Functional definition of NrtR, a remnant regulator of NAD⁺ homeostasis in the zoonotic pathogen Streptococcus suis

ABSTRACT

NAD⁺ is an enzyme cofactor required for the 3 domains of life. However, little is known about the NAD⁺ biosynthesis and salvage pathways in the opportunistic pathogen Streptococcus suis. A genome‐wide search allows us to identify the NAD⁺ salvage pathway encoded by an operon of nadR‐pnuC‐nrtR (from SSU05_1973 to SSU05_1971 on the reverse strand) in the S. suis 05ZYH33 that causes streptococcal toxin shock–like syndrome. The regulator of this pathway is Nudix–related transcriptional regulator (NrtR), a transcription regulator of the Nudix family comprising an N‐terminal Nudix‐like effector domain, and a C‐terminal DNA‐binding winged helix‐turn‐helix–like domain. Intriguingly, the S. suis NrtR naturally contains a single amino acid substitution (K92E) in the catalytic site of its Nudix domain that renders it catalytically inactive but does not influence its ability to bind DNA. Despite its lack of enzymatic activity, DNA‐binding activity of NrtR is antagonized by the effector ADP‐ribose. Furthermore, nrtR knockout in S. suis serotype 2 reduces its capacity to form biofilms and attenuates its virulence in a mouse infection model. Genome mining indicates that nrtR appears in a strain‐specific manner whose occupancy is correlated to bacterial infectivity. Unlike the paradigmatic member of NrtR family having 2 unrelated functions (Nudix hydrolase and DNA binding), S. suis 2 retains a single regulatory role in the modulation of NAD⁺ salvage. This control of NAD⁺ homeostasis contributes to S. suis virulence.—Wang, Q., Hassan, B. H., Lou, N., Merritt, J., Feng, Y. Functional definition of NrtR, a remnant regulator of NAD⁺ homeostasis in the zoonotic pathogen Streptococcus suis. FASEB J. 33, 6055–6068 (2019). www.fasebj.org

ABBREVIATIONS

ADPR

ADP‐ribose

AMP

ampicillin

BLAST

Basic Local Alignment Search Tool

EGS

ethylene glycol bis (succinimidyl succinate)

NA

nicotinic acid

Nam

nicotinamide

NiaR

niacin‐responsive DNA‐binding regulator

NMN

mononucleotide intermediate

NrtR

Nudix‐related transcriptional regulator

RNam

N‐ribosylnicotinamide

SS2

Streptococcus suis serotype 2

Spc^R

spectinomycin resistance

SsNrtR

S. suis NrtR

THB

Todd Hewitt broth

wHTH

winged helix‐turn‐helix

WT

wild type

NAD⁺ and NADP⁺ play crucial roles in cellular metabolism because they act as the primary biologic coenzymes for numerous redox reactions. In addition to its fundamental redox functions, NAD⁺ also serves as a cosubstrate in a number of reactions that deplete its cellular pool, such as ADP‐ribosylation of proteins, DNA repair by DNA ligase in bacteria, NAD⁺‐dependent protein deacetylation by CobB and Sirtuin family (1), calcium signaling (2, 3), and as a precursor of vitamin B₁₂. During oxidative stress, NADH accumulation can be toxic because it provides electrons for the formation of hydroxyl radicals by the Fenton reaction, whereas NADPH plays a protective role (4, 5). These numerous roles of pyridine nucleotides suggest a need for their tight regulation.

In most bacteria, NAD⁺ is synthesized de novo by a 3‐step pathway from L‐aspartate or by salvage pathways from various NAD⁺ precursors, such as nicotinic acid (NA), nicotinamide (Nam), or Nam riboside (NamR) (6). The de novo biosynthesis pathway of NAD⁺ from l‐aspartate to the intermediate of nicotinate mononucleotide proceeds via the consecutive actions of l‐aspartate oxidase, quinolinate synthetase, and quinolinate phosphoribosyltransferase encoded by the nadB, nadA, and nadC genes, respectively (7). Regulation of the de novo biosynthetic pathway of NAD⁺ has been well studied in Salmonella enterica serovar Typhimurium, in which the expression of nadB and nadA genes is controlled by the transcription regulator of NAD biosynthesis (NadR) (7, 8). However, NadR regulators and their corresponding regulons are largely restricted to a compact phylogenetic group of Enterobacteriaceae (9, 10). Another distinct transcriptional regulator of NAD⁺ synthesis is the niacin‐responsive DNA‐binding regulator YrxA (tentatively renamed NiaR), which represses transcription of the de novo biosynthesis operon nadABC and the niacin transporter gene niaP (formerly yceI). NiaR orthologs are also found in bacteria from the Bacillus and Clostridium group as well as Thermotogales (10). In the salvage or recycling pathway, NA, derived from NAD⁺ dissociation or from exogenous sources, is reconverted to NA mononucleotide by an NA phosphoribosyltransferase (PncB in E. coli; previously YueK in Bacillus subtilis) and reinserted in the enzymatic chain downstream of the reaction catalyzed by NadC (11).

Streptococcus suis is a leading agent of bacterial diseases such as septicemia in the swine industry worldwide (12) and also appears to be an opportunistic human pathogen with the involvement of streptococcal toxic shock–like syndrome (13, 14). Among the known 35 serotypes, S. suis serotype 2 (SS2) is the most virulent and the most frequently isolated serotype (12, 14). SS2 is a previously neglected but recently emerging human pathogen that results in a series of occupational and opportunistic infections (12–16). Although the NAD⁺ metabolism is linked to bacterial virulence in certain species such as Pseudomonas and Actinobacillus (17–19), little is known about the regulation and role of NAD⁺ in the pathobiology of SS2. Streptococcaceae lack the de novo NAD⁺ biosynthesis genes (10), and the NiaR regulon is only found in 2 Streptococcus species, S. thermophilus and S. pneumoniae, which branch together with Clostridiales, whereas most other streptococci occur in a distinct Lactobacillales branch (10). Our examinations of the S. suis genome suggest that NAD⁺ biosynthesis utilizes a salvage pathway mediated by the PnuC transporter and NadR proteins previously described for E. coli and S. enterica serovar Typhimurium (2). PnuC‐mediated transport of N‐ribosylnicotinamide (RNam) provides the ultimate precursor for NAD⁺ biosynthesis, whereas the adenylyltransferase activity of NadR can convert mononucleotide intermediate (NMN) directly to NAD⁺ (2). Regulators of the Nudix‐related transcriptional regulator (NrtR) family, currently annotated as ADP‐ribose (ADPR) pyrophosphatases from the Nudix family, are composed of an N‐terminal Nudix‐like effector domain and a C‐terminal, DNA‐binding, helix‐turn‐helix (HTH)–like domain. Here, we further characterize the function of NrtR in NAD⁺ biosynthesis and examine its role in SS2 virulence.

MATERIALS AND METHODS

Bacterial strains and growth conditions

Bacterial strains used in this study included derivatives of E. coli and SS2 (Supplemental Table S1). SS2 and its derivatives were grown at 37°C in Todd Hewitt broth (THB; Becton Dickinson, Franklin Lakes, NJ, USA) liquid medium on THB agar plates containing 5% (vol/vol) sheep blood (16). E. coli DH5α and E. coli BL21 (DE3) were cultured in Luria‐Bertani (LB) medium at 37°C with aeration or on LB agar plates. Necessary, appropriate antibiotics (MilliporeSigma, Burlington, MA, USA) were supplemented as follows: 100 µg/ml spectinomycin for S. suis and 100 µg/ml ampicillin (AMP) or 50 µg/ml of kanamycin for E. coli.

Knockout of nrtR and functional complementation

To test the role of NrtR, nrtR was mutated allelic replacement with a constitutively expressed spectinomycin resistance (Spc^R) cassette. First, the Spc^R gene cassette (amplified from pSET2) was inserted into a pUC19 vector (Promega, Madison, WI, USA) to create the recombinant plasmid pUC19‐Spc. Two DNA fragments [Left Arm (LA) and Right Arm (RA)] flanking the nrtR gene were cloned into pUC19‐Spc to generate the knockout plasmid pUC::nrtR. The pUC::nrtR plasmid was electroporated into the competent cells of strain 05ZYH33 (Supplemental Fig. S1). Spc^R transformants were confirmed using multiplex PCR (Supplemental Table S2). The fidelity of the double crossover recombination was confirmed in the mutants by PCR using flanking primers (P‐F/P‐R) lying outside the homologous regions followed by direct DNA sequencing. To create a complementation construct, a DNA fragment covering the nrtR coding region plus its 286‐bp upstream promoter sequence was PCR amplified from strain 05ZYH33 using the primers CnrtR‐F and CnrtR‐R. The resulting PCR product was cloned into an E. coli‐S. suis shuttle vector, pVA838, to yield plasmid pVA838::nrtR. This plasmid was used for subsequent functional complementation (13).

Plasmids and genetic manipulations

The S. suis nrtR (SsNrtR; SSU05_1971) was amplified by PCR with the primers SsnrtR‐F plus SsnrtR‐R (Supplemental Table S1) and cloned into the expression vector pET28(b) using NdeI and SalI restriction sites, resulting in the recombinant plasmid pET28‐nrtR (Supplemental Table S1). To prepare the NrtR protein, the expression plasmid pET28‐nrtR was transformed into E. coli DH5α and E. coli BL21(DE3). Also, the nrtR‐lacZ fusion reporter plasmid was constructed in 2 steps. First, a promoter‐less lacZ gene was cloned into the low‐copy shuttle vector pVA838 via NdeI and SalI restriction sites, resulting in the recombinant plasmid pVA838‐lacZ. In the second step, the promoter region of nrtR was amplified by PCR using S. suis 05ZYH33 genomic DNA as the template and then cloned into pVA838‐lacZ. The promoterless lacZ construct served as a negative control. To examine the role of nrtR in vivo, the reporter plasmid was separately transformed into the wild‐type (WT) strain and the ΔnrtR mutant. All recombinant plasmids involved in this study were verified by PCR and direct DNA sequencing.

Preparation and purification of NrtR protein

The N‐terminal hexa‐histidine–tagged NrtR protein was overexpressed in BL21(DE3) carrying pET28‐nrtR. Induction of bacterial cultures using 0.3 mM isopropyl β‐D‐1‐thiogalactopyranoside (IPTG) at an OD₆₀₀ nm of 0.6–0.8 at 30°C for 6 h gave partially soluble proteins (20). The bacterial cells were pelleted by centrifugation (4200 g for 20 min.), washed 3 times with ice‐cold PBS buffer (101.4 mM Na₂HPO₄, 1.8 mM KH₂PO₄, 137 mM NaCl, 2.7 mM KCl, 12% glycerol, pH 7.4), and resuspended in PBS buffer containing 20 mM imidazole. Clarified supernatants obtained after lysis in a French pressure cell were generated by centrifugation (16,000 g for 50 min). Supernatants were loaded onto a nickel chelate column (Qiagen, Hilden, Germany) for no longer than 3 h. After washing with 20 column volumes of PBS buffer containing 50 mM imidazole, the NrtR protein was eluted with 200 mM imidazole, dialyzed overnight to remove the residual imidazole, and then concentrated by ultrafiltration [10 kDa cutoff, Amicon Ultra (MilliporeSigma)] (21). The protein purity was judged by 12% SDS‐PAGE and quantified using Bradford method with Coomassie Brilliant Blue G250 (MilliporeSigma). Subsequently, the recombinant NrtR protein was further assayed by gel filtration chromatography as previously described by Feng and Cronan (21). Chemical cross‐linking was applied to define its solution structure of SsNrtR as previously described (22).

EMSAs

To test the functions of the predicted NrtR‐binding sites of S. suis, gel shift assays were used as previously described by Feng and Cronan (20). The double‐strand DNA probe (nadR‐F/R) was produced in vitro by annealing 2 complementary oligonucleotides in the TEN buffer (10 mM Tris‐HCl, 1 mM EDTA, 100 mM NaCl; pH 8.0). After 20 min of incubation of the DNA probe (0.5 pmol) with or without NrtR at room temperature, the DNA‐protein complexes were separated by native 7% PAGE. Finally, gels were stained with fluorescent dye and analyzed by a gel imaging and analysis system (Tanon 2500, Shanghai, China).

β‐galactosidase assays

Midlog‐phase cultures of S. suis carrying the lacZ fusion were analyzed for β‐galactosidase activity by Ye et al. (13). Bacterial lysates were prepared using a French press. The data were recorded in triplicate in at least 3 independent experiments.

Measurement of cytosolic NAD⁺/NADH level

The log‐phase bacterial cells were collected by centrifugation (10,000 g, 4°C, 10 min), suspended with lysis buffer at the ratio (1 × 10 cells/ml), and kept at room temperature for 15 min. Then, the treated cells were destroyed adequately by ultrasonic for 1 min. Following the 10 min of centrifugation at 10,000 g (4°C), the resultant supernatants were subjected to quantitative assays of NAD⁺ and NADH pool with Amplite Colorimetric NAD/ NADH Assay Kit (AAT Bioquest, Sunnyvale, CA, USA). Three individual experiments were conducted according to the manufacturer's protocol, and the data were recorded in triplicate.

Assay of enzymatic activity

An HPLC‐based assay was used to measure the Nudix hydrolase activity of recombinant NrtR (10). The reaction mixture contained 50 mM HEPES, pH 8.0, 5 mM MgCl₂ or 5 mM MnCl₂, 1 mMDTT, and 1 mM nucleoside diphosphate derivatives (ADPR, NADH, and NADPH) as the substrate with varying amounts of purified protein. The reaction was stopped with HClO₄ after incubating at 37°C for 20–30 min.

The instrument for HPLC‐UV analysis was composed of a pump (C18‐T column, 250 × 4.6 cm; Sigma‐Aldrich, St. Louis, MO, USA). An acetonitrile‐phosphoric acid (AN:PA = 2:8, v/v) mixture was used as a mobile phase. The flow rate of mobile phase, column temperature, and detection wavelength were set at 0.8 ml/min, 30°C, and 254 nm, respectively.

Site‐directed mutagenesis

Site‐directed mutagenesis of SsNrtR was performed as described by Feng et al. (23), with the corresponding primers (Supplemental Table S2) and recombinant plasmid pET28b‐nrtR as the PCR template. After digesting by DpnI (20 U/µl; New England Biolabs, Ipswich, MA, USA) at 37°C for 5 h to remove the residual template plasmid, the purified PCR products were transformed into chemically competent cells of E. coli DH5α and selected on kanamycin agar plates. The resulting clones were confirmed by DNA sequencing. The four mutant plasmids pET28b‐nrtR (H223A), pET28b‐nrtR (K246A), pET28b‐nrtR (R248A), and pET28b‐nrtR (K92E) were transformed into E. coli BL21(DE3) to obtain the recombinant proteins NrtR^H223A NrtR^K246A, NrtR^R248A, and NrtR^K92E, respectively.

Biofilm biomass assays

Cultures of SS2 WT and mutant strains grown to midexponential phase (OD₆₀₀ = 0.5) were transferred to 96‐well plates (diluted 1:100) and incubated at 37°C for 12 and 24 h. Adherent biofilms were washed by immersion in sterile distilled water to remove unattached cells. After briefly drying, 200 µl of aqueous 1% crystal violet was added to each well, and the plates were incubated at room temperature for 30 min. Next, the crystal violet solution was removed and the plates washed twice. After drying, the crystal violet dye (stained biofilm) was then solubilized by incubation with 200 µl ethanol per well for 30 min. A total of 100 µl of solubilized dye solution was transferred to the wells of a new microplate, and absorbance was measured at 570 nm in a spectrophotometer (Eon Microplate; BioTek Instruments, Winooski, VT, USA) (24).

Infection studies of mice

To reveal the role of nrtR in bacterial pathogenesis and virulence, 4‐wk‐old BALB/c mice (female, 10 mice/group) were challenged with the nrtR mutant and compared with the groups challenged by either the WT strain or the complemented strain. Sterile THB medium served as a negative control. Deaths were recorded, and moribund animals were humanely euthanized. All experiments on live vertebrates in this study were approved by the Ethics Committee of Research Institute for Medicine of Hangzhou Command and performed in accordance with the relevant guidelines and regulations.

Antibiotic susceptibility assays

In order to test whether NrtR was related to S. suis antibiotic resistance, microbroth dilution method was performed to address minimum inhibitory concentration (MIC) of 05ZYH33 (WT strain) and ΔnrtR (mutant strain), whose level of ≤0.25 µg/ml was treated susceptible. The antibiotics (8 in total) we used are categorized into 5 groups: 1) aminoglycoside antibiotics such as kanamycin, spectinomycin, and streptomycin; 2) tetracycline antibiotics; 3) β‐lactams antibiotics, including AMP and penicillin; 4) amino alcohol antibiotics, such as chloramphenicol; and 5) macrolide antibiotics, such as erythromycin. For MIC of streptomycin, kanamycin, and tetracycline, agar dilution method was applied for double checking.

Phylogenetic tree

Using Basic Local Alignment Search Tool (BLAST), close homologs of NrtR with >60% aa sequence identity were identified, which enabled us to exclude models and uncultured environmental samples and returned 1000 target sequences. Because of a predominance of homologs from Streptococcus species, the BLAST search was repeated with the same parameters excluding hits from Streptococcus (taxid: 1301), and hits with >30% aa sequence identity were collected and combined with the earlier BLAST result. Redundant sequences were eliminated using the Uniqueseq server (https://www.ncbi.nlm.nih.gov/CBBresearch/Spouge/html_ncbi/html/fasta/uniqueseq.cgi) and aligned using Multiple Sequence Comparison by Log Expectation (MUSCLE; https://www.ebi.ac.uk/Tools/msa/muscle/). In total, 49 unique protein sequences were utilized for phylogenetic analysis. jModeltest (via MEGA7) was used to identify the best‐fit protein substitution model, and the best model was used to generate a maximum‐likelihood tree with 1000 bootstrap replicates. A Log Gamma (LG) model with Gamma distribution and Invariant sites was used. Initial trees for the heuristic search were obtained automatically by applying Neighbor‐Join and BioNJ algorithms to a matrix of pair‐wise distances estimated using the Maximum Composite Likelihood approach and then selecting the topology with superior log likelihood value. The results were visualized as a radial phylogram.

Bioinformatic analysis

NrtR homologs from different Streptococcus species were collected by Protein BLAST (BLASTP; http://blast.st‐va.ncbi.nlm.nih.gov/Blast.cgi). The predicted NrtR‐binding sites appear in the Regprecise database (http://regprecise.lbl.gov). Multiple alignments of either NrtR proteins or NrtR‐binding sites were performed using the ClustalW2 program (http://www.ebi.ac.uk/Tools/clustalw2/index.html), and the final output was processed by the ESPript 2.2 server (http://espript.ibcp.fr/ESPript/cgibin/ESPript.cgi) (21). Structural modeling of SsNrtR was conducted with Swiss‐Model (https://www.swissmodel.expasy.org/).

RESULTS

Identification of an NAD⁺ salvage pathway in SS2

ANAD⁺ salvage operon encoding putative orthologs of nadR, pnuC, and nrtR was identified in the SS2 genome (Fig. 1A ). NrtR‐binding sites (http://regprecise.lbl.gov/RegPrecise) were located in the intergenic region upstream of the operon (Fig. 1B, C ), indicating that NrtR may serve as an autoregulator (Fig. 1C, D ). The SsNrtR exhibits high sequence conservation with other orthologs in streptococci and other distantly related organisms (Fig. 2A, B ). The pnuC gene encodes a putative RNam transporter (25), whereas the nadR gene product NadR is predicted to contain 2 enzymatic activities. Its nucleoside kinase activity (NmR‐K; EC 2.7.1.22) phosphorylates NmR to form NMN, whereas its adenylyltransferase activity (NMN‐AT; E.C. 2.7.7.1) catalyzes the conversion of NMN to NAD⁺ (Fig. 1D ) (26, 27). Overall, the pnuC and nadR genes encode all the necessary catalytic functions required for the RNam‐NAD⁺ salvage pathway (Fig. 1D ) (25).

Figure 1.

A working model for the regulation of the NAD⁺ salvage pathway by NrtR in SS2. A) Genetic context of nrtR and its neighboring loci in certain Streptococcus species. B) The sequence of the conserved NrtR‐recognizable motif. C) Sequence analyses of the promoter of the operon nadR‐pnuC‐nrtR. D) Graphic representation of salvage pathway of NAD⁺ cofactor and its regulation by NrtR in SS2. Putative NrtR‐binding sites are marked by red ovals (A). Identical residues are shown as white letters in a red background, similar residues are red letters, and variable residues are black letters (B). The predicted NrtR‐binding site is given in red and underlined letters. S denotes the transcription initiation site (C). D) In this salvage pathway, NAD⁺ is synthesized from the exogenous RNam precursor delivered by the RNam transporter (PnuC) via consecutive reactions catalyzed by 2 separate domains of NadR, Nam mononucleotide adenyl‐transferase (NadR^AT), generating NMN, and Nam riboside kinase (NadR^K), giving NAD.

Figure 2.

Bioinformatic analyses of SsNrtR. A) Schematic representative for SsNrtR. B) Multiple sequence alignment of SsNrtR with the homologs from other Streptococcus species. The SsNrtR protein contains 2 domains: the N‐terminal Nudix domain (in light blue) and the C‐terminal wHTH DNA‐binding domain (in green) (A). The protein sequences of NrtR homologs were aligned using ClustalW2. The resultant photograph was generated using the ESPript 2.2 program. Identical residues are referred to with white letters in a red background. Similar residues are illustrated in black letters with a yellow background. Varied residues are in black letters, and dots represent gaps. Amino acids indicated by the blue arrows are the conservative Nudix signature, GX5EX7REUXEEXGU (in which U is a hydrophobic residue and X is any residue). Amino acids indicated by the green arrows are critical residues of NrtR for DNA binding. The amino acid sequences are from S. thermophilus (WP_011226910.1), S. suis (WP_012027854.1), S. mitis (YP_003447138.1), S. gordonii (WP_012130973.1), and S. sanguinis (YP_001036108.1), respectively.

Characterization of the SsNrtR protein

To further explore the putative function of SsNrtR, a recombinant hexa‐histidine–tagged SsNrtR fusion protein was purified from E. coli BL21 (DE3) harboring pET28b‐nrtR using Nickel NTA agarose affinity chromatography followed by Superdex 200 gel filtration chromatography (Fig. 3A ). The purified protein was then subjected to 12% SDS‐PAGE to test its homogeneity (Fig. 3A ). The recombinant protein preparation was relatively pure and exhibited an intense band of ∼31 kDa, which corresponded to its predicted MW (Fig. 3A ). Sequence analysis suggested that SsNrtR consists of an N‐terminal domain homologous to ADPR pyrophosphatases of the Nudix family and a C‐terminal winged helix‐turn‐helix (wHTH) domain (Figs. 2A and 3B ) involved in DNA binding (28, 29). Earlier studies have indicated that the NrtR family of regulators usually form homodimers (30). In our chemical cross‐linking assay with the cross‐linker ethylene glycol bis (succinimidyl succinate) (EGS), we observed that SsNrtR forms homodimers and behaves as a prevalent form in the presence of increasing concentrations of EGS (Fig. 3C ) in addition to oligomer or multimer formation.

Figure 3.

Biochemical characterization of the SsNrtR protein. A) FPLC profile of the purified NrtR protein of SS2. B) Ribbon illustration for the modeled structure of the SsNrtR protein. C) EGS‐based chemical cross‐linking assays suggest that SsNrtR forms a dimer. SsNrtR denotes SsNrtR. The protein sample was separated by gel filtration with Superdex 200 (GE Healthcare). The expected peak of purified SsNrtR was eluted at the position of ∼9 and ∼16 ml (indicated with arrows). The inset gel denotes the SDS‐PAGE profile of SsNrtR. The apparent MW of recombinant SsNrtR is about 31 kDa. The SsNrtR protein contains 2 domains: the N‐terminal Nudix domain (in aquamarine) and the C‐terminal wHTH DNA‐binding domain (in pink). The predicted secondary structure is shown on the top. OD280, optical density at 280 nm; mAu, milli‐absorbance units; α, α‐helix; β, β‐sheet; M, protein standard marker (Bio‐Rad, Hercules, CA, USA). Glutathione S‐transferase (GST) is positive control, and EGS is the chemical cross‐linker. The concentrations of EGS vary from 0 to 10 µM, and “−” denotes no addition of EGS.

Binding of NrtR to nadR and its reversal by the ligand of ADPR

The position of the putative NrtR‐binding site, which overlaps with the predicted promoter elements of the nadR‐punC‐nrtR operon (Fig. 1C ), strongly suggests that SsNrtR serves as a repressor (Fig. 4A ). This prediction was subsequently demonstrated by genetic manipulation (Fig. 5A, B and Supplemental Fig. S1). Furthermore, its N‐terminal Nudix domain (Figs. 2A and 3B ) is predicted to bind an effector molecule that weakens Protein‐DNA interactions, leading to the derepression of target genes (Fig. 4A ) (10). We compared NrtR‐binding sites in streptococcal genomes using nucleotide BLAST (BLASTN) (Fig. 1A, B ). As a result, 4 Streptococcus species besides S. suis were found to contain 1 or 2 NrtR‐binding sites in their genomes (Fig. 1A, B ). These sites were located in the promoter regions upstream of nadR without exception (Fig. 1A ), whereas 2 or more binding sites were found in various other bacteria upstream of other NAD⁺ synthesis‐related genes or operons, such as nadM, nadE, pncB, nadABC, prs‐nadV (http://regprecise.lbl.gov/RegPrecise). Putative NrtR DNA‐binding signatures were highly conserved among streptococci and to a lesser extent in other species (Fig. 1B ).

Figure 4.

Binding of NrtR to cognate nadR and its reversal by the addition of ADPR. A) A model for NrtR‐mediated transcriptional regulation of an operon nadR‐pnuC‐nrtR. B) NrtR binds to the nadR probe in a dose‐dependent manner. C) Addition of ADPR reverses the binding of NrtR to cognate the nadR promoter. Proteins were added at different concentrations denoted by a triangle (0.0, 0.1, 0.5, 1.0, and 2.0 mM). The 56 bp of DNA probe encompassing the nadR promoter region was obtained by annealing the primers nadR‐F and nadR‐R (Supplemental Table S2). The arrow indicates the shifted DNA‐protein complex. A representative result from 3 independent experiments is shown (B). Functional impairment of interaction between NrtR and the nadR promoter by ADPR. The NrtR‐DNA interaction was reversed by ADPR in a dose‐dependent manner (C).

Figure 5.

In vivo evidence that NrtR acts as a repressor for the nadR‐pnuC‐nrtR operon. A) The SsNrtR represses expression of the nadR‐pnuC‐nrtR operon in the midlog phase (OD₆₀₀ = 0.6). B) Expression of the nadR‐pnuC‐nrtR operon is up‐regulated in the ΔnrtR mutant in the stationary phase (OD₆₀₀ = 1.2). Use of β‐gal assays to elucidate the elevated level of nadR‐lacZ transcriptional fusion in S. suis. The promoter activity of nadR‐pnuC‐nrtR was tested in WT, nrtR deleted strain (ΔnrtR), and the complementary strain (CΔnrtR), respectively. The β‐gal activity was assayed through 3 independent experiments and are expressed in means ± sd. The midlogarithmic phase culture is determined when the OD₆₀₀ (optical density at 280 nm) reaches 0.6 (A), and the stationary phase culture is defined upon its OD₆₀₀ of 1.2 (B). The spectrophotometer UV 8100C (Beijing Labtech Instruments, Beijing, China) was used. The data were given as means ± sd. Significant differences were determined by 1‐way ANOVA followed by Tukey's test. *P < 0.05.

We employed EMSA to test the ability of purified SsNrtR to bind the DNA probe (designated nadR, a 56 bp fragment containing the candidate 21 bp NrtR signal). As expected, SsNrtR efficiently bound the nadR probe in a dose‐dependent manner (Fig. 4B ). Interestingly, we also noticed that SsNrtR‐DNA interaction was interrupted by the addition of ADPR in a dose‐dependent manner (Fig. 4C ). At the level of 300 µM ADPR, NrtR completely lost its ability to bind DNA in our gel filtration assays (Fig. 4C ). It seems likely that ADPR, an intermediate of glycohydrolytic NAD⁺ degradation, acts as an effector molecule impairing the NrtR‐DNA interaction (Fig. 4A, C ) (10). It can be converted to Rib‐P by Nudix hydrolases and further recycled to generate NAD⁺ via phosphoribosyl pyrophosphate (PRPP) formation (10). Because the binding of ADPR to the NrtR Nudix domain promotes dissociation of NrtR‐DNA complexes, ADPR is believed to act as an anti‐repressor of the NAD⁺ salvage operon. Presumably, the accumulation of ADPR inside the cell functions as a signal to replenish the NAD⁺ cofactor pool (Fig. 4A ) (10).

Physiologic role of SsNrtR

To test if SsNrtR has a physiologic role in regulatory transcription of the nadR‐punC‐nrtR operon, we created a ΔnrtR mutant strain using allelic exchange (Supplemental Fig. S1A, B) as well as a complemented strain CΔnrtR (Fig. 5 and Supplemental Fig. S1C). This genetic alteration was confirmed with multiplex PCR as well as direct sequencing of the PCR amplicon (Supplemental Fig. S1B). The disruption of nrtR does not influence bacterial growth (Supplemental Fig. S1C) and its hemolytic activity (Supplemental Fig. S1D). In addition, electron microscopy visualization of nrtR knockout strain did not show any clear morphologic alteration in S. suis (Supplemental Fig. S1E).

The S. suis nadR promoter was fused to the lacZYA operon, giving a LacZ reporter plasmid named pVA838‐FnadR (Supplemental Table S1). Then, this plasmid was introduced into the following strains: the WT, ΔnrtR, and CΔnrtR. Compared with the WT S. suis 05ZYH33, nadR‐lacZ expression was increased in the ΔnrtR background (Fig. 5A, B ), whereas its expression mirrored WT levels in the complemented CΔnrtR strain in both midlogarithmic phase (Fig. 5A ) and stationary phase (Fig. 5B ). Subsequently, direct measurement revealed that level of cytosolic NAD⁺ (and NADH) is increased 2–3‐fold in the ΔnrtR mutant when compared with its parental strain (Fig. 6A, B ). These results are consistent with our EMSA data (Fig. 4A–C ) and provided in vivo evidence that NrtR has physiologic role in the transcription of the nadR‐punC‐nrtR operon (Fig. 5A, B ) as well as NAD⁺/NADH homeostasis in SS2 (Fig. 6A, B ).

Figure 6.

Control of intracellular levels of NAD⁺ and NADH by NrtR in SS2. A) The altered level of intracellular NAD⁺ in the ΔnrtR mutant. B) The SsNrtR negatively regulates the accumulation of cytosolic NADH. It is presented as means ± sd. NAD⁺ is presented as means ± sd. The P value was measured using 1‐way ANOVA along with Tukey's test. *P < 0.05, **P < 0.01, ***P < 0.001.

Residues of NrtR critical for DNA binding

Structural architecture of NrtR from the model bacterium Shewanella oneidensis has been previously described in ref. (30) (Fig. 7A ). The putative residues implicated in the crosstalking of SsNrtR to cognate DNA were inferred from a multiple sequence alignment (Fig. 2B ) and structural modeling (Fig. 7B ). To assess their function in SsNrtR, site‐directed mutagenesis was used to generate 3 mutants of SsNrtR (namely H223A, K246A, and R248A). The mutant versions of NrtR protein were purified to homogeneity (Fig. 7C ) and then tested in our EMSA assays (Fig. 7D–F ). Gel shift assays showed that the H223A mutation of NrtR did not have an obvious effect on its DNA‐binding activity (Fig. 7D ). On the other hand, the point mutant K246A of NrtR loses, in part, its affinity for DNA binding (Fig. 7E ) and easily precipitates at high protein concentration. However, the replacement of arginine 248 in NrtR with alanine dramatically diminished its affinity for DNA binding (Fig. 7F ).

Figure 7.

Structural and functional characterization of SsNrtR. A) The use of structural modeling to probe the SsNrtR binding to its cognate DNA. B) The enlarged view of the interface between SsNrtR and its cognate DNA. The putative residues of SsNrtR with the involvement of DNA interaction refer to H223, K246, and R248, respectively. C) SDS‐PAGE (15%) profile of the SsNrtR protein and its mutants. D) The mutant protein of SsNrtR (H233A) remains DNA‐binding activity. E) The mutant protein of SsNrtR (K246A) easily precipitates in the in vitro EMSA assay. F) Gel filtration revealed that the mutant protein of SsNrtR (R248A) exhibits the undetectable ability to bind cognate nadR probes. G) The mutation of K92E in SsNrtR does not influence its crosstalking to the cognate nadR. The overall structure was given with PyMol (A, B). An EMSA assay was used to probe the interplay between SsNrtR and the nadR probe. Varied level of SsNrtR protein (0.1, 0.5, 1.0, and 2.0 µM) was denoted by a triangle. The minus (−) refers to no addition of protein. The nadR probe is a DNA fragment of 56‐bp long, which is generated through annealing the 2 complementary DNA oligos nadR‐F and nadR‐R (Supplemental Table S2). The arrow indicates the shifted DNA‐protein complex. A representative result is given from 3 independent experiments (D–G).

Enzymatic characterization of NrtR

The N‐terminal domain of NrtR is typical of Nudix hydrolases (Fig. 3B ), which hydrolyze a pyrophosphate bond in a wide range of organic pyrophosphates with varying degrees of substrate specificity (10). Substrates for Nudix hydrolases include nucleotide sugars (such as ADPR; Fig. 8A ) nucleoside di‐ and triphosphates (Fig. 8B ) and dinucleoside polyphosphates. An HPLC‐based assay was used to detect Nudix hydrolase activity of NrtR, but there was no detectable activity when using the nucleoside diphosphate derivatives ADPR (Fig. 8C ), NADH (Fig. 8E ), or NADPH (Fig. 8G ) as substrates. This result suggested that NrtR lost its Nudix activity and generally agrees with an earlier prediction by Rodionov et al. (10). Functional loss of certain members of this family in their Nudix catalytic activity might occur because of the lack of a strictly‐conserved Nudix hydrolase signature sequence, GX₅EX₇_ REUXEEXGU (where U is a hydrophobic residue and X is any residue) (10, 31). We hypothesized that the WT of SsNrtR might have lost its Nudix hydrolase activity as a result of mutation in its Nudix hydrolase signature sequence while retaining its DNA‐binding activity.

Figure 8.

HPLC‐based assays for enzymatic activity of NrtR and its mutant K92E. A) Chemical reaction proposed for the hydrolysis of ADPR by NrtR to yield AMP and Nu‐5‐P. B) Chemical reaction shows that NADH is hydrolyzed to yield NMN and AMP (AMP‐P). C–F) HPLC assays suggested that the WT of NrtR cannot hydrolyze ADPR (C), but the point‐mutation (K92E) of NrtR can restore its enzymatic activity (D). HPLC analyses revealed that the WT of NrtR has no detectable activity on the NADH substrate (E), whereas the revertant (K92E) of NrtR displays the ability of hydrolysis (F). G, H) HPLC evidence that a single substitution (K92E) of NrtR rescues the loss of enzymatic activity. Nu‐5‐P, nucleophile.

The subsequent question we asked is whether or not a revertant of substitution in the Nudix motif of SsNrtR regains the Nudix hydrolase activity. Using site‐directed mutagenesis, we generated a K92E mutant of SsNrtR to restore the consensus Nudix motif. The K92E mutant protein was purified to homogeneity (Fig. 7C ) and found to retain a complete activity of binding cognate DNA (Fig. 7G ). Intriguingly, this substitution of K92E resulted in a significant increase in NrtR hydrolysis of ADPR (Fig. 8D ), NADH (Fig. 8F ), and NADPH (Fig. 8H ). This is consistent with an earlier description of the Nudix hydrolase Slr1690 from Cyanobacterium species (32). Thus, it is plausible that the presence of an intact Nudix signature sequence is essential for the catalytic activity of a typical member of bifunctional NrtR. Moreover, our data validated that the N‐terminal Nudix domain and C‐terminal wHTH binding domain are functionally independent. Thereafter, our data suggest that the loss of Nudix activity in SsNrtR is probably irrelevant for S. suis survival, whereas its transcriptional regulatory role in NAD⁺ salvage is beneficial for its survival.

NrtR is associated with S. suis virulence

Given the numerous roles for NrtR in bacteria, we assayed a variety of potential ΔnrtR phenotypes. We found no obvious changes in growth rate (Supplemental Fig. S1C), hemolysis (Supplemental Fig. S1D), or morphology (Supplemental Fig. S1E). We resorted to mining the sequences of available S. suis clinical isolates in GenBank for nrtR. The results suggested that nrtR is encoded in a strain‐specific manner (i.e., not a component of the core genome) (Supplemental Table S3) and that Nudix catalysis is not a conserved component of S. suis NAD⁺ metabolism. Thus, we were curious whether SsNrtR might play a role in virulence. Intriguingly, S. suis genome comparisons did reveal an obvious correlation between the presence of nrtR and the reported virulence of the strain, especially among SS2 strains (Supplemental Table S3). Consequently, an experimental mouse infection model was tested to assess the potential link of nrtR to bacterial virulence. All specific‐pathogen‐free (SPF) mice inoculated intravenously with the WT strain (positive control) developed serious clinical symptoms, such as high fever, swollen joints, shivering, and even central nervous system failure, within 24 h. Lethality was 100% in <40 h postinfection (Fig. 9A ). In contrast, mice inoculated with the ΔnrtR strain survived longer overall, and 3 mice even survived the entire 4‐d observation period (Fig. 9A ). Functional complementation of plasmid‐borne nrtR (pVA838‐nrtR_ss, in Supplemental Table S1) into the ΔnrtR mutant can restore bacterial virulence to a lesser extent (Fig. 9A ). As expected, the SS2 of interest (WT, ΔnrtR, and CΔnrtR) can be consistently recovered from the infected mice (Supplemental Fig. S2), which is validated by both bacterial hemolytic assays (Supplemental Fig. S2A) and Gram stains (Supplemental Fig. S2B). Next, we assayed the biofilm formation capacity for the WT and ΔnrtR mutant SS2 strains (Fig. 9B ). The result shows a reduced capacity of biofilm formation in the ΔnrtR mutant compared with WT and the complemented strain CΔnrtR. In addition, we noted that the ΔnrtR mutant exhibits obvious resistance to a variety of antibiotics (Supplemental Table S4). Taken together, our results indicated that SsNrtR is relevant to successful infection of SS2.

Figure 9.

The involvement of NrtR in S. suis virulence. A) The removal of nrtR attenuates S. suis infectivity in mice. In the experimental infection, 4 groups of BALB/c mice (10 per group) were challenged with different bacterial strains. It included the WT (positive control), the nrtR deletion mutant (ΔnrtR), and the complementary strain CΔnrtR The negative control denotes THB medium. B) Reduction of biofilm formation by the nrtR inactivation in SS2. The spectrophotometer UV 8100C (Beijing Labtech Instruments) was applied to measure optical density at 570 nm (OD570) here. The data are presented as means ± sd. The P value was fixed using 1‐way ANOVA and Tukey's test. **P < 0.01, ***P < 0.001.

Evolutionary relationships between NrtR and its homologs

The evolutionary relationship between NrtR and its homologs was inferred using phylogenetic analysis. NrtR is seen to be present in many Streptococcus species, with its homologs sharing >60% aa sequence identity. Their next closest relatives are found predominantly in Enterococcus species and archaebacteria, which share only 30–40% aa sequence identity. Consequently, NrtR from S. suis and its close homologs from other Streptococcus species cluster tightly within a distinct clade (subclade I; Fig. 10 ). The homologs from Enterococcus species along with a few other species form their own group (subclade II), which is relatively closer to subclade I than the group formed exclusively of archaebacteria (subclade III; Fig. 10). Therefore, it is possible to believe that ADPR pyrophosphatase might be an ancestor for the NrtR‐like transcriptional regulator.

Figure 10.

Phylogeny of NrtR and its related Nudix hydrolases. An unrooted phylogenetic tree of NrtR and its close homologs is presented as a radial phylogram with 3 distinct subclades highlighted in assorted colors. Subclade I is primarily composed of homologs from the Streptococcus species (in pink), whereas a majority of the homologs in subclade II are from the Enterococcus species (in blue). Subclade III is distinct and exclusively contains members from archaebacteria (in yellow). NrtR is highlighted in bold red font. The evolutionary history of NrtR was inferred using the maximum likelihood method. The trees presented here have been inferred from 1000 bootstrap replicates using an Log Gamma (LG) amino acid substitution model. The percentages of replicate trees in which the associated taxa are clustered in the bootstrap test (1000 replicates) are shown next to the branches. A discrete γ distribution was used to model evolutionary rate differences among sites with some evolutionarily invariable sites. Protein accession numbers of individual members are indicated in this figure.

DISCUSSION

The maintenance of NAD⁺ homeostasis is critical for the 3 domains of life. It seems likely that at least 3 types of regulatory mechanisms have evolved to modulate the synthesis (or salvage) of NAD⁺ cofactor in response to its fluctuation. First, the NadR was initially identified as a negative regulator for NAD⁺ synthesis in S. enterica (2). Second, NiaR (earlier called YrxA) of B. subtilis is a niacin‐responsive repressor of de novo NAD⁺ synthesis (33). Third, comparative genomics suggests that the Nudix domain–containing regulatory protein (designated NrtR here) represents a family of previously‐uncharacterized regulators involved in NAD⁺ homeostasis (10). However, the functional definition of NrtR lacked further experimental evidence.

In this study, we aimed to characterize the role of NrtR in the NAD⁺ metabolism salvage pathway in the zoonotic pathogen SS2 (Fig. 1). The most likely pathway for NAD⁺ biosynthesis in SS2 consists of: 1) transport of RNam across the periplasmic membrane, 2) phosphorylation to NMN, and 3) formation of NAD⁺ by NMN adenylation (Fig. 1). Consistent with this prediction, our results demonstrate that NrtR likely acts as an autoregulatory repressor of the nadR‐punC‐nrtR operon. Furthermore, we anticipate a classic feedback regulation exists for the operon because the product of NAD⁺‐consuming enzymes (ADPR) reverses the NrtR‐mediated repression of the nadR‐punC‐nrtR operon (Figs. 1 and 4).

The paradigm member of NrtR family is believed to possess the Nudix pyrophosphatase enzymatic activity in addition to its DNA‐binding ability (10), and thus it can be classified as a moonlighting protein. A moonlighting protein refers to a single protein with dual independent roles that are not due to gene fusion (35). However, the Nudix hydrolysis activity was lacking in the SsNrtR (Fig. 8), and we were able to restore it using a K92E revertant in the Nudix consensus sequence (Figs. 2 and 8). Given the fact that the redundant Nudix domain‐containing proteins (12 total) are encoded on the chromosome of S. suis 05ZYH33, it is possible that the loss of catalytic activity might render SsNrtR more efficient or energy saving to retain its regulatory function. Because that phylogenic tree unveils clearly that ADPR pyrophosphatase appears in the subclade of archaebacteria that neighbors the subclade I, NrtR Nudix hydrolase in Streptococcus (Fig. 10), we hypothesized that SsNrtR might be an evolutionarily remnant regulator from its archaebacterial ancestor.

In this case, SsNrtR yielded a mutant catalytic domain as a consequence of evolution under unknown selection pressure. Because of the nonessentiality for bacterial viability, SsNrtR as an evolutionary relic is anticipated to benefit from successful infection within its infected hosts. As expected, this was further demonstrated by genetic removal of nrtR. The lack of growth phenotype from the ΔnrtR mutant as well as the lack of nrtR conservation among S. suis strains led us to speculate that SsNrtR could be involved in virulence, and indeed we observed a marked reduction in the pathogenicity of the ΔnrtR mutant using a mouse model of infection. Our conclusion was, in part, verified by a recent observation by Li et al. (34) of attenuation of bacterial virulence by the removal of pnuC. Therefore, this further implied that the salvage pathway encoded by the nadR‐pnuC‐nrtR operon is not the primary mechanism for NAD⁺ generation in S. suis. Rather, this additional NAD⁺ generation pathway in SS2 is likely an evolutionary acquisition supporting its role as a pathogen. The primary mechanism used for endogenous NAD⁺ generation in S. suis still awaits further exploration.

AUTHOR CONTRIBUTIONS

Y. Feng designed research; Q. Wang, N. Lou, and Y. Feng performed research; Q. Wang, B. H. Hassan, N. Lou, J. Merritt, and Y. Feng analyzed data; B. H. Hassan, J. Merritt, and Y. Feng contributed reagents and analytic tools; and Q. Wang, B. H. Hassan, J. Merritt, and Y. Feng wrote the manuscript.

Supporting information

Supplementary Material

Wang, Q. , Hassan, B. H. , Lou, N. , Merritt, J. , Feng, Y. Functional definition of NrtR, a remnant regulator of NAD⁺ homeostasis in the zoonotic pathogen Streptococcus suis . FASEB J. 33, 6055–6068 (2019). www.fasebj.org