Human Milk Oligosaccharide Utilization in Intestinal Bifidobacteria Is Governed by Global Transcriptional Regulator NagR

ABSTRACT

Bifidobacterium longum subsp. infantis is a prevalent beneficial bacterium that colonizes the human neonatal gut and is uniquely adapted to efficiently use human milk oligosaccharides (HMOs) as a carbon and energy source. Multiple studies have focused on characterizing the elements of HMO utilization machinery in B. longum subsp. infantis; however, the regulatory mechanisms governing the expression of these catabolic pathways remain poorly understood. A bioinformatic regulon reconstruction approach used in this study implicated NagR, a transcription factor from the ROK family, as a negative global regulator of gene clusters encoding lacto-N-biose/galacto-N-biose (LNB/GNB), lacto-N-tetraose (LNT), and lacto-N-neotetraose (LNnT) utilization pathways in B. longum subsp. infantis. This conjecture was corroborated by transcriptome profiling upon nagR genetic inactivation and experimental assessment of binding of recombinant NagR to predicted DNA operators. The latter approach also implicated N-acetylglucosamine (GlcNAc), a universal intermediate of LNT and LNnT catabolism, and its phosphorylated derivatives as plausible NagR transcriptional effectors. Reconstruction of NagR regulons in various Bifidobacterium lineages revealed multiple potential regulon expansion events, suggesting evolution from a local regulator of GlcNAc catabolism in ancestral bifidobacteria to a global regulator controlling the utilization of mixtures of GlcNAc-containing host glycans in B. longum subsp. infantis and Bifidobacterium bifidum.

IMPORTANCE The predominance of bifidobacteria in the gut of breastfed infants is attributed to the ability of these bacteria to metabolize human milk oligosaccharides (HMOs). Thus, individual HMOs such as lacto-N-tetraose (LNT) and lacto-N-neotetraose (LNnT) are considered promising prebiotics that would stimulate the growth of bifidobacteria and confer multiple health benefits to preterm and malnourished children suffering from impaired (stunted) gut microbiota development. However, the rational selection of HMO-based prebiotics is hampered by the incomplete knowledge of regulatory mechanisms governing HMO utilization in target bifidobacteria. This study describes NagR-mediated transcriptional regulation of LNT and LNnT utilization in Bifidobacterium longum subsp. infantis. The elucidated regulatory network appears optimally adapted to simultaneous utilization of multiple HMOs, providing a rationale to add HMO mixtures (rather than individual components) to infant formulas. The study also provides insights into the evolutionary trajectories of complex regulatory networks controlling carbohydrate metabolism in bifidobacteria.

INTRODUCTION

Bifidobacteria are Gram-positive, anaerobic, saccharolytic microorganisms that colonize the digestive tracts of humans and various animals (1). Certain Bifidobacterium species, namely, Bifidobacterium longum subsp. infantis, Bifidobacterium longum subsp. longum, Bifidobacterium bifidum, and Bifidobacterium breve often predominate the human neonatal gut microbiota (GM) during breastfeeding (2–5), and their predominance is directly linked with the healthy development of the infant (6, 7). The health-promoting effects attributed to the infant-associated bifidobacteria include protection from enteropathogen colonization (8–10) and modulation of the immune system (11, 12).

Decreased Bifidobacterium abundance is characteristic of immature GMs observed in preterm infants (13) and children suffering from severe acute malnutrition (14, 15). Therapeutic approaches aimed at restoring bifidobacterial population in these affected groups include administering exogenous Bifidobacterium species (e.g., B. longum subsp. infantis) as probiotics and/or food formulas containing prebiotics that would selectively stimulate the growth of autochthonous bifidobacteria in the gut and thus confer beneficial properties to the infant (13, 15–17). Since the prevalence of bifidobacteria in the neonatal gut is often attributed to their ability to selectively utilize dietary human milk oligosaccharides (HMOs) (18–20), these milk glycans are considered “natural” prebiotics and added to infant formulas (21).

HMOs are the third most abundant (5 to 20 g/L) component of human milk after lactose (Lac) and lipids and are not assimilated by the infant (22). HMO building blocks include glucose (Glc), galactose (Gal), N-acetylglucosamine (GlcNAc), l-fucose (Fuc), and N-acetylneuraminic acid (Neu5Ac); these units form more than 100 linear or branched oligosaccharide species (23, 24). Most HMOs contain a Lac core (Galβ1-4Glc) at the reducing end. The Lac core can be elongated at the C-3 position of the galactose residue with a lacto-N-biose (LNB; Galβ1-3GlcNAc) or N-acetyllactosamine (Galβ1-4GlcNAc) unit(s) (23, 24). The resulting HMO structures are denoted as type I and type II chains, respectively, with lacto-N-tetraose (LNT) and lacto-N-neotetraose (LNnT) as archetypes. Type I/II chains and the Lac core are often decorated by Fuc and Neu5Ac residues via various α-glycosidic bonds.

Previous studies have revealed substantial variation in HMO utilization strategies and capabilities within the Bifidobacterium genus (25–29). For example, B. bifidum uses a set of membrane-attached glycoside hydrolases (GHs) that extracellularly degrade HMOs to di- and monosaccharides (30–35) and then imports and catabolizes liberated LNB and Lac (25). In contrast, B. longum subsp. infantis and B. breve import HMOs using ATP-binding cassette (ABC) transporters and then degrade the oligosaccharides taken up to monosaccharides intracellularly using a repertoire of exo-acting GHs (36).

B. longum subsp. infantis, which is widely used in probiotic and synbiotic formulations (8, 15–17, 37), possesses several unique gene clusters (e.g., HMO cluster I or H1) that encode the most elaborate HMO uptake (38–40) and intracellular degradation machinery (41–44) among bifidobacteria (Fig. 1A and B). While the molecular mechanisms of HMO utilization by B. longum subsp. infantis have received considerable attention, the regulatory mechanisms governing this catabolic process remain poorly understood. Previous studies demonstrated that LNT and LNnT induce a profound and surprisingly similar transcriptomic response in B. longum subsp. infantis ATCC 15697, with multiple gene clusters (nag, lnp, H1) (Fig. 1B) being upregulated, pointing to a possible global regulatory mechanism(s) (45, 46). Uncovering this yet unknown mechanism, in addition to fundamental importance, has a potential translational value for the rational selection of individual HMOs for prebiotic and synbiotic formulations.

FIG 1.

Reconstructed NagR regulon in B. longum subsp. infantis ATCC 15697. (A) Schematic representation of LNT, LNnT, and LNB/GNB utilization pathways in B. longum subsp. infantis ATCC 15697. (Step 1) HMOs and their constituents are transported into the cell by various transport systems. (Step 2) Once inside the cell, HMOs are degraded from the nonreducing end by a coordinated action of exo-acting GHs. Breakdown of glycosidic bonds by specific GHs is indicated by light orange circles. (Step 3) Released monosaccharides are converted to fructose-6P and enter the bifid shunt. (B) Gene clusters constituting the reconstructed NagR regulon in B. longum subsp. infantis ATCC 15697. Numbers represent locus tags in the Blon_XXXX format (GenBank accession no. CP001095.1). (C) NagR-binding motif in B. longum subsp. infantis ATCC 15697 based on 11 predicted operators.

Comparative genomic analysis across multiple related genomes is a powerful approach for reconstructing transcriptional regulatory networks (regulons) controlling carbohydrate utilization (47–49). Our earlier bioinformatic analysis conducted on a limited set of Bifidobacterium genomes implicated NagR, a transcription factor (TF) from the ROK family, as a regulator of nag and lnp clusters (Fig. 1B) encoding GlcNAc and lacto-N-biose/galacto-N-biose (LNB/GNB) catabolic pathways, respectively, in B. bifidum, B. breve, B. longum subsp. infantis, and B. longum subsp. longum (48). James et al. further experimentally confirmed this prediction in B. breve UCC2003 (50). However, it was unclear whether and how this knowledge translated to the global regulation of the extensive HMO utilization machinery in B. longum subsp. infantis.

Here, we reconstructed NagR regulons in a substantially larger collection of Bifidobacteriaceae genomes focusing on HMO-utilizing species. This analysis revealed multiple putative NagR-binding sites (operators) in the B. longum subsp. infantis ATCC 15697 genome, suggesting the role of NagR as a global negative regulator of LNB/GNB, LNT, LNnT, and potentially other HMO utilization in this bacterium. This conjecture was corroborated by transcriptome profiling of the nagR knockout mutant and by direct assessment of the binding of recombinant NagR to its predicted operators. The inferred NagR regulon structure in B. longum subsp. infantis indicates that this microorganism is adapted to simultaneous utilization of multiple HMOs, suggesting the use of rationally formulated HMO mixtures rather than individual oligosaccharides as prebiotics. The reconstructed NagR regulons also provide insights into the evolution of complex regulatory networks controlling carbohydrate metabolism in bifidobacteria.

RESULTS

Genomic reconstruction reveals the complexity of the NagR regulon in B. longum subsp. infantis.

We used a position weight matrix (PWM)-based approach to reconstruct the NagR regulon in B. longum subsp. infantis ATCC 15697 and identified 11 potential NagR-binding sites (operators) in promoter regions of genes/operons encoding components of HMO utilization pathways, including six previously unknown NagR operators within in the H1 cluster (Fig. 1A and B and see Table S3 in the supplemental material). Among the new putative NagR regulon members were genes encoding (i) LNnT (type II HMO) ABC transporters (HmoABC, HmoA2B2C2), (ii) substrate-binding components of ABC transporters possibly involved in HMO uptake (HmoA3, HmoA4, HmoA5, HmoA6), (iii) N-acetylneuraminate lyase NanA2 and GH33 family α-sialidase NanH2, and (iv) lactose permeases (LacS, LacS2) (Fig. 1A and Table 1). Overall, the reconstructed regulon contained 29 genes. The inferred NagR-binding motif had a palindrome structure (Fig. 1C), a common feature of ROK family transcriptional regulators (51).

TABLE 1.

Composition of the reconstructed NagR regulon in B. longum subsp. infantis ATCC 15697

Locus tag	Gene	Annotation	Classificationa	Reference
Blon_0879	nagK	Putative ROK family N-acetylglucosamine kinase	EC 2.7.1.59
Blon_0880	nagR	Transcriptional regulator of LNB/GNB and HMO utilization	ROK family	This study
Blon_0881	nagB	Glucosamine-6-phosphate deaminase	EC 3.5.99.6
Blon_0882	nagA	N-Acetylglucosamine-6-phosphate deacetylase	EC 3.5.1.25
Blon_0883	gltF	LNB/GNB ABC transporter substrate-binding protein	TC 3.A.1.1	39
Blon_0884	gltG	LNB/GNB ABC transporter permease protein 1	TC 3.A.1.1
Blon_0885	gltH	LNB/GNB ABC transporter permease protein 2	TC 3.A.1.1
Blon_2171	lnpD	UDP-hexose 4-epimerase	EC 5.1.3.2	59
Blon_2172	lnpC	UTP-hexose-1-phosphate uridylyltransferase	EC 2.7.7.10	59
Blon_2173	lnpB	N-Acetylhexosamine 1-kinase	EC 2.7.1.162	59
Blon_2174	lnpA	1,3-β-Galactosyl-N-acetylhexosamine phosphorylase	EC 2.4.1.211	63
Blon_2175	gltC	LNT/LNB/GNB ABC transporter permease protein 2	TC 3.A.1.1.48
Blon_2176	gltB	LNT/LNB/GNB ABC transporter permease protein 1	TC 3.A.1.1.48
Blon_2177	gltA	LNT/LNB/GNB ABC transporter substrate-binding protein	TC 3.A.1.1.48	39
Blon_2331	lacS2	Lactose MFS permease-2	TC 2.A.2
Blon_2332	lacS	Lactose MFS permease-1	TC 2.A.2	84
Blon_2341	-	Hypothetical protein
Blon_2342	hmoC2	Type II HMO ABC transporter permease protein 2	TC 3.A.1.1
Blon_2343	hmoB2	Type II HMO ABC transporter permease protein 1	TC 3.A.1.1
Blon_2344	hmoA2	Type II HMO ABC transporter-2 substrate-binding protein	TC 3.A.1.1	39
Blon_2345	hmoC	Type II HMO ABC transporter permease protein 2	TC 3.A.1.1
Blon_2346	hmoB	Type II HMO ABC transporter permease protein 1	TC 3.A.1.1
Blon_2347	hmoA	Type II HMO ABC transporter-1 substrate-binding protein	TC 3.A.1.1	39
Blon_2348	nanH2	HMO cluster α-2,3/6-sialidase	GH33	41
Blon_2349	nanA2	N-Acetylneuraminate lyase-2	EC 4.1.3.3
Blon_2350	hmoA3	Putative HMO ABC transporter-1 substrate-binding protein	TC 3.A.1.1	39
Blon_2351	hmoA4	Putative HMO ABC transporter-2 substrate-binding protein	TC 3.A.1.1	39
Blon_2352	hmoA5	Putative HMO ABC transporter-3 substrate-binding protein	TC 3.A.1.1	39
Blon_2354	hmoA6	Putative HMO ABC transporter-4 substrate-binding protein	TC 3.A.1.1	39

^aTransporter Classification Database (TC) numbers for transporters, the CAZy family for GHs, Enzyme Commission (EC) numbers for downstream catabolic enzymes, and the TF family for transcription factors are shown.

TABLE S3

Reconstructed NagR regulons and predicted NagR operators in 25 Bifidobacteriaceae genomes. Download Table S3, XLSX file, 0.02 MB ^{(25.3KB, xlsx)} .

Copyright © 2022 Arzamasov et al.

This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.

ROK family TFs can function as transcriptional activators or repressors (51). To infer the possible mode of action of NagR, we analyzed the position of predicted NagR operators relative to −35 and −10 promoter elements recognized by bacterial RNA polymerase holoenzyme. Ten of 11 predicted NagR operators overlapped with either −10 or −35 sequences (Fig. S1), suggesting that NagR was a potential transcriptional repressor. Therefore, based on the genomic reconstruction, we hypothesized that NagR was a global negative regulator of gene clusters involved in LNB/GNB, LNT, and LNnT utilization in B. longum subsp. infantis ATCC 15697.

FIG S1

Multiple-sequence alignments of promoter regions of NagR-regulated genes in Bifidobacterium spp. Predicted NagR operators are highlighted in yellow. Predicted promoter elements (−35 and −10 sequences) are in boldface black. Experimentally described transcription start sites (TSS) in B. breve UCC2003 (77) are in boldface blue. Predicted Shine-Dalgarno sequences are in boldface red. Coding sequences are highlighted in gray; initiation codons are underlined. Conserved nucleotides are marked with asterisks. Download FIG S1, PDF file, 0.2 MB ^{(174.8KB, pdf)} .

Copyright © 2022 Arzamasov et al.

This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.

Engineered nagR insertional mutant displays comparable yet distinct physiological properties compared to the parental wild-type strain.

To experimentally study the proposed regulatory role of NagR, we generated a nagR knockout (nagR-KO) mutant of B. longum subsp. infantis ATCC 15697 by insertional mutagenesis and verified the insertion position using genomic PCR (Fig. S2A and B). The Sp^r gene insertion at the nagR locus was maintained in the nagR-KO genome for at least 30 generations without antibiotic pressure (Fig. S2C). We used empirical area under the curve (AUC) to integrate information from growth curves of nagR-KO and wild-type (WT) strains cultivated in MRS-CS medium (see Materials and Methods) supplemented with various carbon sources (Fig. 2A). The AUC values for nagR-KO and WT strains grown in MRS-CS-Lac or MRS-CS-HMO were not significantly different (Fig. 2B). In contrast, the nagR-KO mutant had significantly lower AUC when grown in MRS-CS supplemented with sucrose and significantly higher AUC in the medium containing fructooligosaccharides (FOS) (Fig. 2B).

FIG 2.

Growth of B. longum subsp. infantis ATCC 15697 WT and nagR-KO strains in MRS-CS supplemented with various carbon sources (1% [wt/vol]). (A) Growth curves obtained by measuring OD₆₀₀ at specific time points. Data represent the mean ± SD from three biological replicates. Time points where the OD₆₀₀ values for WT and nagR-KO strains were significantly different (*, P_adj < 0.05) were identified using linear regression. Bonferroni correction was used to adjust for multiple testing. (B) Empirical AUC calculated by integrating the areas under growth curves. Data represent the mean ± SD from three biological replicates. Means were compared via Student's t test: ns, not significant (P > 0.05); *, P ≤ 0.05; **, P ≤ 0.01; ***, P ≤ 0.01.

FIG S2

Construction and stability of the B. longum subsp. infantis ATCC 15697 nagR-KO mutant. (A) Schematic representation of insertional inactivation of nagR. The region used for a single-crossover recombination event is in orange. The primers for genomic PCR (green) were designed to anneal outside the region used for recombination. (B) Results of genomic PCR. The amplicon sizes were expected to be 679 bp and 3,164 bp for the WT and nagR-KO strains, respectively. Four clones were analyzed. (C) Stability of the Sp^r gene insertion after ~30 generations without antibiotic pressure. Download FIG S2, PDF file, 2.4 MB ^{(2.4MB, pdf)} .

Copyright © 2022 Arzamasov et al.

This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.

HMO consumption profiling of the WT and nagR-KO mutant grown in MRS-CS-HMO revealed that both strains completely salvaged Lac, LNT, LNnT, 2′/3-fucosyllactose (FL), and difucosyllactose (DFL) after 24 h (Fig. S3A). Compared to the WT strain, the nagR-KO mutant consumed more LNT at the 4- and 8-h time points and displayed delayed consumption of large fucosylated HMOs, namely, lacto-N-fucopentaoses (LNFP) I/II/III, and lacto-N-difucohexaoses (LNDFH) I/II (Fig. S3A). The mutant also expelled less Fuc into the medium (Fig. S3A). Metabolic profiling of supernatants of the WT and nagR-KO strains cultivated in MRS-CS-HMO revealed that both strains released similar quantities of acetate (Fig. S3B). However, the nagR-KO mutant produced significantly more formate and less lactate at the 8-, 12-, and 24-h time points (Fig. S3B). These data demonstrated comparable yet distinct physiological properties of WT and nagR-KO strains, prompting a follow-up transcriptome profiling to infer genes differentially expressed in the mutant.

FIG S3

Metabolic profiling of the B. longum subsp. infantis ATCC 15697 nagR-KO mutant. (A and B) HMO consumption (A) and organic acid production (B) profiles of B. longum subsp. infantis ATCC 15697 WT and nagR-KO strains grown in MRS-CS-HMO. Data points represent the mean of three biological replicates. Error bars depict 95% confidence intervals for the mean. Time points where the metabolite concentrations for WT and nagR-KO strains were significantly different (*, P_adj < 0.05) were identified using linear regression. Bonferroni correction was used to adjust for multiple testing. Download FIG S3, PDF file, 0.1 MB ^{(137.8KB, pdf)} .

Copyright © 2022 Arzamasov et al.

This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.

Comparative transcriptomics corroborates the role of NagR as a global negative regulator of LNB/GNB, LNT, and LNnT utilization in B. longum subsp. infantis.

We used transcriptome sequencing (RNA-seq) to compare transcriptomes of nagR-KO and WT strains of B. longum subsp. infantis ATCC 15697 grown in MRS-CS supplemented with Lac or LNnT, yielding four experimental conditions (Fig. 3A). The choice of carbon sources was based on a previous study, where LNnT had induced many HMO utilization genes and Lac had been used as a comparator (45). Principal component analysis (PCA) of TMM-normalized (52) counts revealed that each experimental condition formed a distinct cluster (Fig. 3B). Specifically, PCA separated samples by carbon source (principal component 1 [PC1], 38.8% of the total variance) and strain (PC2, 21.1% of the total variance).

FIG 3.

RNA-seq of WT and nagR-KO strains grown in MRS-CS supplemented with Lac or LNnT. (A) Schematic representation of the experimental design; (B) PCA of TMM-normalized count data. Each data point represents one sample. (C and D) Volcano plots depicting the log₂ FC of gene expression versus the −log₁₀ P_adj. Panel C compares nagR-KO and WT strains grown in MRS-CS-Lac, whereas panel D compares the WT strain grown in MRS-CS-LNnT and MRS-CS-Lac. Criteria for calling differentially expressed genes were as follows: P_adj < 0.01 and absolute FC > 2. Genes constituting the reconstructed NagR and CscR regulons are colored in orange and purple, respectively.

Linear modeling implemented in the limma framework (53) revealed significant upregulation (fold change [FC] of >2 and adjusted P [P_adj] value of <0.01) of multiple nag, lnp, and H1 cluster genes in the nagR-KO strain grown in MRS-CS-Lac compared to the WT grown in MRS-CS-Lac (Fig. 3C and Table S2A). Overall, 19 out of 29 genes constituting the reconstructed NagR regulon were upregulated. These results demonstrate that NagR, in line with the bioinformatic prediction, functions as a global transcriptional repressor of the nag, lnp, and H1 loci in B. longum subsp. infantis ATCC 15697. In addition, we observed significant upregulation of the malEFG operon encoding an ABC transport system for maltose (39) and significant downregulation (FC <−2 and P_adj < 0.01) of cscA, cscB, and cscR in the nagR-KO mutant (Fig. 3C and Table S2A). The latter genes are involved in sucrose uptake and catabolism (54) and are predicted to be controlled by a local repressor from the LacI family, CscR (48). The downregulation of the csc genes was consistent with significantly decreased AUC of the nagR-KO mutant cultivated in the medium supplemented with sucrose (Fig. 2B). No potential NagR operators were identified in the promoter regions of malEFG and csc cluster genes, suggesting that the effect of the nagR knockout on their expression was indirect.

TABLE S2

List of differentially expressed genes (P_adj < 0.01; absolute log₂ FC > 1) between selected experimental conditions in the RNA-seq experiment. (A) nagR-KO mutant grown in MRS-CS-Lac versus WT grown in MRS-CS-Lac; (B) WT grown in MRS-CS-LNnT versus WT grown in MRS-CS-Lac. Download Table S2, XLSX file, 0.02 MB ^{(20.6KB, xlsx)} .

Copyright © 2022 Arzamasov et al.

This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.

RNA-seq of the WT strain revealed that 25 out of 29 genes constituting the predicted NagR regulon were upregulated during growth in MRS-CS-LNnT compared to MRS-CS-Lac (Fig. 3D and Table S2B). The expression of nag, lnp, and H1 cluster genes was higher in the WT strain cultured in the presence of LNnT than in the nagR-KO mutant cultured in MRS-CS-Lac (Fig. S4). These observations point to the presence of additional transcriptional activation mechanism(s) that may contribute to the upregulation of these gene clusters during the growth of B. longum subsp. infantis in the medium containing LNnT. In addition, we observed a significant upregulation of malK (Fig. S4 and Table S2B), encoding a shared ATPase component that potentially energizes multiple carbohydrate-specific ABC transporters in B. longum subsp. infantis (48). The upregulation of malK is likely tied with the enhanced energetic needs associated with the induction of genes encoding multiple ABC transport systems (GltFGH, GltABC, HmoABC, HmoA2B2C2, HmoA3 to A5).

FIG S4

Heat map depicting the expression of genes from Table S2A and B across four experimental conditions (WT and nagR-KO strains grown in MRS-CS supplemented with Lac or LNnT). Genes constituting the reconstructed NagR regulon are in boldface. Lines depict gene clusters from Fig. 1B. Download FIG S4, TIF file, 1.7 MB ^{(1.7MB, tif)} .

Copyright © 2022 Arzamasov et al.

This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.

Overall, the obtained transcriptomic data are consistent with the bioinformatic regulon reconstruction and demonstrate that NagR regulates LNB/GNB, LNT, and LNnT utilization in B. longum subsp. infantis by repressing genomic loci encoding (i) transporters of respective glycans and (ii) GlcNAc and LNB/GNB catabolic pathways.

Interaction of NagR with predicted operators is dependent on GlcNAc and its phosphorylated derivatives.

To test the interaction between NagR and its predicted binding sites, we cloned the nagR gene from B. longum subsp. infantis ATCC 15697 and expressed the recombinant protein as a fusion with an N-terminal His tag in Escherichia coli BL21(DE3). The electrophoretic mobility shift assay (EMSA) demonstrated that NagR specifically bound DNA fragments (probes) containing candidate NagR operators located upstream of nagK, nagB, gltA, hmoA, hmoA2, and hmoA3 genes (Fig. 4A and B and Fig. 5). Titration with increasing concentrations of NagR revealed that probes nagK, nagB_I, hmoA, hmoA2 had a high affinity for NagR (50% effective concentration [EC₅₀], 13 to 23 nM), whereas probes gltA and hmoA3 had a moderate one (EC₅₀, 140 to 180 nM) (Fig. 5). Reactions with probes nagB_II and hmoA6 did not manifest in robust shifts even at high (>500 nM) protein concentrations (Fig. S5A). These results demonstrate that NagR exhibits different affinities to its various operators. We also observed a significant negative correlation (Pearson R = −0.84, P < 0.05) between calculated EC₅₀ values for probes and expression FCs for corresponding genes in the RNA-seq experiment (nagR-KO mutant versus WT grown in MRS-CS-Lac) (Fig. 3C and 4C), indicating the more tightly NagR bound to an operator, the stronger was the observed upregulation of a respective gene in the nagR-KO strain.

FIG 4.

Interactions of recombinant NagR with predicted operators and screening of NagR effectors. (A) Predicted NagR operators identified in the promoter regions of listed genes. Conserved nucleotides are in boldface. The full sequences of probes used in EMSAs are given in Table S1. (B) EMSA gel showing interactions of NagR with DNA probes containing predicted operators. The NagR concentration was 25 nM, and probe concentrations were 1 nM. The ybdO probe was used as a control for nonspecific binding. (C) Correlation between probe EC₅₀ values determined via EMSA and expression FCs of cognate genes in the RNA-seq experiment (nagR-KO mutant versus WT grown in MRS-CS-Lac). (D) Effect of various GlcNAc metabolism intermediates (10 mM) on the interaction between NagR (25 nM) and the hmoA probe (1 nM). (E) EC₅₀ values of selected NagR effector molecules. The y axis depicts the ratio % shift with effector/% shift without effector. Calculated EC₅₀ values are shown as big circles. Concentrations of NagR and the hmoA probe were 25 nM and 1 nM, respectively.

FIG 5.

EMSA gels depicting titration of DNA probes (1 nM) containing predicted NagR operators with recombinant NagR. Gels were quantified, and the results were approximated by the 4PL equation. The NagR concentration at which half of the probe is shifted (EC₅₀) is shown. Gray shading depicts 95% confidence intervals. The x axis is in the log₁₀ scale.

FIG S5

(A) EMSA gels depicting titration of DNA probes (1 nM) containing predicted NagR operators with recombinant NagR; (B) EMSA gels depicting titration of the NagR-DNA probe complex with various effector molecules. Concentrations of NagR and the hmoA probe were 25 nM and 1 nM, respectively. Download FIG S5, TIF file, 1.6 MB ^{(1.6MB, tif)} .

Copyright © 2022 Arzamasov et al.

This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.

TABLE S1

Oligonucleotides used in this study. IRD oligonucleotides are 5′ labeled with the IRD700 dye, whereas RC oligonucleotides are unlabeled. Predicted NagR operators are in boldface. Download Table S1, PDF file, 0.08 MB ^{(77.7KB, pdf)} .

Copyright © 2022 Arzamasov et al.

This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.

To identify potential NagR effector molecules, we used a probe with a high affinity to NagR (hmoA) and added various GlcNAc metabolism intermediates (GlcNAc, GlcNAc-6P, GlcN-6P, GlcNAc-1P) to the binding reaction mixture. GlcNAc, GlcNAc-6P, and GlcNAc-1P, but not GlcN-6P, disrupted the NagR-DNA complex at a saturating (10 mM) concentration (Fig. 4D). Titration of the effectors revealed that GlcNAc-6P lost its complex-disrupting effect at 1 mM, whereas GlcNAc and GlcNAc-1P lost their effect at 0.1 mM (Fig. S5B). The calculated effector EC₅₀ values were 0.33 ± 0.05 mM for GlcNAc and 0.49 ± 0.06 mM for GlcNAc-1P (Fig. 4E).

Overall, the EMSA results demonstrate that (i) NagR binds its predicted operators with various affinities, and (ii) multiple GlcNAc metabolism intermediates disrupt NagR-DNA interactions in vitro and thus serve as plausible NagR transcriptional effectors in B. longum subsp. infantis.

The NagR regulon was a subject of evolutionary expansion in Bifidobacteriaceae.

We used the PWM-based approach to identify NagR binding motifs and reconstruct regulons in 25 genomes representing various phylogenetic lineages within the Bifidobacteriaceae family to trace the potential evolutionary history of this gene regulatory network. Overall, the sizes and compositions of the reconstructed NagR regulons markedly varied among the studied genomes (Fig. 6A and Table S3). For instance, early diverged Bifidobacterium species isolated from insects and dairy products possessed concise NagR regulons comprised of a single genomic locus (nag) encoding GlcNAc and predicted N,N′-diacetylchitobiose catabolic pathways (Text S1). This observation suggests that in ancestral bifidobacteria, the role of NagR was likely confined to local regulation of the GlcNAc catabolic pathway.

FIG 6.

Evolution of the NagR regulon and binding motif within the Bifidobacteriaceae family. (A) NagR regulon composition in 25 Bifidobacteriaceae strains mapped on a species tree built based on the alignment of 247 core genes. Bootstrap values are shown as purple circles. Black symbols indicate strain isolation sources. Regulon members are colored according to their function: catabolic enzymes are green, GHs are orange, and transporters are blue. (B) NagR-binding motifs mapped on a tree of NagR proteins from 25 Bifidobacteriaceae strains. Bootstrap values are shown as purple circles. NagR paralogs in B. bohemicum and B. bifidum are denoted by the numbers 1 and 2.

TEXT S1

Supplemental methods and results. Download Text S1, DOCX file, 0.09 MB ^{(88.4KB, docx)} .

Copyright © 2022 Arzamasov et al.

This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.

We observed gradual NagR regulon expansion in multiple Bifidobacterium genomes isolated from mammalian hosts (Fig. 6A, Table S3, and Text S1). The largest and most complex NagR regulons were identified in strains isolated from the human neonatal gut, namely, B. longum subsp. infantis ATCC 15697 and B. bifidum PRL2010. While in B. longum subsp. infantis, the NagR regulon expanded to include multiple HMO transporters encoded within the H1 cluster (Fig. 1B), in B. bifidum, the reconstructed regulon contained genes encoding multiple GHs involved in the extracellular degradation of HMOs and mucin O-glycans (Text S1). Additionally, the NagR regulon expansion was accompanied by minor variations in the NagR-binding motif (Fig. 6B and Text S1). Taken together, the comparative regulon reconstruction suggests that NagR evolved from a local regulator of GlcNAc catabolism in ancestral bifidobacteria to a global regulator of utilization of various host glycans (e.g., HMOs) in species isolated from the mammalian neonatal gut.

DISCUSSION

Regulation of HMO utilization in B. longum subsp. infantis.

The predominance of bifidobacteria in the infant gut is linked to their ability to consume and use HMOs as a carbon source (25, 26, 38, 40, 55, 56). B. longum subsp. infantis possesses a unique gene cluster (H1) that encodes multiple catabolic enzymes and components of ABC transporters that endow this species with the ability to utilize a multitude of HMOs (25, 36, 38). Previous studies demonstrated that a pooled HMO mixture and individual HMOs (LNT and LNnT) induce the expression of H1 and nag cluster genes in B. longum subsp. infantis ATCC 15697, suggesting that H1 acts as an HMO-inducible unit and is coregulated with the GlcNAc catabolic pathway (45, 46). However, the regulatory mechanisms underlying this phenomenon were not elucidated.

In this study, we have established NagR-mediated repression of H1, lnp, and nag cluster genes in B. longum subsp. infantis ATCC 15697 by combining PWM-based regulon reconstruction with transcriptome profiling of the nagR-KO mutant. The composition of the NagR regulon suggests that this global TF regulates the utilization of LNB/GNB, LNT, LNnT, and potentially other decorated (e.g., sialylated) type I and II HMOs in B. longum subsp. infantis. We have also demonstrated the concentration-dependent binding of recombinant NagR to its predicted operators in vitro. The EC₅₀ values inferred from EMSAs negatively correlated with fold change values for upregulated genes in the nagR-KO mutant. Thus, the degree of gene repression by NagR is strongly dependent on the affinity of this TF to its cognate operators in the promoter regions of corresponding genes.

We identified GlcNAc and its phosphorylated derivatives, GlcNAc-6P and GlcNAc-1P, as potential NagR transcriptional effectors in B. longum subsp. infantis ATCC 15697. This result is somewhat unexpected since James et al. previously reported GlcNAc-6P, but not GlcNAc, as the NagR transcriptional effector in B. breve UCC2003 (50). This discrepancy may reflect a metabolic adaptation to the more global nature of NagR regulon in B. longum subsp. infantis, although, alternatively, it may be due to the differences in the experimental approach, which in our case was based on the use of purified recombinant NagR for EMSAs versus crude cell lysate as in the previous study (50). EMSA data indicated that the acetyl group of these intermediary metabolites played a crucial role in NagR-effector interactions, whereas phosphorylation of GlcNAc appeared to be dispensable. While GlcNAc and GlcNAc-6P have been described as transcriptional effectors of the ROK family TFs (50, 57, 58), the potential effector role of GlcNAc-1P is novel and unexpected. Although N-acetylhexosamine 1-kinase (LnpB) can phosphorylate GlcNAc to GlcNAc-1P (59), no enzymes that convert the latter to GlcNAc-6P and thus shunt it to the GlcNAc catabolic pathway have been described in prokaryotes (60). In contrast, GlcNAc-1P can be converted to UDP-GlcNAc by GlcNAc-1P uridyltransferase (GlmU) and enter the peptidoglycan biosynthesis pathway (61). Therefore, additional studies are required to assess the biological significance of GlcNAc-1P functioning as a potential NagR transcriptional effector.

Based on the obtained data, we propose a model (Fig. 7) in which the release of GlcNAc during degradation of LNT and LNnT by intracellular GHs results in derepression of nag, lnp, and H1 clusters, including genes encoding LNT and LNnT transporters. This model explains the similarity of transcriptomic responses of B. longum subsp. infantis to LNT or LNnT in batch cultures (45, 46) and gnotobiotic mice (15) and suggests that utilization of any GlcNAc-containing glycan by this bacterium will result in the upregulation of NagR-controlled genes. Among these glycans might be particular fucosylated HMOs (e.g., LNFP I) and milk N-glycans imported by other ABC transport systems (15, 40). Consistent with this notion, a previous study demonstrated that B. longum subsp. infantis upregulates nag and H1 cluster genes (62) when utilizing N-glycosylated human lactoferrin. The proposed model, however, does not explain all transcriptional responses observed during HMO utilization by B. longum subsp. infantis. The transcriptomic profile of the WT strain grown in MRS-CS-LNnT suggests that additional mechanisms may coactivate the expression of NagR regulon genes (particularly within the lnp cluster) under physiologically inducing conditions.

FIG 7.

Model of NagR-mediated regulation of HMO utilization in B. longum subsp. infantis. (Step 1) GlcNAc-containing milk glycans (e.g., LNT and LNnT) are taken up into the cell by various ABC transporters. (Step 2) Once inside the cell, the glycans are degraded by intracellular GHs, and GlcNAc is released. (Step 3) Released GlcNAc interacts with NagR and disrupts the NagR-operator complex, leading to derepression of NagR-controlled genes.

The structure of the NagR-mediated transcriptional network in B. longum subsp. infantis likely reflects the evolutionary adaptation of this bacterium to simultaneous foraging of multiple distinct HMOs and other milk glycans. This notion suggests that using a mixture of LNT and LNnT (and potentially other HMOs) rather than individual oligosaccharides as a prebiotic may be a more efficient solution for selective stimulation of B. longum subsp. infantis growth in the neonatal gut since it considers the nuanced regulatory mechanisms and physiology of the target organism.

Evolution of the NagR regulon in bifidobacteria.

Evolution of B. longum subsp. infantis was shaped by its ecological niche, specifically adaptation to the foraging of dietary milk glycans (e.g., HMOs) abundantly present in the gut of breastfed infants (38). This adaptation led to the emergence of several unique gene clusters, such as H1 (38), controlled by a complex NagR regulatory network. However, other Bifidobacterium species inhabiting the neonatal gut, such as B. breve and B. longum subsp. longum, do not harbor the H1 cluster and have less complex NagR regulons (48, 50). Thus, to elucidate the plausible evolutionary history of the NagR regulon, we reconstructed its content in 25 representative genomes spanning 18 Bifidobacterium species and one Alloscardovia species isolated from various hosts and environments.

The regulon structure in early diverged Bifidobacterium species (e.g., B. asteroides and B. aquikefiri) suggests that NagR potentially functioned as a local regulator of a single gene cluster involved in GlcNAc and possibly N,N′-diacetylchitobiose catabolism in ancestral bifidobacteria. Bifidobacteria colonizing mammalian hosts have acquired various gene clusters encoding the catabolic machinery involved in host glycan utilization. For example, most bifidobacteria isolated from mammals (including humans) harbor the lnp cluster encoding a transporter and a catabolic pathway for LNB and GNB, structural components of various milk (HMOs and glycolipids) and intestinal (mucin O-glycans) glycans, respectively (59, 63–65). The involvement of the lnp cluster in the reconstructed NagR regulons suggests transitioning from a local to a multilocus-controlling TF. Bifidobacterium pseudocatenulatum and two closely related species lacking the lnp genes have NagR regulons potentially expanded to control the LNB/LNT utilization machinery (gltABC-nagK-hex1-nagB-nagA and bga42A genes). The structure of the reconstructed NagR regulon in B. pseudocatenulatum is consistent with a recently published transcriptomic data set in which the expression of these genes was induced by LNFP I (66). Although the import of LNFP I is mediated by an ABC transport system for fucosylated HMOs (55, 66) and not GltABC, the GlcNAc released during the degradation of this oligosaccharide was likely responsible for the derepression of NagR-controlled genes.

Independent NagR regulon expansion events potentially occurred in prevalent infant-associated species with the highest HMO utilization potential: B. longum subsp. infantis and B. bifidum. In two closely related strains of B. longum subsp. infantis, the reconstructed regulon includes multiple HMO transporter genes from the H1 cluster, whereas in B. bifidum, NagR potentially regulates extracellular GHs involved in HMO and mucin O-glycan degradation. These observations provide a fascinating example of how catabolic machinery corresponding to two distinct strategies of HMO utilization (intracellular in B. longum subsp. infantis versus extracellular in B. bifidum) may have converged toward transcriptional control by the same TF. The structure of the reconstructed NagR regulon in B. bifidum PRL2010 partially explains the upregulation of nag, lnp, and genes encoding extracellular GHs during the growth of this strain in a mucin-supplemented medium (56).

Potential expansion of regulons for carbohydrate metabolism genes has been previously described in various bacterial lineages (67, 68), including bifidobacteria (69, 70). However, the underlying rationale of these expansion events to include specific genes/operons was not always straightforward. Here, we hypothesize that the NagR regulon expansion to control genes involved in the catabolism of GlcNAc-containing host glycans in Bifidobacterium might be linked to the ability of NagR orthologs to sense GlcNAc and/or its phosphorylated derivatives (50). Interestingly, other ROK family TFs have been implicated in regulating the utilization of GlcNAc-containing glycans in bifidobacteria. For example, in B. breve UCC2003, while NagR functions as a negative regulator of LNB/GNB and LNT utilization pathways, its paralog, NahR, represses a gene encoding an LNnT transporter (50). Another NagR paralog in this strain, AtsR2, represses a gene cluster involved in utilizing GlcNAc-6S, a mucin O-glycan-constituting saccharide released by 6-sulfo-β-N-acetylglucosaminidase BbhII (71, 72). Different GlcNAc derivatives were reported as effector molecules of these TFs: GlcNAc for NahR and GlcNAc-6S for AtsR2 (50, 71). Finally, a recently identified NagR paralog in B. longum subsp. infantis Bg_2D9, NglR, potentially controls a gene cluster involved in the metabolism of complex N-glycans (15). These data suggest that the evolution of gene regulatory networks governing the utilization of GlcNAc-containing glycans in bifidobacteria was not limited to the NagR regulon expansion. Another scenario might have involved duplication(s) of the nagR gene after speciation followed by the functional divergence of emerged paralogs. Overall, these observations illustrate how bifidobacteria adapted to regulate the foraging of host glycans during the colonization of mammalian hosts and shed more light on how complex regulatory networks emerge and evolve.

MATERIALS AND METHODS

Reagents and bacterial strains.

Reagents were purchased from Alfa-Aesar (Tewksbury, MA, USA), Ambion (Austin, TX, USA), Combi-Blocks (San Diego, CA, USA), Sigma-Aldrich (St. Louis, MO, USA), and Invitrogen (Carlsbad, CA, USA), unless indicated otherwise. Synthetic LNnT (>95% purity) was generously donated by DSM (Heerlen, Netherlands). The HMO mixture was prepared from pooled human milk (55). The experimental protocol was reviewed and approved by the Ethics Committee of Kyoto University (R0046); the study was performed per the Declaration of Helsinki, and informed consent was obtained from all mothers (all subjects). Oligonucleotides were synthesized by Integrated Genomic Technologies (Coralville, IA, USA). Phusion high-fidelity DNA polymerase, restriction enzymes, and Quick Ligase were purchased from New England BioLabs (Ipswich, MA, USA).

The type strain of B. longum subsp. infantis (ATCC 15697 = JCM 1222) was obtained from Japan Collection of Microorganisms (RIKEN BioResource Research Center, Tsukuba, Japan). Escherichia coli DH5α and One Shot TOP10 (Invitrogen) cells were used for genetic manipulations. E. coli BL21(DE3) (New England BioLabs) was used for recombinant NagR overexpression.

Bioinformatic analysis.

A previously established comparative genomics approach was used to identify putative NagR binding sites and reconstruct regulons in B. longum subsp. infantis ATCC 15697 and other selected strains (48, 69). For the initial NagR regulon reconstruction, we built a PWM based on data available in the RegPrecise database (48, 73) using SignalX (74). To improve the identification of NagR operators, we built additional PWMs representing two different NagR-binding motifs specific for distant bifidobacterial lineages. Constructed PWMs were used to search for new potential NagR operators using GenomeExplorer (74) with the following parameters: (i) positions −500 to +50 bp relative to the first codon of a gene and (iii) site score threshold of 4.3. Identified sites were screened using the consistency check and phylogenetic footprinting approaches to filter out false positives (75). NagR-binding motifs were visualized via WebLogo (76). Positions of the −10 and −35 promoter elements were determined via similar PMW-based searches based on data available for B. breve (77) and B. longum (78). Promoter regions were aligned using Pro-Coffee (79). Details on additional genomic analysis of Bifidobacterium strains and phylogenetic inference are available in Text S1 in the supplemental material.

Targeted nagR gene disruption in B. longum subsp. infantis ATCC 15697.

A single-crossover recombination event was used to inactivate the nagR gene (Blon_0880; UniProt entry B7GQA0) in B. longum subsp. infantis ATCC 15697. Briefly, a BamHI-digested, 2.0-kb fragment of pBS423 (80) that carries the pUC ori and a spectinomycin resistance (Sp^r) gene was self-ligated to generate pTK2051, a plasmid incapable of replicating in bifidobacteria. The internal region of nagR was then amplified by PCR using primers NagR_I/NagR_II (Table S1) and genomic DNA as a template. The amplified 0.5-kbp fragment was inserted into the BamHI site of pTK2051 using the In-Fusion Snap Assembly kit (TaKaRa Bio USA, Mountain View, CA, USA). The resulting suicide plasmid was introduced into B. longum subsp. infantis by electroporation (40). To prepare electrocompetent cells, B. longum subsp. infantis ATCC 15697 was grown in 50 mL of Gifu anaerobic medium (GAM; Nissui Pharmaceutical, Tokyo, Japan) to an optical density at 600 nm (OD₆₀₀) of 0.2. Cells were harvested by centrifugation (4,800 × g for 15 min at 4°C), washed with ice-cold 1 mM ammonium citrate buffer (pH 6.0) containing 50 mM sucrose, and resuspended in 400 μL of the same buffer. An aliquot (200 μL) was mixed with 10 μg of the suicide plasmid and then pulsed using a Gene Pulser Xcell system (Bio-Rad Laboratories, Hercules, CA, USA) with 10 kV/cm, 25 μF, and 200 Ω. The pulsed cells were immediately mixed with 800 μL of 1% (wt/vol) Lac-supplemented GAM and incubated at 37°C under anoxic conditions for 3 h before being spread on GAM agar plates containing 1% (wt/vol) Lac and 15 μg/mL spectinomycin (Sp). Colonies that appeared on the plates were subsequently subjected to a genomic PCR analysis at the nagR locus using primers NagR_III/NagR_IV (Table S1 and Fig. S2A). The amplicon was directly sequenced to ensure that the suicide plasmid was integrated into the intended site. The stability of the Sp^r gene insertion in the absence of antibiotic pressure was additionally monitored (Text S1).

Culture conditions.

B. longum subsp. infantis ATCC 15697 was routinely grown in GAM or Lactobacilli MRS broth without dextrose (Alpha Biosciences, Baltimore, MD, USA) with 0.34% (wt/vol) sodium ascorbate and 0.029% (wt/vol) l-cysteine–HCl monohydrate (MRS-CS). The MRS-CS medium was supplemented with Lac, LNnT, or a mixture of neutral HMOs at a final concentration of 1% (wt/vol). Cultures were incubated at 37°C in an AnaeroPack system (Mitsubishi Gas Chemical Company, Tokyo, Japan) or a chamber maintained with a gas mix of 10% H₂, 10% CO₂, and 80% N₂ (Coy Laboratory Products, Grass Lake, MI, USA). Growth was monitored by measuring either culture turbidity in McFarland units using a DEN-1B densitometer (Grant Instruments, Shepreth, United Kingdom) or optical density at 600 nm (OD₆₀₀) using a DU800 spectrophotometer (Beckman Coulter, Brea, CA, USA). E. coli strains were cultured in Luria-Bertani broth at 37°C with vigorous agitation. Where appropriate, growth media were supplemented with spectinomycin (15 μg/mL for B. longum subsp. infantis ATCC 15697 nagR-KO, 75 μg/mL for E. coli DH5α) or kanamycin (50 to 60 μg/mL for all other E. coli strains). Details on measuring the growth, HMO consumption, and organic acid production are described in Text S1.

Transcriptome analysis.

Overnight cultures of B. longum subsp. infantis ATCC 15697 WT and nagR-KO strains grown in MRS-CS-Lac (with Sp in the case of nagR-KO) were harvested, washed with sugar-free MRS-CS, and used to inoculate antibiotic-free MRS-CS medium supplemented with either Lac or LNnT (1% [wt/vol]) at an OD₆₀₀ of 0.02. Samples (2 mL, biological triplicates) were collected at the early to mid-exponential phase (OD₆₀₀ = 0.35) and immediately pelleted in a prechilled centrifuge at 4,800 × g for 5 min. Cell pellets were snap-frozen in liquid nitrogen and stored at −80°C until further use. RNA was extracted as described previously (81) with minor modifications; the detailed protocol can be found in Text S1. rRNA was depleted with the NEBNext rRNA depletion kit for bacteria (New England Biolabs). Barcoded libraries were made with NEBNext Ultra II directional RNA library prep kit for Illumina (New England Biolabs). Libraries were pooled and sequenced (single-end 75-bp reads) on Illumina NextSeq 500 using the High Output V2 kit (Illumina, San Diego, CA, USA). Sequencing data were analyzed as described previously (82) with certain modifications; the details are described in Text S1.

Cloning, expression, and purification of recombinant NagR.

Codon-optimized nucleotide sequence of nagR (Blon_0880) was synthesized by GeneArt gene synthesis (Thermo Fisher Scientific, Waltham, MA, USA), PCR amplified using primers NagR_HisN_F and NagR_HisN_R (Table S1), digested by BamHI and SalI, and ligated into a predigested in-house pET-49b(+) vector conferring resistance to kanamycin. The ligation mixture was introduced into E. coli One Shot TOP10 cells by chemical transformation, and transformants were then selected based on kanamycin resistance. The recombinant NagR was expressed as a fusion with the N-terminal His tag under the control of a T7 promoter in E. coli BL21(DE3). Cells were grown in LB medium (50 mL) at 37°C to an OD₆₀₀ of ~0.6 and then transferred to 16°C. Protein expression was induced by adding 0.2 mM IPTG (isopropyl-β-d-thiogalactopyranoside). Cells were grown at 16°C overnight and collected by centrifugation at 4,800 × g for 15 min. Harvested cells were resuspended in a lysis buffer containing 10 mM HEPES buffer (pH 7.0), 100 mM NaCl, 0.15% Brij-35, and 5 mM β-mercaptoethanol. Cells were lysed by a freeze-thaw cycle, followed by sonication using Misonix sonicator 3000 (Misonix, Inc., Farmingdale, NY, USA). The cell debris was removed; the soluble fraction was loaded onto a Ni-nitrilotriacetic acid (NTA) agarose minicolumn (0.2 mL) (Qiagen, Hilden, Germany). The column was washed with 10 column volumes of At buffer (50 mM Tris-HCl [pH 8.0], 500 mM NaCl, 20 mM imidazole, 0.3% Brij-35, 5 mM β-mercaptoethanol) and 10 column volumes of At buffer with 1 M NaCl. Captured proteins were eluted with 0.6 mL of At buffer with 300 mM imidazole. The eluted protein fraction was concentrated and buffer exchanged into 10 mM Tris-HCl (pH 8.0) with 50 mM NaCl using 30-kDa Amicon Ultra 0.5-mL centrifugal filters (Millipore Sigma, Burlington, MA, USA). The protein concentration was determined by a Qubit protein assay kit (Invitrogen).

Electrophoretic mobility shift assay.

Oligonucleotides containing predicted 21-bp NagR operators and surrounding genomic regions (14 bp from each end) were synthesized by Integrated DNA Technologies. The DNA fragment sequences, sizes, and labels used for testing are given in Table S1. Double-stranded labeled DNA probes were obtained by annealing IRD700-labeled oligonucleotides with unlabeled complementary oligonucleotides (ratio of 1:5) in a mixture of 4 mM Tris-HCl (pH 8.0), 20 mM NaCl, and 0.4 mM EDTA in a Mastercycler PRO thermal cycler (Eppendorf) overnight. Binding reactions were carried out with a final volume of 20 μL in binding buffer containing 10 mM Tris-HCl (pH 7.5), 50 mM KCl, 5 mM MgCl₂, 2.25 mM dithiothreitol (DTT), 0.125% Tween 20, and 2.5% glycerol. DNA probes (1 nM) were incubated with increasing concentrations of the purified NagR (0 to 2,000 nM) for 60 min at room temperature. Reaction mixtures were loaded on a Novex 6% DNA retardation gel (Thermo Fisher Scientific) and run in 0.5× Tris-borate-EDTA buffer (Thermo Fisher Scientific) at 100 V and room temperature for 45 min in XCell SureLock minicell electrophoresis system (Thermo Fisher Scientific). Gels were visualized using Odyssey CLx (Li-COR Biosciences, Lincoln, NE, USA). Bands were quantified in Image Studio v5.2 (Li-COR Biosciences). The resulting data were imported into R and approximated by a 4PL equation in the drc package (83) to calculate EC₅₀ values. In the 4PL model, the lower limit was fixed at 0 and the upper limit at 1. To identify possible NagR effectors, binding reactions were carried out with 25 nM NagR and the addition of 0.1 to 10 mM GlcNAc or its phosphorylated derivatives (GlcNAc-6P, GlcN-6P, GlcNAc-1P). EC₅₀ values for effectors were calculated using the 4PL equation, with the upper limit fixed at 1.

Data availability.

The RNA-seq data set has been deposited in the Gene Expression Omnibus under accession no. GSE196064. Raw EMSA gel quantification, growth, HMO consumption, and metabolic profiling data are available on GitHub (https://github.com/Arzamasov/NagR_manuscript). Code detailing the data analysis steps is available on GitHub and in Text S2.

TEXT S2

Code detailing the data analysis steps. Download Text S2, PDF file, 0.8 MB ^{(784.4KB, pdf)} .

Copyright © 2022 Arzamasov et al.

This content is distributed under the terms of the Creative Commons Attribution 4.0 International license.