plasmid was cured by incubating at 42°C. The successful excision of gat marker was confirmed by PCR and no growth on MG + 0.4% GS Ade/Thi plate. The well-established mini-Tn7 integration system was used to construct the Bp82 ΔsapR complement strain (Norris et al., 2010) in BSL-2 laboratory as described above. Bp82 ΔsapR mutant and ΔsapR/ comp were used in the mammalian cell infection assay.

Bp82, Bp82 Δsap1, Bp82 ΔsapR, and Bp82 ΔsapR complement strains were grown up overnight in LB + Ade broth in a shaking incubator at 37°C. The bacteria were harvested and diluted 100 times in 250 μl fresh LB + Ade and placed in a 96-well plate. This experiment was conducted in triplicate. Growth analysis was carried out in the BioTek EL × 808IU plate reader at 37°C with constant shaking and OD₆₀₀ measurement was recorded every 30 min. Data was analyzed and plotted in Prism.

pBLAC is a derivative of pMLBAD, a low-copy broad-host-range plasmid (Lefebre and Valvano, 2002) that can replicate in Bp (Barrett et al., 2008). The sapR gene is placed immediately downstream of a lac operator and the expression of sapR can be induced by IPTG. pBLAC was constructed by replacing the E. coli araBAD arabinose-inducible promoter and araC, transcriptional regulator gene, of pMLBAD (Lefebre and Valvano, 2002) with lac promoter and lacI-encoded Lac repressor of mini-Tn7-lacZ that has been routinely used in our lab for Bp wildtype 1026b strain gene induction experiments. pBLAC was designed to be used for sapR induction study in Bp82 strains and its usage can be extended to Burkholderia thailandensis (Bt), the experimental surrogate of Bp that is manipulated in a BSL-1 laboratory. New vector pBLAC carries the backbone sequence derived from broad-host-range vector pMLBAD, including mob, gene required for conjugal transfer of plasmid; rep, replication protein gene; ori, origin of replication; the dhfr, dihydrofolate reductase gene encoding trimethoprim resistance; rrnB, transcriptional terminator; MCS, multiple cloning site. SapR cloned into the MCS was transcribed from the lac promoter and regulated by Lac repressor. Trimethoprim (Tp) 150 μg/ml was used in media to maintain the plasmid in Bp82 strains.

RAW264.7 macrophage cells were cultured in the T-75 corning flask to confluence and washed with pre-warmed sterile 1 x PBS twice. Then, cells were scraped off from the flask and seeded in 250 μl of DMEM without AntiAnti at 7 × 10⁴ cells per well into 48-well Corning CellBIND culture plates. Cells were allowed to attach overnight before the infection study. Bp82 strain grown in LB medium + Ade overnight at 37°C was used to infect RAW264.7 macrophages at MOI of 10:1 in the 48-well CellBIND plates. Bacteria were diluted in 1 × PBS for plating and numeration to obtain the initial infecting bacteria number. At 1 h post infection, DMEM containing unattached bacteria were removed and the cell monolayers were washed three times with pre-warmed sterile 1 × PBS. Monolayers were lysed with 500 μl 1 × PBS + 0.2% Triton X-100, and serial diluted in 1 × PBS. Fifty μl of 10⁰ (undiluted lysate), 10⁻¹, 10⁻², 10⁻³ dilutions were plated onto LB + Ade 12-well agar plates and colonies were enumerated after 48 h incubation at 37°C as the attached bacteria number. The attachment efficiency was determined by dividing the attached number by the initial infecting bacteria number. The experiment was carried out in triplicate and the numbers represent the average of all three replicates with the error bars representing the SEM. The student t-test was used to determine the statistical significance between attachment efficiencies of the different strains.

Intracellular replication assay was carried out as previously described (Norris et al., 2011) with slight modifications. Briefly, RAW264.7 murine macrophage cells were infected with Bp82, mutant and complement strains at an MOI of 10:1 in an aminoglycoside protection assay. RAW264.7 cells were cultured and seeded into a 48-well Corning CellBIND culture plates as described above. After allowing the infection to progress for 1 h, the medium containing bacteria was removed, and the monolayers were washed three times with 1 × PBS to remove any unattached bacteria. Fresh DMEM + Ade with 1,000 μg/ml each of amikacin and kanamycin was added to the monolayers to kill any non-internalized bacteria and inhibit extracellular bacterial growth for the remainder of the assay. During the assay, medium was removed from the wells at 2, 8, and 24 hpi (hours post infection), and the infected cell monolayers were washed, lysed and diluted as described above. Bacteria colonies were numerated after 48 h incubation at 37°C, and data was calculated and plotted in Prism. All the assays were performed in triplicate, and the SEM were calculated for each.

Burkholderia pseudomallei wildtype strain1026b and sapR::T24 insertion mutant, were grown overnight in 3 ml LB broth, subcultured to grow to mid exponential phase for RNA extraction. Total RNA from 12 ml culture of each strain was isolated using TRIzol^® reagent and purify through QIAGEN RNeasy plus mini kit according to the manufacturer’s instructions. Sequencing was performed by the Tufts University Core Facility Genomics on the Illumina HiSeq 2500 platform after rRNA depletion and library generation. More than 20 million reads with > 96% alignment rate to Bp genome were generated for each triplicate sample from each strain. The fastq files were uploaded to Rockhopper tool (McClure et al., 2013) designed specifically for analysis of bacterial RNA-Seq data. The normalized expression values generated in the Rockhopper analysis, in which the sapR mutant was compared to the wild type control, were used for the differential gene expression analysis. Genes with a false discovery rate (q-value) of < 0.01 were considered to be significantly differentially expressed and were designated for further analysis. A subset of these significantly differentially expressed genes with log₂FC ≤ −1 and ≥ 1 were used for Clusters of Orthologous Groups of proteins (COG) functional analysis and subcellular localization analysis for their gene products. The subcellular localization was computationally predicted through PSORTb V3.0 (Yu et al., 2010). The list of predicted gene across the whole genome was downloaded from Burkholderia.com to use in the analysis. Graphs were generated in Prism. Differentially regulated gene clusters (q-value < 0.01) identified from RNA-Seq were mapped along two Bp chromosomes using WoPPER web server (Puccio et al., 2017) to visualize their physical localizations.

To find regulators of sap1 using high-throughput transposon mutagenesis and deep sequencing, we must first determine the approximate sap1 regulatory region by identifying its transcriptional start site. Using the neural network based promoter prediction server of the Berkeley Drosophila Genome Project (BDGP; Reese, 2001), we found two possible transcriptional start sites. The first possible transcriptional start site is 182 bp before the sap1 start codon and has a confidence value of 0.81. The second possible transcriptional start site is 281 bp before the sap1 start codon and has a confidence value of 1.00. Transcription of sap1 could occur from one or both of these possible transcriptional start sites. To experimentally determine if one or both of these transcriptional start sites is used, we employed the well-established non-radioactive SMART RACE method (Tabansky and Nurminsky, 2003). As shown in Figure 1A, one SMART product approximately 500 bp in size was obtained. Sanger sequencing determined that the sap1 transcriptional start site is 281 bp upstream of its start codon, confirming the high confidence prediction from BDGP (Figures 1B,C). From this information, we determined the promoter of sap1 (P_sap1) and identified the −10 and − 35 elements (Figure 1C). The sap1 promoter region encompassing 200 bps flanking either side of the predicted sap1 transcriptional start site was utilized to identify potential regulators of sap1 (Figure 1C).

We designed a high-throughput strategy to identify the regulator(s) for sap1 using transposon mutagenesis of the select-agent exempt strain Bp 1026b Δasd (B0011) and deep sequencing (Figure 2; Gallagher et al., 2011; Norris et al., 2011). A reporter strain was constructed with a single-copy integration of a P_sap1-lacZ fusion at a neutral site on the chromosome of B0011 using the mini-Tn7 system (Koch et al., 2001; Choi et al., 2005; Kang et al., 2009). Random mutagenesis was then performed on this B0011 reporter fusion strain, utilizing the mariner transposon on pBT20-gat (Kulasekara et al., 2005). To ensure comprehensive coverage of both chromosomes, a total of 230,000 transposon mutant colonies, approximately 40 × coverage of Bp genome, were screened for variations in β-galactosidase activity. Colonies with increased β-galactosidase activity would have dark blue color indicating that the transposon insertions disrupted genes that repress expression from the sap1 promoter (Figure 2B). On the other hand, if a gene that activates expression from the sap1 promoter was mutated, lower β-galactosidase activity would be observed, and colonies would be less blue than the basal level (Figure 2B). Several hundreds of both dark blue and white colonies were patched to ensure a consistent β-galactosidase activity pattern. Eventually, 205 dark blue colonies and 220 white colonies were pooled separately and the transposon junctions were identified by Tn-seq⁴⁹. From the pool of white colonies and through this analysis, we identified BP1026B_II1046, a hypothetical protein with a DNA binding motif. BP1026B_III046 shows similarities to MurR/RipR DNA binding transcriptional regulators (COG1737) with an N-terminal helix-turn-helix like HTH_6 domain (Pfam: PF01418) and a C-terminal a SIS (Sugar ISomerase) domain (cd05013), which is found in many phosphosugar isomerases and phosphosugar binding proteins (Sørensen and Hove-Jensen, 1996; Finn et al., 2016; Lu et al., 2020). In addition, both Phyre2 (Kelley et al., 2015) and I-TASSER (Zhang, 2009; Roy et al., 2012; Yang and Zhang, 2015) predicted that BP1026B_II1046 was homologous to Vibrio vulnificus NanR further indicating that BP1026B_II1046 is a regulator. We, therefore, designated BP1026B_II1046 as sapR for the remainder of this study.

We used qRT-PCR to verify SapR regulatory activity on sap1 (Table 1). RNA was isolated from Bp 1026b wildtype strain and a sapR mutant strain from the Bp 1026b T24 transposon mutant library (Borlee et al., 2017). In the absence of SapR, the sap1 gene was down-regulated with a log₂FC (fold change) of −0.78 compared to wildtype Bp 1026b, consistent with the lower β-galactosidase activity in our initial transposon mutagenesis screening (Table 1). We performed beta-gal assay (Son et al., 2008) in the P_sap1-lacZ reporter strains of Bp82 and Bp82 ΔsapR mutant to further quantify the regulatory activity of SapR. The beta-galactosidase activity was significantly reduced in the ΔsapR mutant, and the reduction level agreed with our RNA-seq and RT-PCR results (Supplementary Figure S2). Additionally, from our previous Bp single cell ‘TRANSITomic’ analysis that revealed dynamic gene-expression flux as Bp transit through RAW264.7 macrophage at defined stages: vacuole entry; cytoplasmic escape and replication; and membrane protrusion (Heacock-Kang et al., 2021), we identified that both sap1 and sapR were up-regulated in the endocytic vesicle. This finding suggests a coordinated expression of sap1 and sapR during intracellular infection (Supplementary Figure S4). Together these data indicate that SapR is an activator of the sap1 gene.

To further validate the SapR regulatory function of sap1, attachment assays that measure the efficiency of bacterial adherence to the host cell surface were carried out, using the select-agent excluded Bp82 strain derived from wildtype strain 1026b. RAW264.7 murine macrophages were infected with strains Bp82, Bp82 ΔsapR, Bp82 ΔsapR-attTn7::sapR complement (Bp82 ΔsapR/comp), and Bp82 Δsap1 at an MOI of 10:1. The ΔsapR mutant was significantly defective in attachment to RAW264.7 monolayers, showing 48.16% of Bp82 attachment ability (Figure 3A). This is comparable to the attachment defect of the Δsap1 mutant exhibiting 46.72% wildtype attachment (Figure 3A). This result shows that the ΔsapR mutant has a similar phenotype to the Δsap1 mutant during host cell infection and suggests that SapR and Sap1 act in a similar pathway. The attachment defect of the ΔsapR mutant was restored by single copy complementation of the sapR gene.

In addition, we over expressed SapR in trans to compare the attachment efficiencies with Bp82 and Bp82 ΔsapR mutant. RAW264.7 murine macrophages were infected with Bp82/pBLAC, Bp82 ΔsapR/pBLAC, and the over-expressed sapR strain Bp82 ΔsapR/pBLAC-sapR all under IPTG induction (Figure 3B). The strain ΔsapR /pBLAC-sapR that over expresses SapR demonstrated a significantly increased ability to attachment to RAW264.7 cells, at approximately 140% of Bp82 (Figure 3B). This observation that increased expression of sapR leads to increased attachment provides further evidence that SapR is an activator of sap1. Taken together, these data suggest that both SapR and Sap1 have roles in the host cell attachment process.

To investigate if the attachment deficiencies of the ΔsapR mutant lead to subsequent decreased in intracellular replication, an intracellular replication assay was carried out in RAW264.7 macrophage cells. RAW264.7 cells were infected with an MOI of 10:1 with Bp82, Bp82 ΔsapR, Bp82 ΔsapR-attTn7::sapR complement (Bp82 ΔsapR/comp), and Bp82 Δsap1 strains (Figure 4A). At 2 h post infection (hpi), no significant difference was observed initially amongst all the strains indicating a similar number of bacteria were uptake by macrophage cells. As the infection progressed, the intracellular replication capability of Bp82 ΔsapR was significantly impaired, replicating at 45% of Bp82 at 8 hpi and 48% of Bp82 at 24 hpi (Figure 4B). The defect of the ΔsapR mutant was restored by single copy complementation of the sapR gene (Figures 4A,B). These results are similar in the Bp82 Δsap1 mutant showing defects at 8 h and 24 hpi, replicating at 35 and 18 percent of Bp82 (Figure 4B). In vitro growth analysis showed that these mutants have similar growth patterns to wildtype and complement strains (Figure 4C). Taken together, the results indicate that the defect in intracellular replication of the Bp82 ΔsapR mutant is due to impaired pathogenesis rather than defect in in vitro fitness.

We have shown that SapR controls the expression of sap1 and that there is a functional correlation between these two proteins. To further understand the role of SapR in the context of Bp pathophysiology, we sought to determine the regulation network of SapR. We compared the transcriptome profile from Bp 1026b wildtype and 1026b sapR::T24 mutant strains growing under the same conditions using RNA-Seq analysis. The sap1 gene was differentially expressed with a log₂FC of −0.56 and q-value (the false discovery rate) of 9.56E-93, further confirming that SapR is a positive regulator of sap1 (Table 1). To gain a better understanding of the SapR regulatory network, we did a further investigation on the role of SapR regulating membrane associate components which was summarized in Table 2. This table listed interesting findings of differentially expressed membrane associate components that are either known to associate with Bp pathogenesis or newly identified membrane-associated clusters from WoPPER analysis in our study.

In total, 483 genes were regulated by SapR using a statistical cut off of q-value < 0.01 and a log₂FC cut off of ≤ −1 or ≥ 1 (Supplementary Table S1; Figure 5A). Among these 483 differentially expressed genes (DEGs), there are 67 predicted RNA transcripts, and 416 annotated open reading frames. Of the 416 protein coding genes, 56% are assigned a subcellular location and 44% are unknown (Figure 5B). Some of these SapR-regulated proteins are associated with the cytoplasmic membrane (14%), periplasmic space (3%), outer membrane (3%) and extracellular space (2%; Figure 5B). Cytoplasmic membrane associated proteins showed a large enrichment to 24% when increasing the stringency to a log₂FC cut off of ≤ −2 or ≥ 2 (Figure 5C). BP1026B_I1015, BP1026B_I1020 and BP1026B_II1220, BP1026B_II1225 associated with cell envelope for nitrogen metabolism were regulated by SapR with the mean of log₂FC > 1 and < −1 (Table 2). SapR also regulated known virulence factors associated with pathogenesis. SapR regulates components of type III and type VI secretion systems (T3SS3 and T6SS). BP1026B_II0111 (tssM-2) and BP1026B_II2265 (tssM-6) encoding an inner membrane protein TssM essential for the function of a sub-complex of the T6SS. Both genes were significantly up-regulated in the sapR mutant with log₂FC of 1 (Table 2).

BP1026B_II1619 (BopA), an effector of T3SS3 responsible for phagosome escape (Gong et al., 2011), and two translocators of T3SS3, BP1026B_II1628 (BipB) involved in stimulating cell-to-cell fusion and BP1026B_II1627 (BipC) bound to host F-actin to facilitate internalization of Bp into host cells (Kang et al., 2016; Vander Broek et al., 2017), were down-regulated in the sapR mutant. In addition, SapR regulates many outer membrane porins (BP1026B_II0620, BP1026B_II0774, BP1026B_II1369, BP1026B_I0046 and BP1026B_I3218) generally responsible for cellular permeability and antibiotic resistance (Choi and Lee, 2019). Furthermore, an autotransporter BP1026B_I1153 (batA; Lafontaine et al., 2019) was down-regulated in the sapR mutant. We also observed altered expression of genes near the sapR gene, namely BP1026B_II1040–BP1026B_II1045, amongst the highest down-regulated genes in the SapR mutant (Table 2; Supplementary Figure S3). These include a dipeptidase, dipeptide and ABC transport systems, and a periplasmic binding protein that could be involved in peptidoglycan metabolism. The sapR mutant was complemented by inserting a copy of sapR back into the genome via the mini-Tn7 system. This demonstrates that the transposon mutation of sapR is unlikely to cause polar effects on the adjacent genes. This evidence strongly suggests that SapR regulates many proteins associated with the cell envelope that are also tied to Bp pathogenesis.

Among the 483 DEGs, 249 were up regulated and 234 were down regulated (Figure 6; Supplementary Table S1). Approximately 50% of DEGs are classified as hypothetical proteins, proteins of unknown function, proteins with no COG prediction, or hypothetical RNA transcripts (Figure 6). The rest of DEGs encode core functions such as lipid, inorganic ion, carbohydrate, or amino acid transport and metabolism, energy production and conversion, secondary metabolites biosynthesis, and cell wall/membrane/envelope biogenesis (Figure 6). Among those functional categories, lipid transport and metabolism and energy production and conversion account for the largest proportions of the analyzed DEGs (Figure 6). This suggests that SapR controls many pathways of unknown function and several known pathways including lipid metabolism and energy utilization.

To understand what genomic regions are activated and repressed by SapR, we analyzed the RNA-seq data set using WoPPER (Puccio et al., 2017). WoPPER is a web tool based on Position RElated genomics Data Analysis (PREDA) with the integration of gene expression data and the physical position of genes along the genome to identify differentially expressed chromosomal regions in bacteria (Puccio et al., 2017). WoPPER analysis identified 72 gene clusters up-regulated and 73 gene clusters down-regulated in the sapR mutant strain (Supplementary Figure S5). This analysis highlights the broad reach that SapR has on gene expression throughout the entire genome and indicates that SapR has the ability to activate and repress transcription. This analysis also identifies two nitrogen metabolism gene clusters that are differently regulated by SapR (Supplementary Figure S5; Table 2; Supplementary Table S2). One of these gene clusters (BP1026B_I1015–BP1026B_I1020) is located on chromosome I (Cluster ID: 21) and up-regulated, while the second (BP1026B_II1220, BP1026B_II1222-BP1026B_II1225) is located on chromosome II (Cluster ID: 30) and down-regulated in the sapR mutant. These nitrate reduction gene clusters share 69–79% sequence homology (Mangalea et al., 2017) but have different expression patterns controlled by SapR indicating possible distinct roles for each. This evidence suggests that Bp may exploit different nitrogen metabolism operons to survive in diverse environments.