YbdO Promotes the Pathogenicity of Escherichia coli K1 by Regulating Capsule Synthesis

Escherichia coli K1 is the most popular neonatal meningitis-causing Gram-negative bacterium. As a key virulence determinant, the K1 capsule enhances the survival of E. coli K1 in human brain microvascular endothelial cells (HBMECs) upon crossing the blood–brain barrier; however, the regulatory mechanisms of capsule synthesis during E. coli K1 invasion of HBMECs remain unclear. Here, we identified YbdO as a transcriptional regulator that promotes E. coli K1 invasion of HBMECs by directly activating K1 capsule gene expression to increase K1 capsule synthesis. We found that ybdO deletion significantly reduced HBMEC invasion by E. coli K1 and meningitis occurrence in mice. Additionally, electrophoretic mobility shift assay and chromatin immunoprecipitation–quantitative polymerase chain reaction analysis indicated that YbdO directly activates kpsMT and neuDBACES expression, which encode products involved in K1 capsule transport and synthesis by directly binding to the kpsM promoter. Furthermore, ybdO transcription was directly repressed by histone-like nucleoid structuring protein (H-NS), and we observed that acidic pH similar to that of early and late endosomes relieves this transcriptional repression. These findings demonstrated the regulatory mechanism of YbdO on K1 capsule synthesis, providing further insights into the evolution of E. coli K1 pathogenesis and host–pathogen interaction.


Introduction
Newborn bacterial meningitis is an acute inflammation of the meninges, subarachnoid space, and brain vasculature caused by bacteria or bacterial products and associated with substantial mortality and morbidity worldwide [1][2][3][4]. As defined by organism isolation from cerebrospinal fluid (CSF) cultures, Escherichia coli K1 is the most common Gramnegative bacterium that causes meningitis in newborns [5] and represents~80% of the CSF isolates identified in meningitic neonates [6]. The molecular mechanisms involved in the pathogenesis of E. coli K1 leading to this severe disease are not fully understood [7].
Most cases of E. coli-caused meningitis occur as a consequence of hematogenous spread [8], and circulating E. coli enters the brain parenchyma by penetrating the bloodbrain barrier (BBB), which is the most critical pathogenic step [9]. Several studies in humans and experimental animals suggest that high-degree bacteremia (e.g., >10 5 CFU/mL of neonatal animal blood [10]; >10 3 CFU/mL of neonatal infant blood [11]) is a prerequisite for meningeal invasion based on its enabling E. coli K1 escape from host defenses to cause meningitis. Penetration of the brain by E. coli K1 involves its binding to and invasion of human brain microvascular endothelial cells (HBMECs) that constitute the BBB. E. coli K1 binds to HBMECs by interacting with CD48 and sialoglycoproteins through type 1

E. coli K1 Invasion of HBMECs Increases ybdO Expression
To investigate the regulatory mechanisms associated with E. coli K1 invasion, we performed transcriptome analysis to reveal differences in the gene-expression profile between E. coli K1 cultured in brain-heart infusion (BHI) media alone and when invading HBMECs. We identified downregulation and upregulation of the expression of 1335 and 1359 genes, respectively, in HBMEC-invading E. coli K1 relative to BHI-cultivated E. coli K1 (Table S3). Classification of the differentially upregulated genes using Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed significant enrichment of the KEGG categories ABC transporters, two-component systems, and propanoate metabolism ( Figure S2). KEGG categories that were significantly enriched for the downregulated genes mainly included cofactor biosynthesis, amino sugar and nucleotide sugar metabolism, and oxidative phosphorylation ( Figure S2).
Interestingly, many known virulence genes in E. coli K1 were upregulated, including genes encoding type 1 and S fimbriae, which are associated with HBMEC binding, and cnf1, ibeA, aslA, and sitA, which are associated with HBMEC invasion ( Figure 1A) [37]. These results identified substantially upregulated expression of virulence factors upon E. coli K1 invasion of HBMECs, and that various transcription factors (TFs) likely mediate their expression.
1359 genes, respectively, in HBMEC-invading E. coli K1 relative to BHI-cultivated E. coli K1 (Table S3). Classification of the differentially upregulated genes using Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed significant enrichment of the KEGG categories ABC transporters, two-component systems, and propanoate metabolism ( Figure S2). KEGG categories that were significantly enriched for the downregulated genes mainly included cofactor biosynthesis, amino sugar and nucleotide sugar metabolism, and oxidative phosphorylation ( Figure S2).
Interestingly, many known virulence genes in E. coli K1 were upregulated, including genes encoding type 1 and S fimbriae, which are associated with HBMEC binding, and cnf1, ibeA, aslA, and sitA, which are associated with HBMEC invasion ( Figure 1A) [37]. These results identified substantially upregulated expression of virulence factors upon E. coli K1 invasion of HBMECs, and that various transcription factors (TFs) likely mediate their expression. Comparative transcriptome analysis revealed significant changes in the expression of twelve genes encoding putative transcriptional regulators possibly involved in virulence regulation ( Figure 1B). To determine the roles of these putative TFs in E. coli K1 invasion of HBMECs, we constructed isogenic in-frame-deletion mutants of potential TFs. For the invasion assay, we chose the ΔompA mutant strain and the non-invasion strain HB101 as a positive control for decreased invasion and negative control, respectively [38]. Among these TF mutants, the invasion rate of four (ΔyfeC, ΔyihW, ΔycjZ, and ΔybdO) To validate RNA-seq results, ybdO expression was analyzed by qRT-PCR. Data were obtained from three independent experiments and analyzed using Student's t-test. * p < 0.05, ** p < 0.01, *** p < 0.001.
Comparative transcriptome analysis revealed significant changes in the expression of twelve genes encoding putative transcriptional regulators possibly involved in virulence regulation ( Figure 1B). To determine the roles of these putative TFs in E. coli K1 invasion of HBMECs, we constructed isogenic in-frame-deletion mutants of potential TFs. For the invasion assay, we chose the ∆ompA mutant strain and the non-invasion strain HB101 as a positive control for decreased invasion and negative control, respectively [38]. Among these TF mutants, the invasion rate of four (∆yfeC, ∆yihW, ∆ycjZ, and ∆ybdO) significantly decreased, with that of the ∆ybdO mutant decreasing the most ( Figure 1C). Quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis revealed a 2.97-fold increase in ybdO expression during HBMEC invasion as compared with that observed in BHI cultures ( Figure 1D), which agreed with the RNA-seq results. These findings suggested that YbdO might play an important regulatory role in E. coli K1 invasion of HBMECs.

YbdO Contributes to E. coli K1 Invasion of HBMECs
To determine whether YbdO affects HBMEC invasion by E. coli K1, we constructed a complementary strain (cybdO) for virulence evaluation. We found that the invasion rate of the ∆ybdO mutant, as indicated by the number of bacteria entering HBMECs, was reduced by 2.90-fold as compared with that observed in the wild-type (WT) strain (Figure 2A), whereas the invasion rate of the complementary strain (cybdO) was comparable with that of the WT strain ( Figure 2A). Additionally, the growth rate of the ∆ybdO mutant in BHI broth was the same as that of the WT strain ( Figure 2C), demonstrating that the decreased invasion rate was not due to different growth rates. These results indicated that YbdO promotes the invasion of HBMECs by E. coli K1. Because binding occurs before invasion, we examined the ability of the ∆ybdO mutant strain to bind to HBMECs, finding that the binding rate was comparable with that of the WT strain ( Figure 2B), indicating that YbdO does not affect E. coli K1 binding to HBMECs. Consistent with the lack of effect on binding, ybdO expression did not change upon HBMEC binding as compared with that of the mutant grown in BHI medium ( Figure S3). , 23, x FOR PEER REVIEW 4 of 16 significantly decreased, with that of the ΔybdO mutant decreasing the most ( Figure 1C). Quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis revealed a 2.97-fold increase in ybdO expression during HBMEC invasion as compared with that observed in BHI cultures ( Figure 1D), which agreed with the RNA-seq results. These findings suggested that YbdO might play an important regulatory role in E. coli K1 invasion of HBMECs.

YbdO Contributes to E. coli K1 Invasion of HBMECs
To determine whether YbdO affects HBMEC invasion by E. coli K1, we constructed a complementary strain (cybdO) for virulence evaluation. We found that the invasion rate of the ΔybdO mutant, as indicated by the number of bacteria entering HBMECs, was reduced by 2.90-fold as compared with that observed in the wild-type (WT) strain ( Figure  2A), whereas the invasion rate of the complementary strain (cybdO) was comparable with that of the WT strain ( Figure 2A). Additionally, the growth rate of the ΔybdO mutant in BHI broth was the same as that of the WT strain ( Figure 2C), demonstrating that the decreased invasion rate was not due to different growth rates. These results indicated that YbdO promotes the invasion of HBMECs by E. coli K1. Because binding occurs before invasion, we examined the ability of the ΔybdO mutant strain to bind to HBMECs, finding that the binding rate was comparable with that of the WT strain ( Figure 2B), indicating that YbdO does not affect E. coli K1 binding to HBMECs. Consistent with the lack of effect on binding, ybdO expression did not change upon HBMEC binding as compared with that of the mutant grown in BHI medium ( Figure S3).

YbdO Promotes Meningitis in Mice
Because E. coli K1 invasion of HBMECs is associated with crossing the BBB, we investigated the contribution of YbdO to E. coli K1 pathogenesis by infecting mice with highlevel bacteremia and examining the CSF cultures for bacterial meningitis. CSF samples were collected and cultured to illustrate the onset of bacterial meningitis. Induction of meningitis by the ΔybdO strain resulted in a significantly reduced rate of meningitis occurrence (54.17%; n = 24) relative to that induced by the WT strain (84.62%, n = 26) (p = 0.0334) ( Figure 3A). By contrast, the rate of meningitis occurrence induced by the cybdO strain (76.92%, n = 26) was comparable to that induced by the WT strain ( Figure 3A), indicating that YbdO promotes meningitis in mice. Because different levels of bacteremia affect the percentage of meningitis occurrence, we further investigated whether ybdO deletion affected the level of bacteremia. The ΔybdO mutant and WT induced similar levels of bacteremia (WT, 7.89 ± 0.04 mean log CFU/mL of blood vs. ΔybdO, 7.54 ± 0.31 mean log CFU/mL of blood) ( Figure 3B), suggesting that YbdO has no effect on E. coli K1 replication

YbdO Promotes Meningitis in Mice
Because E. coli K1 invasion of HBMECs is associated with crossing the BBB, we investigated the contribution of YbdO to E. coli K1 pathogenesis by infecting mice with highlevel bacteremia and examining the CSF cultures for bacterial meningitis. CSF samples were collected and cultured to illustrate the onset of bacterial meningitis. Induction of meningitis by the ∆ybdO strain resulted in a significantly reduced rate of meningitis occurrence (54.17%; n = 24) relative to that induced by the WT strain (84.62%, n = 26) (p = 0.0334) ( Figure 3A). By contrast, the rate of meningitis occurrence induced by the cybdO strain (76.92%, n = 26) was comparable to that induced by the WT strain ( Figure 3A), indicating that YbdO promotes meningitis in mice. Because different levels of bacteremia affect the percentage of meningitis occurrence, we further investigated whether ybdO deletion affected the level of bacteremia. The ∆ybdO mutant and WT induced similar levels of bacteremia (WT, 7.89 ± 0.04 mean log CFU/mL of blood vs. ∆ybdO, 7.54 ± 0.31 mean log CFU/mL of blood) ( Figure 3B), suggesting that YbdO has no effect on E. coli K1 replication in mouse blood. Accordingly, we found that ybdO expression in mouse blood did not change significantly relative to that of the mutant grown in BHI medium ( Figure S3). Collectively, these results suggested that YbdO promotes BBB penetration by E. coli K1 to cause meningitis in mice.
Int. J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 5 in mouse blood. Accordingly, we found that ybdO expression in mouse blood did change significantly relative to that of the mutant grown in BHI medium ( Figure S3). lectively, these results suggested that YbdO promotes BBB penetration by E. coli K cause meningitis in mice.

YbdO Promotes HBMEC Invasion by E. coli K1 by Enhancing Capsule Production
To further investigate the mechanism of YbdO-mediated promotion of HBMEC vasion by E. coli K1, we performed RNA-seq analysis using the WT and ΔybdO strain determine the downstream target genes of YbdO. We identified upregulated and do regulated expression of 58 and 108 genes, respectively, in the ΔybdO mutant (Table KEGG analysis revealed that these differentially expressed genes (DEGs) have div functions ( Figure S4). Of particular relevance to E. coli K1 pathogenesis is the expres of several K1 capsule genes, including kpsT, which encodes ABC transporters (K1 cap region 3), and neuSEAB, which is involved in O antigen nucleotide sugar biosynthesis capsule region 2), was downregulated by ybdO deletion, whereas the expression of g of K1 capsule region 1 were not significantly affected ( Figure 4A). This suggested YbdO regulates the expression of genes of capsule regions 2 and 3 but not those of re 1. This result was further confirmed using qRT-PCR analysis ( Figure 4B), the resul which correlated well with RNA-seq data and confirmed the validity of the latter. identified no other known E. coli K1 virulence genes regulated by YbdO based on R seq analysis (Table S4). The positive effect of YbdO on the expression of K1 capsule g suggested that YbdO might promote E. coli K1 pathogenesis by increasing K1 capsule duction.

YbdO Promotes HBMEC Invasion by E. coli K1 by Enhancing Capsule Production
To further investigate the mechanism of YbdO-mediated promotion of HBMEC invasion by E. coli K1, we performed RNA-seq analysis using the WT and ∆ybdO strains to determine the downstream target genes of YbdO. We identified upregulated and downregulated expression of 58 and 108 genes, respectively, in the ∆ybdO mutant (Table S4). KEGG analysis revealed that these differentially expressed genes (DEGs) have diverse functions ( Figure S4). Of particular relevance to E. coli K1 pathogenesis is the expression of several K1 capsule genes, including kpsT, which encodes ABC transporters (K1 capsule region 3), and neuSEAB, which is involved in O antigen nucleotide sugar biosynthesis (K1 capsule region 2), was downregulated by ybdO deletion, whereas the expression of genes of K1 capsule region 1 were not significantly affected ( Figure 4A). This suggested that YbdO regulates the expression of genes of capsule regions 2 and 3 but not those of region 1. This result was further confirmed using qRT-PCR analysis ( Figure 4B), the results of which correlated well with RNA-seq data and confirmed the validity of the latter. We identified no other known E. coli K1 virulence genes regulated by YbdO based on RNA-seq analysis (Table S4). The positive effect of YbdO on the expression of K1 capsule genes suggested that YbdO might promote E. coli K1 pathogenesis by increasing K1 capsule production.
3, x FOR PEER REVIEW 6 of 16 Deletion of ybdO resulted in decreased expression of genes of regions 2 and 3, resulting in a decreased E. coli K1-invasion rate. To further verify whether YbdO contributes to E. coli K1 invasion and virulence by activating K1 capsule gene expression, we constructed a ybdO/neuDB double mutant strain (ΔybdOΔneuDB) and a complementary strain (ΔybdOΔneuDB strain complemented with ybdO, ΔybdOΔneuDB+P-ybdO) and performed HBMEC-invasion assays. The rate of HBMEC invasion by ΔybdOΔneuDB (31.95%) was lower than that by ΔybdO (53.33%, p < 0.05), indicating that deletion of neuDB in ΔybdO further decreased the invasion ability of the strain ( Figure 4C). Additionally, the rate of Deletion of ybdO resulted in decreased expression of genes of regions 2 and 3, resulting in a decreased E. coli K1-invasion rate. To further verify whether YbdO contributes to E. coli K1 invasion and virulence by activating K1 capsule gene expression, we constructed a ybdO/neuDB double mutant strain (∆ybdO∆neuDB) and a complementary strain (∆ybdO∆neuDB strain complemented with ybdO, ∆ybdO∆neuDB+P-ybdO) and performed HBMEC-invasion assays. The rate of HBMEC invasion by ∆ybdO∆neuDB (31.95%) was lower than that by ∆ybdO (53.33%, p < 0.05), indicating that deletion of neuDB in ∆ybdO further decreased the invasion ability of the strain ( Figure 4C). Additionally, the rate of HBMEC invasion by ∆ybdO∆neuDB+P-ydbO (35.55%) was comparable with that by ∆ybdO∆neuDB ( Figure 4C), indicating that the K1 capsule plays an important role in the process of E. coli K1 invasion of HBMECs. These results confirmed that YbdO enhances E. coli K1 invasion by activating K1 capsule gene expression.
To investigate whether YbdO directly enhances capsule production, we performed immunofluorescence microscopy of the K1 capsule using anti-sialic acid, with capsule production quantified based on the fluorescence intensity of the fluorescein isothiocyanate (FITC)-labeled K1 capsule [39]. The fluorescence intensity of the FITC-labeled K1 capsule produced by ∆ybdO cultured in BHI medium did not differ significantly from that of the WT and ybdO-overexpressing strains ( Figure S5), indicating that YbdO did not increase capsule production in culture medium. To further investigate the effect of YbdO on K1 capsule production during HBMEC invasion by E. coli K1, we stained the K1 capsule following E. coli K1 invasion of HBMECs with an immunofluorescent. Subsequent immunofluorescence microscopy revealed a 1.51-fold reduction in the fluorescence intensity of the capsule produced by the ∆ybdO mutant strain following HBMEC invasion as compared with that of the mutant before invading the HBMECs. By contrast, the fluorescence intensity of the capsule produced by the WT and cybdO strains increased 1.74-and 1.81-fold, respectively, after HBMEC invasion as compared with that of the strains before invading the HBMECS ( Figure 4D), indicating that YbdO is required for E. coli K1 production of K1 capsules in host cells.
To evaluate whether YbdO directly or indirectly regulates the expression of regions 2 and 3 genes, we performed electrophoretic mobility shift assays (EMSAs) using purified 6×-His-tagged YbdO and the potential promoter region of region 3 (300-bp upstream of the initial gene kpsM). The results showed that YbdO specifically binds to the kpsM promoter ( Figure 5A), whereas no binding between YbdO and the lacZ fragment (used as a negative control) was detected under the same conditions ( Figure 5A). Additionally, chromatin immunoprecipitation-quantitative PCR (ChIP-qPCR) analysis showed that the kpsM promoters were exceedingly enriched in the YbdO-ChIP samples, with their relative quantity in these samples 241-fold higher than that in the mock-ChIP control samples ( Figure 2B). By contrast, the fold enrichment of lacZ (negative control) did not significantly differ between the YbdO-ChIP and mock-ChIP samples ( Figure 5B). These results demonstrated that YbdO directly binds to the promoter region of the kpsM gene both in vitro and in vivo, suggesting that YbdO directly activates the expression of regions 2 and 3 genes by directly binding to the kpsM promoter. To investigate whether YbdO directly enhances capsule production, we performed immunofluorescence microscopy of the K1 capsule using anti-sialic acid, with capsule production quantified based on the fluorescence intensity of the fluorescein isothiocyanate (FITC)-labeled K1 capsule [39]. The fluorescence intensity of the FITC-labeled K1 capsule produced by ΔybdO cultured in BHI medium did not differ significantly from that of the WT and ybdO-overexpressing strains ( Figure S5), indicating that YbdO did not increase capsule production in culture medium. To further investigate the effect of YbdO on K1 capsule production during HBMEC invasion by E. coli K1, we stained the K1 capsule following E. coli K1 invasion of HBMECs with an immunofluorescent. Subsequent immunofluorescence microscopy revealed a 1.51-fold reduction in the fluorescence intensity of the capsule produced by the ΔybdO mutant strain following HBMEC invasion as compared with that of the mutant before invading the HBMECs. By contrast, the fluorescence intensity of the capsule produced by the WT and cybdO strains increased 1.74-and 1.81fold, respectively, after HBMEC invasion as compared with that of the strains before invading the HBMECS ( Figure 4D), indicating that YbdO is required for E. coli K1 production of K1 capsules in host cells.
To evaluate whether YbdO directly or indirectly regulates the expression of regions 2 and 3 genes, we performed electrophoretic mobility shift assays (EMSAs) using purified 6×-His-tagged YbdO and the potential promoter region of region 3 (300-bp upstream of the initial gene kpsM). The results showed that YbdO specifically binds to the kpsM promoter ( Figure 5A), whereas no binding between YbdO and the lacZ fragment (used as a negative control) was detected under the same conditions ( Figure 5A). Additionally, chromatin immunoprecipitation-quantitative PCR (ChIP-qPCR) analysis showed that the kpsM promoters were exceedingly enriched in the YbdO-ChIP samples, with their relative quantity in these samples 241-fold higher than that in the mock-ChIP control samples ( Figure 2B). By contrast, the fold enrichment of lacZ (negative control) did not significantly differ between the YbdO-ChIP and mock-ChIP samples ( Figure 5B). These results demonstrated that YbdO directly binds to the promoter region of the kpsM gene both in vitro and in vivo, suggesting that YbdO directly activates the expression of regions 2 and 3 genes by directly binding to the kpsM promoter.

H-NS Represses ybdO Gene Expression by Directly Binding to the ybdO Promoter
ybdO and other horizontally-transferred genes are under the control of the global regulator H-NS in E. coli K12 [36]. Therefore, we compared the homology of the ybdO promoter between E. coli K12 and K1 and found similarities in their high-AT ratio regions ( Figure S6). To investigate whether H-NS regulates ybdO expression in E. coli K1, we constructed an Δhns mutant strain and performed qRT-PCR to determine changes in ybdO

H-NS Represses ybdO Gene Expression by Directly Binding to the ybdO Promoter
ybdO and other horizontally-transferred genes are under the control of the global regulator H-NS in E. coli K12 [36]. Therefore, we compared the homology of the ybdO promoter between E. coli K12 and K1 and found similarities in their high-AT ratio regions ( Figure S6). To investigate whether H-NS regulates ybdO expression in E. coli K1, we constructed an ∆hns mutant strain and performed qRT-PCR to determine changes in ybdO expression. Deletion of hns in E. coli K1 resulted in 2.2-and 2.4-fold increases in ybdO expression in strains grown in BHI medium and after HBMEC invasion, respectively, relative to that observed in the WT strain ( Figure 6A). These results indicated that H-NS was able to repress ybdO expression in E. coli K1. To evaluate whether H-NS directly represses ybdO expression, we performed EMSAs using purified 6×-His-tagged H-NS and the promoter region of ybdO (corresponding to 300 bp upstream). The results showed that H-NS binds specifically to the promoter region of ybdO in vitro ( Figure 6B) but did not bind the lacZ fragment (negative control) [40] under the same experimental conditions ( Figure 6B). These results indicated that H-NS represses ybdO gene expression by directly binding to the ybdO promoter.

23, x FOR PEER REVIEW 8 of 16
expression in strains grown in BHI medium and after HBMEC invasion, respectively, relative to that observed in the WT strain ( Figure 6A). These results indicated that H-NS was able to repress ybdO expression in E. coli K1. To evaluate whether H-NS directly represses ybdO expression, we performed EMSAs using purified 6×-His-tagged H-NS and the promoter region of ybdO (corresponding to 300 bp upstream). The results showed that H-NS binds specifically to the promoter region of ybdO in vitro ( Figure 6B) but did not bind the lacZ fragment (negative control) [40] under the same experimental conditions ( Figure 6B). These results indicated that H-NS represses ybdO gene expression by directly binding to the ybdO promoter.

Acidic pH Is a Host Cue to Induce ybdO Expression by Reducing H-NS Repression
We then investigated possible host signals that contribute to the induction of ybdO expression. H-NS release from targeted DNA-binding regions is reportedly induced by an acidic pH [41,42], which is a typical feature of early and late endosomes (pH ~ 6.5 and ~5.5, respectively) [43]. Notably, E. coli K1 traverses HBMECs through transcytosis via early and late endosomes [8]. To determine whether an acidic pH affects ybdO and kpsM expression, we simulated normal blood [44] and early and late endosomes using M9 medium at pH 7.4, 6.5, and 5.5. qRT-PCR results revealed 1.8-and 2.3-fold increases in ybdO expression upon exposure of the WT strain to pH 6.5 and pH 5.5, respectively, as compared with expression observed at pH 7.4 ( Figure 7A). These results indicated that an acidic pH promotes ybdO expression. Furthermore, ybdO expression was not induced by an acidic pH in the Δhns mutant strains ( Figure 7B), indicating that an acidic pH induces ybdO expression via H-NS release from the ybdO promoter region. PCR products were added to the binding buffer at 20 ng each, and H-NS protein was added to the reaction buffer (lanes 1-5) at 0, 0.125, 0.25, 0.5, and 1 µM. Data were obtained from three independent experiments and analyzed using Student's t-test. *** p < 0.001.

Acidic pH Is a Host Cue to Induce ybdO Expression by Reducing H-NS Repression
We then investigated possible host signals that contribute to the induction of ybdO expression. H-NS release from targeted DNA-binding regions is reportedly induced by an acidic pH [41,42], which is a typical feature of early and late endosomes (pH~6.5 and~5.5, respectively) [43]. Notably, E. coli K1 traverses HBMECs through transcytosis via early and late endosomes [8]. To determine whether an acidic pH affects ybdO and kpsM expression, we simulated normal blood [44] and early and late endosomes using M9 medium at pH 7.4, 6.5, and 5.5. qRT-PCR results revealed 1.8-and 2.3-fold increases in ybdO expression upon exposure of the WT strain to pH 6.5 and pH 5.5, respectively, as compared with expression observed at pH 7.4 ( Figure 7A). These results indicated that an acidic pH promotes ybdO expression. Furthermore, ybdO expression was not induced by an acidic pH in the ∆hns mutant strains (Figure 7B), indicating that an acidic pH induces ybdO expression via H-NS release from the ybdO promoter region. dium at pH 7.4, 6.5, and 5.5. qRT-PCR results revealed 1.8-and 2.3-fold increases in ybdO expression upon exposure of the WT strain to pH 6.5 and pH 5.5, respectively, as com pared with expression observed at pH 7.4 ( Figure 7A). These results indicated that an acidic pH promotes ybdO expression. Furthermore, ybdO expression was not induced by an acidic pH in the Δhns mutant strains ( Figure 7B), indicating that an acidic pH induces ybdO expression via H-NS release from the ybdO promoter region.  Similar to ybdO expression, we observed 2.14-and 2.27-fold increases in kpsM transcript levels in the WT strain at pH 6.5 and pH 5.5, respectively, relative to those at pH 7.4 ( Figure 7C). This suggested that E. coli K1 might increase capsule production to resist acidic conditions. Notably, activation of kpsM expression by an acidic pH was abolished by ybdO deletion (Figure 7D), indicating that an acidic pH promotes kpsM expression through YbdO. Collectively, these results suggested that acidic conditions in endosomes are a host signal sensed by E. coli K1 to reverse the H-NS-mediated repression of ybdO expression and thereby enhance capsule gene expression.

Discussion
Pathogens often exploit cunning strategies to adapt to the niche of the host. E. coli K1 is a critical common pathogen responsible for high morbidity and mortality in infants [1]; however, the mechanisms involved in regulating E. coli K1 virulence remain poorly understood. Here, we demonstrated that the transcriptional regulator YbdO promoted E. coli K1 invasion of HBMECs and meningitis in mice by directly activating the expression of K1 capsule region 2 and 3 genes by binding to the kpsM promoter. Additionally, ybdO expression was directly repressed by H-NS, with this inhibitory effect abolished under acidic conditions similar to those found in endosomes. Therefore, these findings identified a relationship between YbdO and both acidic pH and K1 capsule regulation.
The ability of E. coli K1 to biosynthesize and assemble capsular polysaccharides is conferred by the K1 capsule genes [29,30]. Previous studies show that upregulation of capsular genes is often accompanied by increased capsular production [34,45,46]. In this study, we found that acidic pH induced capsule gene expression in the E. coli K1 WT strain ( Figure 7C). It is reasonable to assume that acidic pH would increase the production of K1 capsular polysaccharides, although this requires further experimental confirmation. Therefore, we proposed a model of the YbdO-dependent K1 capsule regulatory pathway ( Figure 8). Briefly, when E. coli K1 enters endosomes, the acidic pH relieves the H-NSmediated repression of ybdO expression, after which activated ybdO expression promotes upregulation of K1 capsule gene expression, thereby increasing K1 capsule production to counteract the unfavorable environment in endosomes. capsular polysaccharides, although this requires further experimental confirmation. Therefore, we proposed a model of the YbdO-dependent K1 capsule regulatory pathway ( Figure 8). Briefly, when E. coli K1 enters endosomes, the acidic pH relieves the H-NSmediated repression of ybdO expression, after which activated ybdO expression promotes upregulation of K1 capsule gene expression, thereby increasing K1 capsule production to counteract the unfavorable environment in endosomes. In the blood, the pH is neutral (7.4), and H-NS binds to the ybdO promoter and represses ybdO expression, thereby inhibiting activation of capsule gene expression. In the endosomes of HBMECs, the pH is acidic (5.5-6.5), which induces H-NS detachment from the promoter to activate ybdO expression, allowing subsequent activation of capsule gene expression by YbdO to promote capsule production. In the blood, the pH is neutral (7.4), and H-NS binds to the ybdO promoter and represses ybdO expression, thereby inhibiting activation of capsule gene expression. In the endosomes of HBMECs, the pH is acidic (5.5-6.5), which induces H-NS detachment from the promoter to activate ybdO expression, allowing subsequent activation of capsule gene expression by YbdO to promote capsule production.
Previous transcriptome analysis of E. coli K1 bound to HBMECs detected a total of 227 genes that were differentially expressed in E. coli K1 associated with HBMEC as compared with expression observed in non-associated bacteria in the supernatant [47]. These genes are mainly involved in the presentation of cell-surface molecules, cellular function, and nitrogen metabolism. We constructed the transcriptome of E. coli K1 following its invasion of HBMECs and detected a total of 2700 DEGs relative to expression observed in bacteria cultured in BHI medium. These results demonstrated that E. coli K1 invasion of HBMECs is a complex process involving the expression of numerous genes, thereby offering insight into the mechanisms associated with E. coli K1 interactions with relevant target tissues. Complete elucidation of these transcriptional changes will provide additional information concerning E. coli K1-HBMEC interactions that are critical to understanding the pathogenesis of E. coli meningitis.
Our RNA-seq results showed that 166 regulator genes were differentially expressed (58 upregulated and 108 downregulated) in the ∆ybdO mutant as compared with the WT strain (Table S4), indicating YbdO as a global transcriptional regulator. In addition to K1 capsule genes, we found that ybdO deletion resulted in the downregulated expression of many known E. coli virulence genes, including ompF (encoding outer membrane porin F), dnaJ (encoding chaperone protein DnaJ), and ibpB (encoding the small heat-shock protein IbpB) (Table S4). Therefore, YbdO may exert its regulatory functions on E. coli K1 virulence and pathogenesis by affecting the expression of these virulence genes; however, confirmation of this will require further experimental studies.
RNA-seq analysis of genes exhibiting low expression can lead to relatively large fold biases and inaccurate results [48]. To confirm the RNA-seq results, we performed qRT-PCR analysis to validate the changes in expression of 14 capsule genes. For most of the genes examined, the fold change detected by real-time PCR was higher than that detected by RNA-seq. This was understandable, given that RNA-seq is generally less sensitive than qRT-PCR for quantifying gene expression [49,50]. Additionally, we found that some RNAseq results were inconsistent with qRT-PCR results, especially those involving changes in neuD expression ( Figure 4A,B). This could be because transcription levels of neuD in both WT and the ∆ybdO mutant were extremely low based on read counts (12 in WT and 15 in ∆ybdO) from RNA-seq results.
The typical expression of capsules in E. coli is regulated by temperature [34], whereas genes involved in capsule polysaccharide biosynthesis are upregulated in response to low oxygen and low iron levels [31]. We found that E. coli regulated the expression of the K1 capsule in response to an acidic environment in HBMECs. Acidic pH causes hydrolysis of the K1 capsule for capsule sloughing. Bacteria that have already produced polysialic acid (PSA) during growth under different conditions release their preformed PSA when placed in an acidic environment, such as in a phagolysosome [51][52][53]. Specifically, the Kl capsule produced in cultures at pH 7.0 was optimally released at pH 5.0. Therefore, it appears that K1 capsule production is increased in E. coli K1 to resist the acidic environment of endosomes and escape to fuse with lysosomes for survival and crossing of the BBB; however, the precise mechanisms involved in this process remain unclear and merit further investigation.
The K1 capsule increases the recovery of viable intracellular E. coli but attenuates binding to and internalization in HBMECs [8]. Therefore, the capsule is a double-edged sword for pathogens, and the spatial and temporal regulation of capsule production is of great significance for virulence. A previous study reported that bacterial adhesins shielded by the capsule also affect the interaction of Neisseria meningitidis with epithelial and endothelial cells, and that downregulation of capsule gene expression and removal of the sialic acid from the capsule are necessary for meningococcal interactions with host cells [54]. During urinary tract infection, uropathogenic E. coli downregulates the expression of capsule genes in urine for optimal adhesion to epithelial cells and switches from low or no capsule expression to increased capsule expression upon invasion of epithelial cells to allow the formation of intracellular bacterial communities [55]. In the present study, we found that ybdO deletion significantly decreased the capsule production and survival of intracellular E. coli K1 but had no obvious effect on E. coli K1 cultured in medium. We speculated that E. coli K1 expresses normal or low K1 capsule extracellularly to efficiently bind and invade HBMECs, followed by activation of the expression of K1 capsule genes by YbdO intracellularly to resist acidic challenges.
In conclusion, these findings enhance our understanding of how E. coli K1 utilizes environmental cues to facilitate HBMECs invasion and provide a paradigm for environmental signal sensing and virulence regulation that can be used to study other human bacterial pathogens.

Bacterial Strains, Plasmids, and Growth Conditions
The bacterial strains and plasmids used in this study are listed in Supplementary Table  S1. The oligonucleotide primers used in this study are listed in Supplementary Table S2. E. coli K1 RS218 was used as the WT strain. Mutant strains were generated using the λ Red recombinase system of pSim6 and primers carrying the 50 bp homologous regions flanking the start and stop codons of the gene to be deleted, as previously described [56]. Bacteria were generally grown in BHI media at 37 • C; however, strains containing the temperature-sensitive plasmid pSim6 were cultured at 30 • C. For qRT-PCR analysis, bacteria were grown overnight in BHI media and inoculated at 1:100 into fresh M9 medium at different pH values (5.5-7.4; 37 • C; 200 rpm) until reaching the stationary phase [41]. The working concentrations of the antibiotics ampicillin, kanamycin, and chloramphenicol were 100 µg/mL, 50 µg/mL, and 25 µg/mL, respectively. All bacterial strains were frozen at −80 • C using 20% (v/v) glycerol.

E. coli Binding and Invasion Assays with HBMECs
HBMECs were a generous gift from Dr. K. S. Kim (Johns Hopkins University, Baltimore, MD, USA) and cultured in Roswell Park Memorial Institute (RPMI)-1640 medium with 10% fetal bovine serum (FBS), 10% Nu-serum, 2 mM glutamine, 1% MEM nonessential amino acids, 1 × MEM vitamin, 100 U/mL penicillin, 100 µg/mL streptomycin, and 1 mM sodium pyruvate. The E. coli strain was grown to the exponential phase at an optical density of 600 nm (OD 600 ) of 0.6, collected via centrifugation, and resuspended in RPMI-1640 medium containing 10% FBS. HBMECs infected with a multiplicity of infection (MOI) of 100 were incubated at 37 • C in a 5% CO 2 incubator for 90 min. The monolayers were then washed with warm phosphate-buffered saline (PBS) and incubated with experimental medium containing gentamicin (100 mg/mL) for 1 h at 37 • C to kill the extracellular E. coli. HBMECs were washed, lysed using 0.5% Triton X-100 in PBS, and cultured for determination of the CFUs. A binding assay was performed similar to the invasion assay, except with the omission of the gentamicin-treatment step [13].

Animal Model of E. coli Bacteremia and Hematogenous Meningitis
E. coli bacteremia and hematogenous meningitis were induced in BALB/c mice, which were~14-days old (Vital River Laboratory Animal Technology Co., Beijing, China), as described previously [57]. All experiments were conducted according to protocols approved by the Institutional Animal Care Committee at Nankai University (Tianjin, China). Each mouse received E. coli (1 × 10 6 CFU) in the exponential phase in 100 µL of PBS via tail vein injection. After 4 h, blood and CSF specimens were collected for determination of the CFUs and for RNA extraction. Meningitis was defined as a positive culture in CSF [58]. For determination of the CFUs, the bacteria in the blood specimens were subjected to serial 10-fold dilutions in PBS and enumerated by plating on BHI agar plates. For RNA extraction, the mice were sacrificed, and blood specimens were collected to extract RNA using TRIzol reagent (Invitrogen, Carlsbad, CA, USA).

qRT-PCR
qRT-PCR was performed using an ABI QuantStudio 5 real-time PCR system (Applied Biosystems, Foster City, CA, USA). The E. coli strains were cultured overnight and subsequently subcultured (1:100) in fresh BHI medium to the exponential phase. Bacteria were pelleted via centrifugation, and RNA samples were isolated using TRIzol (Invitrogen), reverse transcribed using a PrimeScript RT reagent kit (Takara, Shiga, Japan), and processed for qRT-PCR. Each qRT-PCR was performed using Power SYBR Green PCR master mix (Applied Biosystems). The fold change in the expression of the target gene relative to that of the housekeeping gene (dnaE) was determined using the 2 −∆∆Ct method [59]. At least three biological replicates were performed for each qRT-PCR analysis.

RNA-seq
E. coli K1 was collected in the exponential phase in BHI or during HBMEC invasion, and total RNA was isolated using TRIzol reagent (Invitrogen) according to manufacturer instructions. RNA was quantified and qualified using a Bioanalyzer 2100 (Agilent Technologies, Santa Clara, CA, USA), a NanoPhotometer spectrophotometer (Implen GmbH, Munich, Germany), and 1% agarose gel electrophoresis. For library preparation, we used 3 µg of total RNA per sample. rRNA was depleted from total RNA using a Ribo-off rRNA depletion kit (Vazyme, Nanjing, China), and libraries were constructed and analyzed by NOVO-GENE, Inc. (TianJin, China). DEGs in HBMEC-invading E. coli K1 were identified using the DESeq R package. Available online: https://www.nuget.org/packages/PuppeteerSharp (accessed on 13 April 2022). Their expression was compared with that of BHI-cultivated E. coli K1. The resulting p-values were adjusted using the Benjamini-Hochberg test for controlling the false discovery rate. Genes with an adjusted p < 0.05 were considered as differentially expressed. The other transcriptomes of the WT and ∆ybdO mutant strains were processed and compared using the same methods.

ChIP-qPCR Analyses
For the FLAG-tagged plasmids, the coding DNA sequence was amplified from the RS218 genome using PCR and cloned into the pBAD24 plasmid to allow ybdO overexpression. The WT strain containing pBAD-YbdO was cultured in LB medium supplemented with ampicillin and 0.1% arabinose until the mid-log phase (OD 600 = 0.6) and then treated with 1% formaldehyde for 10 min at 25 • C. Cross-linking was stopped by the addition of 125 mM glycine. Bacterial pellets were washed twice with PBS buffer, resuspended in immunoprecipitation buffer [IP, 50 mM, (pH 7.5), HEPES-KOH, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, and protease inhibitor cocktail (Medchem Express LLC, Monmouth Junction, NJ, USA)], and then subjected to sonication to produce 250 to 500 bp DNA fragments. Insoluble cellular debris was removed via centrifugation at 4 • C, and the supernatant was used as the input sample in the immunoprecipitation experiments. Both the mock and immunoprecipitated samples were separately incubated with isotype and anti-FLAG antibodies and then incubated with protein A beads in an immunoprecipitation buffer. Washing, cross-link reversal, and purification of the ChIP DNA were then conducted. To measure the enrichment of the potential YbdO-binding targets in the immunoprecipitated DNA samples, the percent input and fold enrichment were determined using SYBR green PCR master mix. The relative target levels were calculated using the ∆Ct method, with lacZ used as a negative control according to a previous study [60]. The results were reported as the average enrichment of three biological replicates.

Immunofluorescence Assays
Immunofluorescence analysis was performed as described previously [39]. Briefly, bacteria were subcultured at a 1:100 ratio in BHI and incubated overnight at 37 • C with shaking at 180 rpm until an OD 600 of 0.6 was obtained. The bacteria were then diluted on coverslips to allow HBMEC infection at an MOI of 100 in the exponential phase. After 2.5 h of incubation, the coverslips were washed and fixed with formaldehyde, and the cells were permeabilized with 0.1% Triton X and stained with fluorescein AF647-labeled phalloidin to visualize the actin filaments. E. coli was stained with the FITC-labeled anti-E. coli K and O antigen antibody (Abcam, Cambridge, UK). The K1 capsule was stained with an anti-PSA antibody and AF488-labeled secondary antibody (Abcam, Cambridge, UK). HBMEC nuclei were stained with 6-diamidino-2-phenylindol. Invasion assays were performed for each cell line, with three slides per experiment.

EMSAs
The 6×-His-tagged H-NS (N-terminus) and YbdO (C-terminus) proteins were expressed in E. coli BL21/DE3 containing pET-H-NS and pET-YbdO plasmids, respectively, and purified from soluble extracts using a Ni-NTA-chelating column (Thermo Scientific, Waltham, MA, USA), as previously described [41] Protein concentrations were determined using a bicinchoninic acid protein assay, and the proteins were aliquoted and stored at −80 • C. The PCR fragments containing the promoter regions of kpsM and ybdO and the lacZ fragment (negative control) [40] were amplified using the genomic DNA of the RS218 strain as a template. The fragments were then gel-purified, and 20 ng of the DNA fragments was incubated with purified protein (0-2 µM) in 20 µL of a solution containing band-shift buffer [10 mM Tris (pH 7.5), 1 mM EDTA, 100 mM KCl, 0.1 mM DTT, 5% (v/v) glycerol, and 0.01 mg/mL bovine serum albumin] [61] at 25 • C for 30 min. Native 8% (w/v) polyacrylamide gels were used to separate the samples in 0.5 × Tris-borate-EDTA, and the gels were then stained with GelRed (Genestar, Beijing, China).

Growth Assay
To determine the growth curve of each strain, overnight cultures were washed with PBS three times and diluted (1:1000) in BHI broth without antibiotics. A 200 µL aliquot was added to a 96-well flat-bottom microplate and incubated at 37 • C with shaking at 180 rpm for 24 h, as previously described [62]. The absorbance at 600 nm was recorded. Experiments were independently performed three times.

Statistical Analysis
Statistical analysis was conducted using GraphPad Prism software (v8.3.0; GraphPad Software, San Diego, CA, USA). The mean ± SD from three independent experiments is shown in the figures. Differences between two mean values were evaluated using a two-tailed Student's t-test. Statistical significance was assessed using the two-sided Fisher's exact test for the animal meningitis experiments. Statistical significance was set at a p < 0.05.

Institutional Review Board Statement:
All animal experiments in this study were conducted according to protocols approved by the Institutional Animal Care Committee at Nankai University (Tianjin, China; protocol code 2021-SYDWLL-000028). Optical density at 600 nm