Interactions of the Metalloregulatory Protein SloR from Streptococcus mutans with Its Metal Ion Effectors and DNA Binding Site

ABSTRACT

Streptococcus mutans is the causative agent of dental caries, a significant concern for human health, and therefore an attractive target for therapeutics development. Previous work in our laboratory has identified a homodimeric, manganese-dependent repressor protein, SloR, as an important regulator of cariogenesis and has used site-directed mutagenesis to map functions to specific regions of the protein. Here we extend those studies to better understand the structural interaction between SloR and its operator and its effector metal ions. The results of DNase I assays indicate that SloR protects a 42-bp region of DNA that overlaps the sloABC promoter on the S. mutans UA159 chromosome, while electrophoretic mobility shift and solution binding assays indicate that each of two SloR dimers binds to this region. Real-time semiquantitative reverse transcriptase PCR (real-time semi-qRT-PCR) experiments were used to determine the individual base pairs that contribute to SloR-DNA binding specificity. Solution studies indicate that Mn²⁺ is better than Zn²⁺ at specifically activating SloR to bind DNA, and yet the 2.8-Å resolved crystal structure of SloR bound to Zn²⁺ provides insight into the means by which selective activation by Mn²⁺ may be achieved and into how SloR may form specific interactions with its operator. Taken together, these experimental observations are significant because they can inform rational drug design aimed at alleviating and/or preventing S. mutans-induced caries formation.

IMPORTANCE This report focuses on investigating the SloR protein as a regulator of essential metal ion transport and virulence gene expression in the oral pathogen Streptococcus mutans and on revealing the details of SloR binding to its metal ion effectors and binding to DNA that together facilitate this expression. We used molecular and biochemical approaches to characterize the interaction of SloR with Mn²⁺ and with its SloR recognition element to gain a clearer picture of the regulatory networks that optimize SloR-mediated metal ion homeostasis and virulence gene expression in S. mutans. These experiments can have a significant impact on caries treatment and/or prevention by revealing the S. mutans SloR-DNA binding interface as an appropriate target for the development of novel therapeutic interventions.

INTRODUCTION

Cariogenesis is a disease that derives from interactions involving the human dentition, an acidogenic microflora, including Streptococcus mutans, and dietary factors. Specifically, S. mutans, which is among the most cariogenic of the oral microbes, assumes an obligate biofilm lifestyle upon colonizing the tooth surface during the first year of life, shortly after tooth eruption. Commonly found in the human oral cavity as the most abundant species in the dental plaque biofilm, S. mutans metabolizes exogenous dietary carbohydrates to generate energy for itself via fermentation and releases acid as a metabolic byproduct. The buffering capacity of saliva combined with appropriate oral hygiene practices can maintain the plaque pH at near neutrality and so curtail the demineralization process that marks the onset of tooth decay. However, the prevalence of dietary carbohydrate in the Western diet continues to provide S. mutans and other acidogenic and aciduric oral microbes with a competitive advantage, encouraging their proliferation and shifting the ecology of the plaque microflora from one that was associated with oral health to one that is more prone to disease (1).

In all organisms, metal ion homeostasis is mediated by mechanisms of transport and sequestration that ensure sufficient levels of essential micronutrients, such as iron and manganese, while avoiding the toxic consequences of excess cellular concentrations. In bacteria, metal ion-dependent transcriptional regulators are chiefly responsible for modulating homeostasis and often play important roles in the ability of pathogens to survive in a human host, making them especially attractive therapeutic targets. Iron and manganese play a central role in a variety of cellular processes as well as in the colonization, survival, and proliferation of bacteria living on or within a mammalian host, with the success of the bacterial pathogen hinging upon its ability to overcome the metal ion deprivation in the host that characterizes so-called “nutritional immunity.” In the context of the oral cavity, nutritional immunity prevails during periods of famine over a 24-h day. During these nonmealtimes, availability of metal ions is a limiting factor, and S. mutans must therefore scavenge for these essential micronutrients by turning on its metal ion uptake machinery. During a mealtime, however, the introduction of foodstuffs into the mouth renders metal ions transiently plentiful, and in response, S. mutans downregulates its metal ion transporters to avoid excessive metal ion uptake and its associated toxicity.

Previous work in our laboratory revealed expression of a multitude of S. mutans virulence genes that is coordinated with conditions of metal ion deprivation and repletion (2–4). On the basis of these observations, we hypothesize that expression of S. mutans virulence attributes, including the genes that encode metal ion uptake and transport, is responsive to metal ion availability, which fluctuates during periods of feast and famine in the oral cavity. Importantly, we and others had previously identified a 25-kDa metalloregulatory protein called SloR that modulates the expression of a sloABC manganese transport operon on the S. mutans UA159 chromosome (5, 6). We also recognized that SloR-mediated gene expression includes modulation of additional virulence attributes that allow S. mutans to adhere to the dentition, generate acid, and tolerate acid and oxidative stress (2, 3). Continued investigations of the sloABC locus revealed SloR binding to a palindromic SloR recognition element (SRE) that is located upstream of and in a location proximal to the sloABC promoter (4). We noted that such SloR-SRE binding is manganese dependent and that expression of the sloABC operon is derepressed in UA159 cells grown in a manganese-deficient medium and in a UA159-derived SloR insertion-deletion mutant called GMS584. Importantly, monoinfection of a germfree rat model with SloR-deficient S. mutans resulted in a hypercariogenic phenotype (6), presumably the result of virulence gene derepression that is concomitant with derepression of the sloABC metal ion transporter.

SloR is a member of the DtxR family of metal-dependent regulators. The namesake of the family, DtxR, is an iron-dependent transcriptional repressor in Corynebacterium diphtheriae that regulates the tox gene, which encodes the diphtheria toxin, as well as several operons involved in iron uptake (7, 8). Each subunit of DtxR is composed of three domains: an N-terminal DNA binding domain, a central dimerization domain promoting quaternary structure, and a C-terminal domain that has a fold similar to the Src homology 3 (SH3) domain and that is related to FeoA, a protein involved in iron uptake in bacteria (9). Two metal ions are bound per subunit of DtxR, and two dimers bind overlapping 21-bp recognition sites, spanning 33 bp in total (10, 11). As a member of the DtxR family of metalloregulators, SloR is predicted to interact with DNA as a homodimer, and previous work has revealed that residues at positions equivalent to those involved in metal ion and DNA binding in DtxR are likewise critical for SloR function (12).

The goal of the present study was to further advance our understanding of the interactions made by SloR with its DNA recognition element and metal ion effector(s). To that end, we have investigated the nature of the complex that is formed by SloR with its operator using electrophoretic mobility shift assays (EMSAs), solution binding assays monitored by fluorescence anisotropy, real-time semiquantitative reverse transcriptase PCR (real-time semi-qRT-PCR) experiments, and DNase I protection assays. The crystal structure of SloR bound to the noncognate metal ion, Zn²⁺, provides an atom-level view of the interactions between the repressor and metal ions, and our findings are supported by the results of activation studies that monitored the affinity of SloR for the SRE in the presence of Mn²⁺ and Zn²⁺.

MATERIALS AND METHODS

Bacterial strains, plasmids, and primers.

The bacterial strains and plasmids used in this study are described in Table 1. Details of the oligonucleotide primers designed with the assistance of MacVector 12.0 software and purchased from Sigma-Aldrich (St. Louis, MO) are presented in Table 2, and the primers were designed on the basis of the genome sequence of Streptococcus mutans UA159 (serotype c).

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristicsa	Source or reference
Strains
E. coli BL21(DE3)	fhA2 [lon] ompT gal(λ) (DE3) [dcm] ΔhsdS	NE BioLabs
E. coli TOP10	F− mcrA Δ(mrr-hsdRMS-mcrBC) ϕ80lacZΔM15 ΔlacX74 recA1 araD139 Δ(ara leu)7697 galU galK rpsL (Strr) endA nupG	Life Technologies
S. mutans UA159	Wild type, serotype c	ATCC 700610
S. mutans GMS584	UA159 derived, sloR deficient, Emr	2
S. mutans GMS602	UA159 derived, contains IFDC2 cassette, Emr, 4-Cl-Phes	This study
S. mutans GMS603	UA159 derived, A-to-G transition at SRE position 3	This study
S. mutans GMS604	UA159 derived, A-to-G transition at SRE position 4	This study
S. mutans GMS605	UA159 derived, T-to-C transition at SRE position 5	This study
S. mutans GMS607	UA159 derived, A-to-G transition at SRE position 7	This study
S. mutans GMS609	UA159 derived, C-to-T transition at SRE position 9	This study
S. mutans GMS613	UA159 derived, A-to-G transition at SRE position 13	This study
S. mutans GMS614	UA159 derived, C-to-T transition at SRE position 14	This study
S. mutans GMS615	UA159 derived, T-to-C transition at SRE position 15	This study
S. mutans GMS616	UA159 derived, T-to-C transition at SRE position 16	This study
S. mutans GMS619	UA159 derived, T-to-C transition at SRE position 19	This study
S. mutans GMS620	UA159 derived, T-to-C transition at SRE position 20	This study
S. mutans GMS601u	UA159 derived, A-to-G transition 1 bp upstream of SRE	This study
S. mutans GMS602u	UA159 derived, T-to-C transition 2 bp upstream of SRE	This study
S. mutans GMS603u	UA159 derived, A-to-G transition 3 bp upstream of SRE	This study
S. mutans GMS604u	UA159 derived, T-to-C transition 4 bp upstream of SRE	This study
S. mutans GMS605u	UA159 derived, A-to-G transition 5 bp upstream of SRE	This study
S. mutans GMS606u	UA159 derived, A-to-G transition 6 bp upstream of SRE	This study
S. mutans GMS607u	UA159 derived, T-to-C transition 7 bp upstream of SRE	This study
S. mutans GMS608u	UA159 derived, C-to-T transition 8 bp upstream of SRE	This study
S. mutans GMS609u	UA159 derived, T-to-C transition 9 bp upstream of SRE	This study
S. mutans GMS610u	UA159 derived, A-to-G transition 10 bp upstream of SRE	This study
S. mutans GMS611u	UA159 derived, C-to-T transition 11 bp upstream of SRE	This study
S. mutans GMS612u	UA159 derived, A-to-G transition 12 bp upstream of SRE	This study
S. mutans GMS601d	UA159 derived, T-to-C transition 1 bp downstream of SRE	This study
S. mutans GMS602d	UA159 derived, A-to-G transition 2 bp downstream of SRE	This study
S. mutans GMS603d	UA159 derived, T-to-C transition 3 bp downstream of SRE	This study
S. mutans GMS604d	UA159 derived, A-to-G transition 4 bp downstream of SRE	This study
S. mutans GMS605d	UA159 derived, T-to-C transition 5 bp downstream of SRE	This study
S. mutans GMS606d	UA159 derived, T-to-C transition 6 bp downstream of SRE	This study
S. mutans GMS607d	UA159 derived, A-to-G transition 7 bp downstream of SRE	This study
S. mutans GMS608d	UA159 derived, G-to-A transition 8 bp downstream of SRE	This study
Plasmids
pES03	pET101/D-TOPO (Invitrogen) derived, expressing wild-type SloR, Ampr	12
pGEM-T Easy	TA cloning vector, Ampr	Promega
pSMT3	Champion pET SUMO expression plasmid, Knr	13
pSMT3_SloR	pSMT3 derived, harbors sloR ORF cloned in frame for cotranslation with N-terminal His tag, Knr	This study

^aEm^r, erythromycin resistant; Amp^r, ampicillin resistant; Kn^r, kanamycin resistant; 4-Cl-Phe^s, 4-chloro-phenylalanine sensitive.

TABLE 2.

List of primers used in this study

Primer application	Primer/oligonucleotide name	Nucleotide sequence (5′ to 3′)a	Annealing temp (°C)b	Amplicon size (bp)b
Cloning	slorF	GAGGCTggatccACACCTAATAAAGAAGATTACCTTAAAA	67	675
	slorR	GAGGCTaagcttTTAGTAGGCTTTCTTTTCAATGTATATTTGACTGGC
	S1MF	GAACAGATTGGTggaatgACACCTAATAAAGAAG	71
	S1MR	CTTCTTTATTAGGTGTCATTCCACCAATCTGTTC
	M13-F	TGTAAAACGACGGCCAGT	50
	M13-R	AGGAAACAGCTATGACCAT
Overlap extension PCR	UpF	TGGATGCCTATTTGATTGTTG	51.6	1,407
	DnR	AAGCCTTGTATAAATTCTGTAATCAT
	UpR.d3c	AAATTAAGTCAAGTTAATcTTTATATTAGAT	51.2	625
	DnF.d3d	ATCTAATATAAAgATTAACTTGACTTAATTT	50.5	812
	UpR.d4c	AAATTAAGTCAAGTTAAcTTTTATATTAGAT	51.2	625
	DnF.d4d	ATCTAATATAAAAgTTAACTTGACTTAATTT	50.5	812
	UpR.d7c	AAATTAAGTCAAGTcAATTTTTATATTAGAT	51.2	625
	DnF.d7d	ATCTAATATAAAAATTgACTTGACTTAATTT	50.5	812
	UpR.d13c	AAATTAAGcCAAGTTAATTTTTATATTAGAT	51.2	625
	DnF.d13d	ATCTAATATAAAAATTAACTTGgCTTAATTT	50.5	812
	UpR.d14c	AAATTAAaTCAAGTTAATTTTTATATTAGAT	51.2	625
	DnF.d14d	ATCTAATATAAAAATTAACTTGAtTTAATTT	50.5	812
	UpR.d15c	GTAATATAACCTAATATAAAAATTAgGTCAAG	51.2	644
	DnF.d15d	CTTGACcTAATTTTTATATTAGGTTATATTAC	50.7	794
	UpR.d16c	GTAATATAACCTAATATAAAAATTgAGTCAAG	51.2	644
	DnF.d16d	CTTGACTcAATTTTTATATTAGGTTATATTAC	50.7	794
	UpR.d19c	GTAATATAACCTAATATAAAAgTTAAGTCAAG	51.2	644
	DnF.d19d	CTTGACTTAAcTTTTATATTAGGTTATATTAC	50.7	794
	UpR.d20c	GTAATATAACCTAATATAAAgATTAAGTCAAG	51.2	644
	DnF.d20d	CTTGACTTAATcTTTATATTAGGTTATATTAC	50.7	794
	UpR.d1uc	AAATTAAGTCAAGTTAATTTTcATATTAGAT	51.2	625
	DnF.d1ud	ATCTAATATgAAAATTAACTTGACTTAATTT	50.5	812
	UpR.d2uc	AAATTAAGTCAAGTTAATTTTTgTATTAGAT	51.2	625
	DnF.2ud	ATCTAATAcAAAAATTAACTTGACTTAATTT	50.5	812
	UpR.d3uc	AAATTAAGTCAAGTTAATTTTTAcATTAGAT	51.2	625
	DnF.d3ud	ATCTAATgTAAAAATTAACTTGACTTAATTT	50.5	812
	UpR.d5uc	TTTTATAcTAGATGTAATTAATTCCGAGG	51.9	607
	DnF.d5ud	CCTCGGAATTAATTACATCTAgTATAAAA	50.8	828
	UpR.d6uc	TTTTATATcAGATGTAATTAATTCCGAGG	51.9	607
	DnF.d6ud	CCTCGGAATTAATTACATCTgATATAAAA	50.8	828
	UpR.d7uc	TTTTATATTgGATGTAATTAATTCCGAGG	51.9	607
	DnF.d7ud	CCTCGGAATTAATTACATCcAATATAAAA	50.8	828
	UpR.d8uc	TTTTATATTAaATGTAATTAATTCCGAGG	51.9	607
	DnF.d8ud	CCTCGGAATTAATTACATtTAATATAAAA	50.8	828
	UpR.d10uc	TTTTATATTAGAcGTAATTAATTCCGAGG	51.9	607
	DnF.d10ud	CCTCGGAATTAATTACgTCTAATATAAAA	50.8	828
	UpR.d11uc	TTTTATATTAGATaTAATTAATTCCGAGG	51.9	607
	DnF.d11ud	CCTCGGAATTAATTAtATCTAATATAAAA	50.8	828
	UpR.d12uc	TTTTATATTAGATGcAATTAATTCCGAGG	51.9	607
	DnF.d12ud	CCTCGGAATTAATTgCATCTAATATAAAA	50.8	828
5′ RACEe	sloA_GSP1	CGAGCCAAGAGCATACG
	sloA_GSP2	GCATAGTCTGCCAAATCAACC
	sloA_nested GSP	CCATTGGGGCCAATAATACC
Real-time semi-qRT-PCR	SloA.qPCR.F	CTGACCGAAGGCAAGGACTTT	49.6	112
	SloA.qPCR.R	GCTCAAGAAAGAAGGAAAGACCAT
	Hk11-RT-CP-F	GCTGGCTAATAATGTCATCAAGC	50.4	88
	Hk11-RT-CP-R	CTCAACAGTTACTTCAATCTCCTCC
Sequencing	SREseq.F	CGACCTAACTTTGACAGTGC	49.2	528
	SREseq.R	CATAGTCTGCCAAATCAACC
DNA footprinting	SloA.F1	ATCGGTGAATCGCACTGTCG	51.5	364
	SloA.R	GCCATCAATAAAACTTGTCCCTTC
Fluorescence anisotropyf	SRE24	5′-F-CTAAAATTAACTCGAGTTAATTTG
	SRE24-R	CAAATTAACTCGAGTTAATTTTAG
Gel shift	MW_SloA50_Ff	AATTACATCTAATATAAAAATTAACTTGACTTAATTTTTATATTAGGTTA		50
	MW_SloA50_R	TAACCTAATATAAAAATTAAGTCAAGTTAATTTTTATATTAGATGTAATT
	MW_SloA42_Ff	ACATCTAATATAAAAATTAACTTGACTTAATTTTTATATTAG		42
	MW_SloA42_R	CTAATATAAAAATTAAGTCAAGTTAATTTTTATATTAGATGT
	MW_SloA34_Ff	ACATCTAATATAAAAATTAACTTGACTTAATTTT		34
	MW_SloA34_R	AAAATTAAGTCAAGTTAATTTTTATATTAGATGT
	MW_SloA22_Ff	AAAATTAACTTGACTTAATTTT		22
	MW_SloA22_R	AAAATTAAGTCAAGTTAATTTT
	F.Sham.42f	GGATGATTACACATCATCGTGAATCTACGATGATGTGTAATC		42
	R.Sham.42	GATTACACATCATCGTAGATTCACGATGATGTGTAATCATCC
	recA.Gel.LN.F	CGGTTATCCAAAAGGGCGTATC	62.9	212
	recA.Gel.LN.R	CCTGTTCTCCTGAATCTGGTTGTG
	F.GMS612u.42	gCATCTAATATAAAAATTAACTTGACTTAATTTTTATATTAG	65	42
	R.GMS612u.42	CTAATATAAAAATTAAGTCAAGTTAATTTTTATATTAGATGc
	F.GMS605u.42	ACATCTAgTATAAAAATTAACTTGACTTAATTTTTATATTAG	64	42
	R.GMS605u.42	CTAATATAAAAATTAAGTCAAGTTAATTTTTATAcTAGATGT
	F.GM607.42	ACATCTAATATAAAAATTgACTTGACTTAATTTTTATATTAG	65	42
	R.GMS607.42	CTAATATAAAAATTAAGTCAAGTcAATTTTTATATTAGATGT
	F.GMS613.42	ACATCTAATATAAAAATTAACTTGgCTTAATTTTTATATTAG	66	42
	R.GMS613.42	CTAATATAAAAATTAAGcCAAGTTAATTTTTATATTAGATGT
	F.GMS607d.42	ACATCTAATATAAAAATTAACTTGACTTAATTTTTATATTgG	66	42
	R.GMS607d.42	CcAATATAAAAATTAAGTCAAGTTAATTTTTATATTAGATGT

^aLowercase lightface type represents restriction enzyme cut sites. Lowercase italicized type represents a mutated BamHI restriction site for Ser-to-Met conversion. Lowercase boldface type represents nucleotides that were modified with respect to the wild-type sequence.

^bAnnealing temperatures and amplicon sizes are relevant to the primer pair.

^cAnnealing temperatures and amplicon sizes derive from pairing with UpF.

^dAnnealing temperatures and amplicon sizes derive from pairing with DnR.

^ePCR amplification parameters are detailed in 5′RACE System for Rapid Amplification of cDNA Ends (catalog no. 18374-058) by Invitrogen/Life Technologies.

^fSingle-stranded oligonucleotides were annealed by gradual cooling from 95°C to 4°C in a programmed PCR machine.

SloR expression and purification.

The sloR gene was amplified by PCR from plasmid pES03 (Table 1) with primers slorF and slorR harboring BamHI and HindIII sites, respectively (Table 2), for subcloning into the vector pSMT3 (Table 1). The sloR open reading frame (ORF) was cloned in frame for cotranslation with an N-terminally His-tagged SMT3 domain so that it could be cleaved from the SloR protein with Ulp protease (13). Because of the cloning strategy, there was an N-terminal serine encoded with the cleaved SloR product. The serine was made to revert to the native N-terminal methionine via primer-directed mutagenesis using primers S1MF and S1MR (Table 2). The resulting plasmid, pSMT3_SloR, confers kanamycin resistance and places expression of the SMT3_SloR fusion under the control of the lacO operator and T7 RNA polymerase promoter. The plasmid was transformed into BL21(DE3) Escherichia coli cells (Table 1) which were grown in Luria broth containing 50 μg/ml kanamycin. For expression, 1-liter cultures were grown and monitored spectrophotometrically in a Nanodrop ND-1000 spectrophotometer (Thermo Fisher Scientific, Carlsbad, CA) until they reached an optical density at 600 nm (OD₆₀₀) of 0.6 to 1.0, and expression was induced at 37°C upon the addition of 0.15 g IPTG (isopropyl-β-d-thiogalactopyranoside) per liter. Following 2 to 3 h of additional incubation, cells were harvested by centrifugation in a Sorvall RC-5B centrifuge at 3,000 × g and then washed with lysis buffer (25 mM HEPES [pH 7.0], 300 mM NaCl, 10 mM imidazole, 5% glycerol) prior to storage at −80°C.

To purify SloR, cell pellets obtained from 2 liters of culture were lysed by sonication in a Branson 250 Digital Sonifier (Branson Ultrasonics Corp., Danbury, CT) and the clarified lysate was applied to a Co²⁺ affinity column that was washed further with lysis buffer and then with 300 mM imidazole in lysis buffer to elute the SMT3-SloR fusion protein from the column. Fractions containing purified fusion protein were combined, and 200 μg of Ulp protease harboring an N-terminal hexahistidine tag was added to effect cleavage. The mixture was dialyzed overnight against 1 liter of lysis buffer containing 1 mM β-mercaptoethanol. The dialyzed protein solution was then applied to a Ni²⁺ affinity column, and the fractions containing SloR in the void volume were combined and dialyzed against several exchanges of storage buffer (25 mM HEPES [pH 8.0], 300 mM NaCl, 10% glycerol), with the last round of dialysis proceeding in the presence of 10 g/liter Chelex resin. The SloR concentration was measured using a calculated extinction coefficient of 19,370 M⁻¹ cm⁻¹. Typically, 20 to 50 mg of purified protein was obtained per preparation. SloR solutions were flash-frozen in liquid nitrogen and stored at −80°C.

Expression of SloR containing selenomethionine (SloR-SeMet) was carried out in M9 minimal medium with glucose as a carbon source (14). Overnight growth of BL21(DE3) cells harboring the pSMT3-SloR plasmid in LB with 50 μg/ml kanamycin were pelleted and resuspended in M9 medium prior to transfer to 1-liter cultures in 2.8-liter Fernbach flasks. Growth proceeded at 37°C until an OD₆₀₀ of 0.4 was reached. Cultures were then supplemented with 100 mg of lysine, phenylalanine, threonine, and isoleucine, 50 mg of leucine and valine, and 60 mg of selenomethionine (Sigma-Aldrich), and 150 mg of IPTG per liter of culture was added 15 min later. Expression continued for 12 h, after which the cells were harvested by centrifugation as described above. Purification of SloR-SeMet proceeded as described above except that 1 mM Tris(carboxyethyl)phosphine was present in all solutions.

5′ RACE.

The transcription start site of the S. mutans sloABC operon was elucidated using a 5′ rapid amplification of cDNA ends (RACE) kit (Invitrogen, Carlsbad CA) to reveal the positioning of the predicted SRE relative to the −10 and −35 regions of the sloABC promoter. Total intact RNA was prepared as described below from mid-logarithmic-phase S. mutans UA159 cultures (OD₆₀₀, 0.4 to 0.7). The sloABC transcript was enriched from the total RNA pool by reverse transcription with primer SloA_GSP1 (where “GSP1” is “gene-specific primer 1”) (Table 2) and SuperScript II reverse transcriptase, and the cDNA product was dC tailed at its 5′ end in accordance with the recommendations of the supplier (Invitrogen). The cDNA was then PCR amplified with primer sloA_GSP2 (Table 2), the Abridged Anchor Primer (AAP) provided in the kit, and Platinum HiFi Taq polymerase according to the recommendations of the supplier (Invitrogen). The resulting PCR products were diluted 1:100 in Tris-EDTA (pH 8.0) buffer and subjected to a second round of amplification with a sloA nested GSP (Table 2) as described above. The amplified cDNAs were purified using a QIAquick PCR purification kit (Qiagen, Valencia, CA), quantified on a Nanodrop Lite spectrophotometer (Thermo Fisher Scientific), cloned into pGEM-T Easy vector (Promega) according to established protocols, and transformed into TOP10 E. coli (Life Technologies). Transformants demonstrating resistance to ampicillin on L-agar plates were screened by colony PCR using the M13F and M13R universal primer set (Table 2) and Platinum HiFi Taq polymerase according to the recommendations of the supplier (Invitrogen). Positive transformants were grown overnight in LB broth supplemented with 100 μg/ml ampicillin for plasmid DNA isolation using a QIAspin miniprep kit (Qiagen). The nested GSP (Table 2) was used to sequence the recombinant plasmid DNA by the use of a BigDye sequencing kit and a 3130 genetic analyzer (Applied Biosystems). The transcription start site was identified relative to the dC tail, and the −10 and −35 regions of the promoter were annotated accordingly.

Preparation of SRE mutant variants.

PCR amplicons with overlapping regions spanning the sloABC SRE were amplified from UA159 genomic DNA by the use of primers UpF and UpR.dN to generate a 600-bp 5′ product and primers DnR and DnF.dN to generate an 800-bp 3′ product (Table 2). Specifically, primers UpR.dN and DnF.dN were degenerated by one base pair and used to introduce the desired mutation into the overlapping region during a first round of overlap extension PCR (OE-PCR). During the final round of PCR, primers UpF and DnR were used to amplify a 1.4-kb PCR product that harbored the desired point mutation.

We used an established markerless mutagenesis approach to introduce SRE variants into the S. mutans UA159 chromosome (15). This approach employs an IFDC2 cassette containing a mutated phenylalanine tRNA synthetase, which incorporates toxic 4-chloro-phenylalanine into proteins during translation. S. mutans strain GMS602 harbors a copy of the IFDC2 cassette on the S. mutans chromosome that displaces the sloABC SRE and promoter region so that allelic exchange between sequences flanking the IFDC2 cassette and the 1.4-kb PCR product may occur. Overnight cultures of GMS602 were diluted with sterile Todd-Hewitt broth plus 0.5% yeast extract (THYE) to an OD₆₀₀ of 0.1 and incubated for 3 h in the presence of 150 ng of competence-stimulating peptide (CSP) and 5 μl of the 1.4-kb PCR product. Following incubation, the cell suspension was diluted 1:25 in 1× PBS, spread onto fresh brain heart infusion (BHI)–4-chloro-phenylalanine plates, and incubated for 48 h. Transformants were selected and patched onto fresh BHI–4-chloro-phenylalanine plates as well as onto BHI-erm¹⁵ plates to confirm loss of the IFDC2 cassette.

Confirmation of SRE variants.

Genomic DNA was isolated from transformants demonstrating substantial growth on BHI–4-chloro-phenylalanine plates but no growth on BHI-erm¹⁵ plates and was then used as a template for PCR using primers UpF and DnR (Table 2). For several SRE variants, whole-cell lysates were used as a template in the PCR (16). In either case, a 1.4-kb product indicated loss of the 3.1-kb IFDC2 cassette. The PCR product was purified using a PCR purification kit (Qiagen), and Sanger sequencing was used to confirm the presence of the mutation with primers SRE.seqF and SRE.seqR (Table 2).

EMSAs.

SloR-DNA binding was monitored in electrophoretic mobility shift assays (EMSAs) as previously described (12). Specifically, 22-bp, 34-bp, 42-bp, and 50-bp oligonucleotides, each harboring the minimal computationally predicted 22-bp SloR recognition element (SRE), were annealed and then assessed for SloR binding on 12% nondenaturing polyacrylamide gels, as were variants of the 42-bp SRE-containing DNA fragment bearing specific point mutations. In an effort to quantify SloR binding to the latter variants, band intensities were assessed by pixel counting using Kodak 1D 3.6 software (New Haven, CT) and SloR binding was calculated as a percentage of the total double-stranded DNA (dsDNA) target shifted in each lane. The ability of SloR to bind to a 42-bp “sham” palindrome with a unique nucleotide sequence was also monitored in separate EMSAs. All of these SloR binding targets were purchased as single-stranded oligomers (Table 2) from Sigma Genosys (St. Louis, MO), adjusted to a final concentration of 10 pmol/μl, and then either end labeled immediately or heated to 95°C for 5 min followed by gradual cooling to 4°C in an S1000 thermal cycler (Bio-Rad, Hercules, CA) to facilitate annealing. The resulting mixture containing predominantly double-stranded and some residual single-stranded oligomers was resolved on a 3% agarose gel, purified by gel extraction according to the recommendations of the supplier (Zymogen Research), and end labeled with [γ-³²P]dATP (PerkinElmer) in the presence of T₄ polynucleotide kinase (New England BioLabs). The end-labeled DNAs were centrifuged through a TE Select-D G-25 spin column (Roche Applied Science) to remove the unincorporated radiolabel and were diluted 2-fold before incubation in the presence of purified SloR protein in a 16-μl binding reaction mixture as described previously (12). EDTA was added to certain reaction mixtures at a final concentration of 15 mM to confirm the presence of SloR:SRE binding that is metal ion dependent. The samples were live-loaded at 40 V onto 12% nondenaturing polyacrylamide gels (3 ml 20× bis-Tris borate [pH 7.4], 74 μl 100 nM MnCl₂, 1.5 ml 100% glycerol, 24 ml 30% acrylamide [37.5:1 acrylamide-bis], 31 ml Millipore H₂O, 300 μl 15% ammonium persulfate [APS], 90 μl TEMED [N,N,N′,N′-tetramethylethylenediamine]) and then run for 450 V · h. Gels were exposed to Kodak BioMax film for up to 48 h at −80°C in the presence of an intensifying screen before being developed.

Fluorescence anisotropy.

Measurements of SloR binding to DNA were explored under equilibrium conditions using a 6-carboxyfluorescein-labeled 24-bp fragment of DNA containing the 22-bp sloABC SRE sequence (SRE24). Specifically, the fluorescently labeled duplex was prepared by annealing 5′-6-carboxyfluorescein (FAM)-CTAAAATTAACTCGAGTTAATTTG with a 10% molar excess of its unlabeled complementary strand. SloR-DNA complex formation was detected by measurement of fluorescence anisotropy using a Beacon 2000 fluorescence polarization spectrometer (Panvera LLC, Madison, WI). All assays were performed in FA buffer (25 mM HEPES [pH 8.0], 250 mM NaCl, 10% glycerol) with added metal ion salts as described above.

An assay of SloR binding to SRE24 was also performed using 50 nM duplex DNA in a 1-ml total sample volume in the presence of 1 mM Mn²⁺. The higher concentration of DNA shifts the equilibrium strongly toward complex formation and promotes a nearly stoichiometric association of the added SloR to the DNA duplex. Data were analyzed for best fit with a sequential binding model (17). The first and second dissociation constants were fitted using R statistical software (18). The stoichiometry of binding was obtained by manually testing values until the best fit, as measured by the residual standard error, was obtained.

Metal ion activation of SloR was also assayed by monitoring the anisotropy in fluorescent emission from 1 nM SRE24 in a 1-ml sample volume. Using R statistical software, data were fitted to the equation below, where [SloR] is the concentration of added SloR dimer, r is the measured anisotropy, r_min is the anisotropy of free SRE24, Δr is the difference in anisotropy between bound and unbound SRE24, and K_d is the dissociation constant of SloR from SRE24 (equivalent to the second binding event observed in the stoichiometry assay described above):

$$r=\frac{\mathrm{\Delta }\hspace{0.17em}\mathit{\text{r}}[\text{SloR}]}{K_d+[\text{SloR}]}+r_{\min }$$

In the case of trials performed using 10 μM Zn²⁺, an additional term, K_ns[SloR], was added to the equation to account for nonspecific (ns) interactions of SloR with its target DNA.

RNA isolation and cDNA synthesis.

S. mutans total RNA was isolated from UA159, GMS584, and a variety of different GMS602-derived S. mutans strains that harbor SRE variants in their chromosomes (Table 1), using a modification of a protocol described by Chen et al. (19). Briefly, 14 ml of mid-logarithmic-phase S. mutans cells (OD₆₀₀ = 0.5 to 0.6) was centrifuged at 11,000 × g for 5 min, resuspended in 1 ml sterile semidefined medium (SDM) and 2 ml of RNAprotect Bacteria Reagent (Qiagen), and allowed to incubate at room temperature for 5 min. Cell suspensions were centrifuged as described above, resuspended in 250 μl of 50 mM Tris–10 mM EDTA, and transferred to lysing matrix B tubes (MP Biomedicals, Santa Ana, CA) containing 10 μl of 10% (wt/vol) SDS, 300 μl of acid-buffered phenol, and 200 μl of zirconium beads. The cell suspensions were subjected to two 30-s rounds of mechanical disruption in a Bio101 FastPrep machine (MP Biomedicals), after which the cell slurry was centrifuged at 17,000 × g for 10 min. Supernatant (100 μl) was purified on an RNeasy column and treated with DNase I according to the manufacturer's instructions (Qiagen) to digest contaminating genomic DNA. Total RNA was examined for integrity on a 1% agarose gel and diluted to 100 ng/μl as determined on a Nanodrop Lite spectrometer (Thermo Fisher Scientific). RNA was reverse transcribed using a Revert-Aid First-Strand cDNA synthesis kit (Thermo Fischer Scientific) along with appropriate RNA-free and reverse transcriptase-free controls according to the manufacturer's instructions.

Real-time semi-qRT-PCR.

Experiments were performed using approximately 5 ng/μl of cDNA per 10 μl of reaction mixture. Thermal cycling was carried out using a CFX96 real-time system (Bio-Rad) and SsoFast Evagreen Supermix with SYBR green as the intercalating dye. The hk11 gene was used for normalization, as its expression is influenced neither by the SloR protein nor by mutations within the sloABC SRE. sloA- and hk11-specific primers (500 nM) (Table 2) were used in each reaction to amplify approximately 100 bp of PCR products internal to each gene's coding sequence on the cDNA. Thermal-cycling conditions were programmed at 95°C for 3 min followed by 40 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 15 s. Primer efficiencies were calculated using a standard curve of UA159 serial cDNA dilutions (1:5, 1:25, 1:125, and 1:625). RT-negative and no-DNA-template controls were included in all experiments to confirm the absence of contaminating genomic DNA.

DNase I footprinting.

Primers SloA.F1 and SloA.R (Table 2) were used to PCR amplify a 364-bp fragment from the S. mutans UA159 chromosome containing the sloABC promoter region and its promoter-proximal SRE. The reaction products were purified using a Qiaquick PCR column according to the manufacturer's instructions (Qiagen) and stored in nuclease-free H₂O at −20°C. Radiolabeled SloA.F1 or SloA.R was either used immediately to generate an end-labeled 364-bp amplicon or else stored at −20°C for no more than 3 days. Each of the resulting amplicons was purified as described previously in preparation for the binding reaction. Binding reactions involved adding native SloR protein, purified by anion exchange and heparin column chromatography (12), to 10 μl of 5× binding buffer and nuclease-free H₂O in a final volume of 45 μl. The final SloR concentration was 10 μM in a final volume of 50 μl containing 8.4 mM NaH₂PO₄, 11.6 mM Na₂HPO₄, 50 mM NaCl, 5 mM MgCl₂, 10 μg/ml bovine serum albumin, 200 μg/ml salmon sperm DNA, and 7.5 μM MnCl₂. The reaction mixtures were incubated for 10 min at room temperature before addition of 5 μl (∼20 μg/μl) of the radiolabeled amplicon as a binding target. The binding-reaction mixtures were mixed gently by the use of a vortex mixer, centrifuged briefly in a microcentrifuge, and then incubated for 30 min at 37°C to provide sufficient time for SloR-DNA binding. RQ1 DNase I and 5× RQ1 DNase I buffer were then added, and the reaction mixtures were incubated for 1 min at 37°C prior to the addition of prewarmed stop buffer (20 mM EGTA [pH 8.0]). The digested DNAs were purified by phenol chloroform extraction and precipitated overnight at −20°C in 2 vol of 100% ethanol. The DNAs were pelleted by centrifugation for 30 min at 4°C and 15,800 × g, and the pellets were washed with 70% ethanol followed by another round of centrifugation for 15 min at 4°C. The ethanol was decanted, and the DNA pellets were air dried prior to resuspending them on ice for 30 min in 5 μl of Stop/Loading buffer (0.1% [wt/vol] bromophenol blue, 0.1% [wt/vol] xylene cyanol, 10 mM EDTA, 95% [vol/vol] formamide). The end-labeled DNAs representing the coding and template strands were then loaded into separate wells on an 8% urea-containing polyacrylamide gel for 1,600 V · h, and the gel was exposed to Kodak BioMax film for up to 48 h at −80°C in the presence of an intensifying screen before being developed. DNA sequencing reaction mixtures were prepared with an Affymetrix Thermo Sequenase cycle sequencing kit in accordance with the recommendations of the supplier (USB, Cleveland, OH) and resolved alongside the DNA footprinting reaction mixtures.

SloR crystallization and structure solution.

Crystals of SloR could be grown using well solutions containing 10% to 15% polyethylene glycol (PEG) 3350, 40 mM lithium sulfate, and either 1 mM manganese chloride or 1 mM zinc chloride. Drops were composed of equal volumes of the well solution and SloR solution (10 mg/ml in storage buffer). Rod-shaped crystals grew over a few days to between 100 μm and 300 μm in length. Data were collected using synchrotron radiation at beamline 4.2.2 at the Advanced Light Source (Lawrence Berkeley National Laboratory, CA) from crystals that had been transferred to solutions containing 20% to 30% ethylene glycol in addition to the well solution components and had then been flash cooled in liquid nitrogen. In general, these crystals diffracted poorly. However, the selenomethionine derivative of SloR produced crystals that diffracted well in the presence of zinc ions (though not in the presence of manganese ions).

Structure solution and refinement were performed using the PHENIX suite of crystallographic software (20), and modeling was performed using Coot (21). The structure of SloR bound to zinc was obtained using data collected from a crystal of SloR-SeMet and the AutoSol routine employing SAD, based on anomalous scattering from selenium and zinc, for phase determination. An initial model was obtained via AutoBuild, and subsequent model building was performed in Coot, with rounds of refinement to the maximum-likelihood target proceeding against unmerged Bijvoet pairs using phase information. B-factor refinement took place on individual atoms, with four translation-libration-screw (TLS) groups identified by the TLSMD server (22).

Protein structure accession number.

Crystallographic data for the S. mutans SloR-SeMet protein have been deposited in the Research Collaboratory for Structural Bioinformatics (RCSB) Protein Data Bank (www.rcsb.org) with PDB ID number 5CVI.

RESULTS

The SRE overlaps the sloABC promoter.

We defined the transcription start site for the sloABC operon in 5′ RACE experiments so that its relationship with the SloR recognition element could be revealed. Specifically, the transcription start site is located 26 bp downstream of the 3′ end of the 22-bp SRE and 12 nucleotides upstream of the sloA start codon (Fig. 1). The palindromic sequence that comprises the 22-bp SRE completely overlaps the −35 promoter region (TTGACT), and its 3′ end is situated in close proximity (within 9 bp) of the −10 promoter element (TATATT).

FIG 1.

Anatomy of the sloABC promoter region. The sloABC transcription start site (shown at +1) is located 12 bp upstream of the sloA start codon and 26 bp downstream of the 3′ end of the 22-bp SloR recognition element (SRE). The predicted ribosome binding site (RBS), sloA start codon (Start), and −10 and −35 promoter regions are also designated underneath the sequence.

The 22-bp SRE is sufficient to accommodate the binding of two SloR dimers.

To elucidate the details of SloR binding to the 22-bp SRE in vitro, we performed EMSAs, the results of which resolved two band shifts when SloR was presented with SRE-containing DNA fragments of various sizes as binding targets (Fig. 2). These band shifts (designated C1 and C2) are consistent with SloR binding to its target sequence (F) as a low-molecular-weight dimer-DNA complex (C1) and a higher-molecular-weight double-dimer–DNA complex (C2). In addition, while a DNA fragment as small as a 22-bp SRE can accommodate binding of C2, this SloR double-dimer complex appears to be stabilized and hence is the favored species in binding to larger (i.e., 34-bp, 42-bp, and 50-bp) SRE-containing fragments (Fig. 2).

FIG 2.

SRE-containing fragments as small as 22 bp support SloR binding as single and double dimers. (a) Reaction mixtures containing 4.4 μM SloR and SRE-containing DNA fragments ranging from 22 bp to 50 bp (F) were resolved on a 12% nondenaturing gel for 750 V · h. Two different SloR binding species (C1 and C2) were evident when SloR was present in the reaction mixture, consistent with SloR binding to the target sequence as a single dimer (C1) and as a double dimer (C2). Double-dimer binding (C2) appears to be favored with increasing SRE length, consistent with stabilization of the SloR-SRE complex. A metal ion chelating reagent, EDTA, was added to selected reaction mixtures to reveal abrogation of the band shift and so validate that SloR-DNA binding is metal ion dependent. w/t, wild type. (b) Positioning of the 22-bp SRE within the 22-bp, 34-bp, 42-bp, and 50-bp SloR binding targets that were used as probes in the EMSA shown above.

The results of an assay of SloR binding to fluorescently labeled duplex DNA containing the SRE are also consistent with two SloR dimers binding to the operator (Fig. 3). The plot of anisotropy versus SloR dimer concentration suggests that two distinct binding events take place sequentially (17). The first, with a small increase in anisotropy, saturates with high affinity (K_d1 is less than 1 nM and is essentially unmeasurable in the fit determinations) after the addition of approximately 1 eq (1.16 ± 0.12; n = 3) of SloR dimer to a solution of 50 nM DNA. Subsequently, a second site is saturated with lower affinity (K_d2 = 24 ± 10 nM) after addition of a second 1-eq volume of SloR dimers.

FIG 3.

Titration of 50 nM SRE24 with SloR in the presence of 1 mM Mn²⁺ monitored by fluorescence anisotropy. The fitted curve derives from a sequential binding model. The fit indicates extremely tight binding by the first equivalent of SloR (K₁ < 1 nM) and a dissociation constant for the second equivalent of 24 ± 10 nM. The overall stoichiometry of binding is 2.3 ± 0.2 dimers per 24-bp DNA duplex.

SloR binds to double-stranded DNA (dsDNA) and makes base-specific contacts with the SRE.

It became evident from EMSAs that the palindromic SloR binding targets migrate as two distinct bands in 12% nondenaturing polyacrylamide gels and that SloR, when present in the reaction mixture, shifts only the upper one. For instance, Fig. 4a shows the 42-bp SRE-containing oligonucleotide that migrates as two distinct bands in the absence of any protein, consistent with the annealed dsDNA oligonucleotide that migrates more slowly through the gel matrix and the unannealed single-stranded DNA (ssDNA) oligonucleotide that migrates more rapidly through the gel. We noted that adding SloR to the reaction mixture decreased the intensity of the upper DNA band but had no impact on the lower band, indicating that SloR recognizes and binds to the oligonucleotide only when it is in its double-stranded configuration. To determine if this SloR-oligonucleotide interaction is dependent on the presence of a secondary structure that might result in vitro from the palindrome that is inherent in the resident 22-bp SRE and/or if SloR binding is sequence specific, we artificially synthesized a palindromic oligonucleotide equal in size to but with a sequence entirely different from that of the wild-type 42-bp SRE-containing DNA fragment. When this so-called sham 42mer (Table 2) was used as a target sequence in EMSAs with SloR, no band shifts were observed and hence no SloR-DNA binding was found (Fig. 4b). These observations indicate that SloR must make base-specific contacts with its DNA binding target and that any contacts that SloR makes with the sugar phosphate backbone of DNA are not sufficient alone to stabilize the SloR-DNA interaction.

FIG 4.

EMSA reveals SloR binding to double-stranded DNA that is sequence specific. A 42-bp oligonucleotide that harbors the wild-type 22-bp SRE and a 42-bp sham oligonucleotide that lacks any sequence that might resemble an SRE were allowed to interact with SloR protein provided at 10 nM or 60 nM prior to resolution on a 12% nondenaturing gel for 450 V · h. (a) SloR binds to the wild-type SRE sequence and shifts DNA in its double-stranded but not its single-stranded configuration. A metal ion chelating reagent, EDTA, added to selected reaction mixtures abrogated the band shift, thereby validating SloR-DNA binding that is metal ion dependent. (b) SloR does not bind to the 42-bp sham sequence, indicating that SloR recognizes and makes base-specific contacts with nucleotides within the SRE binding target.

Impact of SRE variants on sloABC transcription.

We tested variants with specific mutations within and flanking the 22-bp sloABC SRE in real-time semi-qRT-PCR studies to determine which base pair contacts are important for SloR-mediated gene transcription. Specifically, we monitored expression of the sloA gene in the SloR-proficient S. mutans UA159 strain, which harbors the wild-type 22-bp SRE that is in a promoter-proximal position with respect to the sloABC operon, and in SloR-proficient S. mutans mutants that harbor transition mutations at nucleotide positions within the 22-bp SRE and at positions up to 12 bp upstream and 8 bp downstream of the 22-bp SRE. The results of these experiments revealed significant derepression of sloA expression in variants with mutations at positions 3, 4, 7, 14, 16, and 20 within the 22-bp SRE, with an A/G transition at position 7 and T/C transitions at positions 14, 16, and 20 yielding the greatest derepression compared to the wild-type results (Fig. 5). Transitions positioned 12, 5, and 2 nucleotides upstream (12u, 5u, and 2u) and 2 and 7 nucleotides downstream (2d and 7d) of the 22-bp SRE also resulted in considerable if not significant derepression of sloA expression. Interestingly, SRE variants with mutations positioned at 9u, 4u, and 13 resulted in significant hyperrepression of sloA transcription. Taken together, these findings implicate nucleotides within the sloABC promoter-proximal SRE as important for SloR binding and support the idea of SloR binding to nucleotides within a 42-bp SRE that extends at least 12 nucleotides upstream and 7 nucleotides downstream of the canonical 22-bp SRE that was predicted computationally (4).

FIG 5.

Impact of SRE variants on sloABC transcription. Results of real-time semi-qRT-PCR experiments performed with the wild-type UA159 S. mutans strain and its derivative SloR-proficient SRE variants are shown. The means ± standard deviations of the results are shown for each of 3 independent experiments, each performed in triplicate. Six nucleotide substitutions within the 22-bp SRE and at least four substitutions in the upstream and downstream flanking sequences resulted in significant derepression of sloABC transcription compared to the wild-type results. *, P < 0.05 (Student's t test). Three nucleotide substitutions, at positions 9u, 4u, and 13, resulted in expression that was hyperrepressed. These findings support the idea of an expanded SRE that includes additional nucleotides that flank the canonical 22-bp SRE.

SloR binding to SRE variants is compromised.

To validate the base-specific contacts, which we identified in real-time semi-qRT-PCR experiments and confirmed to be important in SloR-SRE binding, we performed EMSAs with the wild-type 42-bp SRE-containing sequence and with several of its aforementioned mutant derivatives as SloR binding targets. The experimental results indicated that the A/T base pair at position 7 within the 22-bp SRE is especially important for SloR-SRE binding since substituting the adenine at this position with a guanine abolished the band shift (Fig. 6). In addition, SloR binding to an SRE variant with mutations at positions 12u, 5u, 7d, and 13 was also compromised compared to that seen with the wild-type strain, suggesting that the A/T base pairs at these positions are also important for the SloR-SRE interaction. Taken together, these findings support the idea of base-specific contacts between the SloR protein and the A/T base pairs that reside at these positions within and flanking the canonical 22-bp SRE.

FIG 6.

SloR binding to SRE variants is compromised relative to wild-type results. SloR binding targets that harbor the SRE that is proximal to the sloABC promoter were mixed with 60 nM purified SloR and resolved on 12% nondenaturing gels for 450 V · h in EMSAs. (a) Binding targets included the wild-type 42-bp SRE and 42-bp SRE variants with transitions at position 12u, 5u, 7, 13, or 7d. (b) The autoradiograph was scanned and analyzed with Kodak 1D software (version 3.6) to quantify the band intensities; the results are represented as percentages of the average total amount of double-stranded DNA shifted in each lane.

DNA footprinting validates an extended 42-bp SRE for SloR binding.

The results of EMSAs confirmed direct manganese-dependent binding of SloR to SRE-containing DNA fragments as small as 22 bp and indicated that specific mutations within and flanking the 22-bp SRE can either compromise or abolish SloR binding. To further inform a SloR-SRE binding model, we set out to determine the specific nucleotides that are recognized and bound by SloR in DNase I footprinting experiments. These studies were performed with a 364-bp DNA amplicon that harbors the SRE that is a promoter-proximal location with respect to the sloABC genes and which includes 120 nucleotides upstream of the sloA coding sequence. The 5′ end of the coding or template DNA strands was radiolabeled and incubated in the presence of 10 μM purified SloR protein and 125 μM Mn²⁺ to reveal the SloR binding pattern on both faces of the DNA. The experimental results revealed a DNA footprint that included a robust 42-bp region of protection on both the coding and template DNA strands (Fig. 7; see also Fig. S1 in the supplemental material) and which included at least two pairs of hexameric inverted repeats (TAATNT), with the repeats in each pair separated by the 8 bp that defines the −35 promoter element. Hypersensitive sites were localized both within and adjacent to the 42-bp footprint and, interestingly, aligned closely with the −35 and −10 promoter regions.

FIG 7.

SloR protects an extended region of DNA that includes the 22-bp SRE as well as the sloABC −35 promoter region. A 364-bp amplicon that harbors the sloABC promoter region and its promoter-proximal SRE was 5′ end labeled on the coding strand and digested with RQ1-DNase in the presence or absence of 10 μM purified SloR protein before resolution on an 8% denaturing polyacrylamide gel. Protection from DNase I digestion includes a 42-bp region of DNA that localizes upstream of the sloA +1 transcription start site and extends through the −35 promoter region. The positioning of the 42-bp SRE within the footprinted region is shown by the gray and black bar, with the black region representing the original 22-bp SRE that was determined computationally. The positions of the −10 and −35 promoter sequences, which align with regions of DNase I hypersensitivity on the DNA footprint, are indicated. The bent arrow reveals the position of the +1 transcription start site as well as the direction of transcription on the coding strand. The vertical arrows indicate the positioning of inverted hexameric repeats that fall within the 42-bp SRE.

Mn²⁺ is a better activator of SloR than Zn²⁺.

The ability of various divalent metal ions to activate SloR for DNA binding was monitored in vitro using anisotropy of fluorescence from 1 nM solutions of fluoresceinated duplex DNA fragments containing the SRE recognition sequence (Table 2). We tested Mn²⁺, Co²⁺, Zn²⁺, and Cd²⁺ as potential activating metals for SloR (Table 3). No DNA binding activity was detected under metal-free conditions. Saturating behavior was observed for SloR-DNA interactions in the presence of 10 μM and 1 mM Mn²⁺, with a slight improvement in DNA binding activity at the higher concentration. The observed dissociation constant of 32 ± 1 nM under the 1 mM Mn²⁺ condition is comparable to the second dissociation constant measured with stoichiometric concentrations of DNA (Fig. 3). That suggests that the first, high-affinity association seen in latter titrations was completed at low concentrations of SloR under the assay conditions that were used to monitor metal ion activation. SloR was likewise strongly activated by 1 mM Co²⁺ and somewhat more weakly activated by 1 mM Cd²⁺ (see Fig. S2 in the supplemental material). On the other hand, 10 μM Zn²⁺ was a poor activator of SloR, while the use of 1 mM Zn²⁺ led to anomalous anisotropy readings from the unbound DNA sample. The unusual readings at 1 mM were likely due to nonspecific interactions between Zn²⁺ and the DNA duplex.

TABLE 3.

Dissociation constants for the SloR-SRE24 interaction measured using fluorescence anisotropy in FA buffer and in the presence of different metal ions^a

Added metal ion(s)	Kd (nM)
10 μM Mn2+	58 ± 2
1 mM Mn2+	32 ± 1
10 μM Zn2+	420 ± 20
1 mM Mn2+ + 10 μM Zn2+	95 ± 2
1 mM Co2+	31 ± 6
1 mM Cd2+	62 ± 7

^aEach value represents the average of the results of at least three trials, and the standard error of the mean is reported for each.

The poor activation of SloR that we observed in the presence of 10 μM Zn²⁺ could have arisen from either of two possibilities: (i) Zn²⁺ could have been binding to the same metal ion binding sites as Mn²⁺ (but such binding fails to promote the conformation of SloR with high DNA binding affinity), or (ii) Zn²⁺ could have been binding to other sites on SloR or on the DNA duplex that interfered with the formation of the complex. To investigate these possibilities, we tested the ability of SloR to bind DNA in the presence of 10 μM Zn²⁺ and 1 mM Mn²⁺ (Table 3; see also Fig. S2 in the supplemental material). If Zn²⁺ were capable of supporting the fully active conformation of SloR, then comparable binding affinity would be expected with or without the addition of manganese, since the interference suggested by the second possibility would be present in both cases. However, SloR bound much more tightly when manganese was added, suggesting that 10 μM Zn²⁺ was not interfering with DNA binding in a nonspecific fashion. Rather, the first possibility, that zinc is a weak activator of SloR that competes with manganese for functional metal binding sites in each subunit, is the more likely of the two.

Structure of the SloR-Zn²⁺ complex.

The crystal structure of the selenomethionine derivative of SloR bound to zinc (SloR-Zn²⁺) was solved at 2.8 Å with a dimer in the asymmetric unit (Table 4). All predicted residues were identified in the electron density map except for the two N-terminal residues of the B chain. The SloR dimer strongly resembles the DtxR metalloregulator from Corynebacterium diphtheria (7, 8) and ScaR, a metalloregulator of manganese uptake in Streptococcus gordonii (23), in quaternary structure and domain organization within each subunit. An N-terminal DNA binding domain encompassed residues 1 to 72, a dimerization domain that formed a contact with the second subunit of the dimer encompassed residues 73 to 139, and a C-terminal domain bearing structural homology to the FeoA protein encompassed residues 140 to 215 (Fig. 8a). The DNA binding domain includes a winged helix-turn-helix (H-T-H) motif from residues 24 to 61, with the presumed recognition helix spanning residues 35 to 48. Residues Ser34, Pro36, Ser39 Glu40, Lys43, and Lys44 all occupy the outer face of the recognition helix and are positioned to interact with base pair edges and/or the sugar phosphate backbone. The dyad-related recognition helices are separated by 30.0 Å, as measured between the α-carbons of Glu40 of each subunit.

TABLE 4.

Data collection and refinement statistics for the SloR-Zn²⁺complex

Parameter	Result(s)a
Data collection statistics
Space group	P41212
Unit cell dimensions (Å)	a = b = 60.4, c = 298.8
Beam line	ALS 4.2.2
Wavelength (Å)	0.9785
Resolution range (Å)	59.23–2.80 (2.96–2.80)
No. of observations	198,579
No. of unique reflections	25,853
Completeness (%)	99.7 (98.9)
I/σ(I)	25.0 (5.1)
Rmerge (%)b	9.8 (53.2)
Wilson B factor (Å2)	49.9
Phasing
No. of sites in asuc	11 Se, 8 Zn
Resolution range	15.0–2.8
Figure of merit	0.38
Refinement statistics
Rcryst/Rfree (%)d	20.0/25.8
No. of subunits per asuc	2
No. of atoms
Protein	3,475
Solvent	32
Solute	12
RMSD bonds (Å)e	0.003
RMSD angles (°)	0.57
Ramachandran plotf
Favored regions (%)	97.4
Allowed regions (%)	100

^aNumbers in parentheses reflect the values for the highest-resolution shell.

^bR_merge = Σ_hΣ_j|I_h,_j − <I_h>|/Σ_jΣ_h|I_h,_j|, where I_h,_j is the jth observation of reflection h.

^casu, asymmetric unit.

^dR_cryst = Σ_h||F_o| − |F_c||/Σ_h|F_o|, where F_o and F_c are the observed and calculated structure factors for reflection h. R_free was calculated similarly for 10% of the data not used in refinement.

^eRMSD, root mean square deviation.

^fData were determined using Molprobity.

FIG 8.

(a) Quaternary and tertiary structures of SloR. The biologically active dimer of SloR appears in the asymmetric unit of SloR-Zn²⁺. One subunit is pink, and the second is colored by domain as follows: N-terminal DNA binding domain, cyan; dimerization domain, yellow; C-terminal FeoA-like domain, green. Zinc ions in conserved positions are shown as gray spheres with coordinating residues shown in the rightmost subunit. (b) Zinc binding in SloR. Zn²⁺ ions were observed to bind SloR in each of three sites on both subunits: the primary site (Zn_p), the ancillary site (Zn_a), and the secondary site (Zn_s). The zinc binding residues are labeled, as are two solvent molecules (W1 and W2). Atoms are colored by element (nitrogen, blue; oxygen, red; sulfur, orange; zinc, gray), with carbons colored by domain as described for panel a. Metal ligand interactions are represented as orange dotted lines, and hydrogen bonds that link coordination shells are shown as green dotted lines.

In total, eight zinc ions were bound by the SloR dimer. Six of these metal ions occupied sites shared by both subunits and with other DtxR family members, while two Zn²⁺ ions appeared to have bound adventitiously due to the relatively high (1 mM) concentration used in crystal growth. The presence of three bound zinc ions in both subunits was, even so, unexpected. Results of previous work on ScaR and SloR indicate the presence of two binding sites, the primary and secondary sites, while DtxR possesses two bound metal ions at the so-called primary and ancillary sites (7, 8). In SloR, we observed bound zinc ions at the primary (Asp7, Glu99, His103, and Glu102), secondary (Glu80, Cys123, and His125), and ancillary (His76, His95, and Asp160) sites (Fig. 8b). The naming of these sites is intended to capture analogy to the equivalent positions in DtxR and ScaR, though the residues contributing to each site differ somewhat between SloR and its two homologues. All three zinc ions were bound with tetrahedral coordination geometry, with the fourth ligand position in the secondary and ancillary sites occupied by a solvent molecule. Each of the sites possessed one ligand that was part of the secondary coordination shell of another site. In the primary site, Glu102 accepts an H-bond from His76 of the ancillary site, and the solvent ligand of the ancillary site donates an H-bond to Glu80 of the secondary site, creating a networked set of interactions among the metal binding sites (Fig. 8b). The zinc ions bound in the secondary and ancillary sites are 5.8 Å apart, while the distance from the ancillary site to the primary site zinc is 8.3 Å, comparable to the 9.2-Å separation of cobalt ions observed in DtxR (10).

DISCUSSION

Accumulating evidence sheds light on the SloR-SRE binding interface.

As a member of the DtxR family of metalloregulators, SloR is predicted to interact with DNA as a homodimer, particularly when metal ions are sufficiently available to occupy its primary, secondary, and/or ancillary metal ion binding sites. The winged H-T-H motifs at the N-terminal DNA binding domains of the SloR protein, and the Ser34, Pro36, and Glu40 residues in the recognition helix of the H-T-H motif in particular, are predicted to recognize base pairs that fall within adjacent major grooves on the DNA (12). Electrophoretic mobility shift assays demonstrated that SloR directly binds to the sloA promoter region, producing two DNA-protein complexes (Fig. 2). Considered in combination with the results of a fluorescence anisotropy assay (Fig. 3), the idea of binding of two SloR dimers to SRE-containing target sequences in the sloA promoter region is supported. This is reminiscent of two binding sites that are present in the S. gordonii scaC control region to which the ScaR metalloregulator binds (23, 24). Moreover, the results of DNA footprinting experiments support the idea of the symmetrical binding of SloR to the coding and template strands of DNA and reveal a large region of protection that includes at least 42 bp on both strands (Fig. 7; see also Fig. S1 in the supplemental material). Although a 42-bp SloR footprint seems large compared with that of C. diphtheria DtxR, which protects 33 bp through the action of two overlapping dimers (11), it is not unlike that of the MtsR metalloregulator in group A streptococci, which protects a 62-bp region of DNA, and that of the ScaR regulator in S. gordonii, which protects a 46-bp region (23–26).

Regions of enhanced DNase I cleavage are especially evident at the S. mutans −35 and −10 promoter regions on the coding strand, with the former separating the 42-bp protected region into two distinct footprints (Fig. 7). Hypersensitive sites often become more accessible to DNase I digestion as a consequence of DNA bending; hence, we propose that the SloR-SRE binding interaction at the sloABC promoter region induces alterations in DNA topology. Taking those results together with the data from studies of the DNase I-resistant regions that overlap the −35 promoter element, we believe that promoter occlusion is a plausible mechanism for sloABC repression by SloR and that it involves changes in DNA topology that prevent promoter access to RNA polymerase binding and downstream gene transcription. Inhibition of transcription elongation may also contribute to transcriptional repression by SloR, in light of another putative region of protection on the coding and template strands that span additional nucleotides downstream of the sloABC transcription start site (Fig. 7; see also Fig. S1 in the supplemental material).

Our findings also support the idea of a preference for SloR binding to thymine and adenine nucleotides, given that 12 of 13 nucleotide substitutions in the SRE that impacted sloABC transcription were conversions of A/T to G/C (Fig. 5). Many of these adenine/thymine base pairs, including those at positions 5u, 7, 16, and 20, fall within the SloR-DNA footprints that we observed on both the coding and template DNA strands (Fig. 7; see also Fig. S1 in the supplemental material), and moreover, they comprise part of the hexameric inverted-repeat units in this region. Taking the data collectively, it is reasonable to envision SloR binding as a homodimer corresponding to an apparent TAATNT consensus sequence that defines each hexameric repeat within the 42-bp SRE (Fig. 9). Indeed, the native sloA promoter/operator region includes at least two pairs of these hexameric repeats, with SloR presumably binding to each pair in 1:1 stoichiometry. In S. mutans, the hexameric repeats are separated by 2 bp, comprising a 6-2-6 binding motif. This organization is reminiscent of the 46-bp ScaR target sequence in S. gordonii, which includes two palindromic elements that straddle the scaC promoter (24). In fact, similar repeat sequences and promoter structures appear to be conserved in the control regions of the fim, psa, and mts operons across several streptococcal species (23–28), suggesting that they may share similar binding mechanisms with sloA.

FIG 9.

Binding model of the SloR-SRE interaction. SloR binds as two homodimers to a 42-bp SRE that overlaps the sloABC −35 promoter element on the S. mutans UA159 chromosome. The SloR homodimers engage in base-specific contacts within palindromic regions I and II on the DNA in a 1:1 stoichiometry. Each palindrome is comprised of two inverted hexameric repeats (TAATNT) separated by 2 bp (arrows), thereby defining a 6-2-6 binding motif (shown in red) for SloR. We propose that such binding occludes access of RNA polymerase to the −35 promoter region and so represses transcription of the downstream sloABC metal ion transport operon. The 42-bp region of DNA that is protected from DNase I digestion in DNA footprinting experiments is designated by the bracket, and the “lightning bolts” represent regions of DNase I hypersensitivity.

Although we have crystallized SloR in the absence of DNA, homology modeling provides some perspective on the ability of SloR to exert selectivity over a 22-bp region indicated by the studies reported here. A low-resolution crystal structure of MntR bound to a 26-bp DNA duplex containing its 22-bp operator sequence (unpublished observations) can be used to position a SloR dimer with respect to its DNA binding site. The dyad-related “wings” of the winged H-T-H motif define a 71-Å region of interaction between SloR and duplex DNA that permits interactions across the 22-bp stretch (see Fig. S3 in the supplemental material). In addition, it is possible that adjacent dimers that localize to palindromic regions I and II could make contact with one another and so protect base pairs that localize upstream and/or downstream of the computationally derived 22-bp SRE (Fig. 9). The recognition helix (residues 32 to 49) likely embeds in the major groove, which may allow contacts with bases that are 4 to 7 nucleotides away from the center of symmetry for the palindrome, while the wing stretches across the minor groove, permitting contacts with nucleotides at positions 10 and 11 bp from the palindromic center. The N termini of each subunit are in close proximity to the sugar phosphate backbone at positions 2 and 3.

Previous work in our laboratory highlighted the roles of Pro36, Ser39, and Glu40 in the ability of SloR to act as a transcriptional repressor (12). Our new homology model positions these residues directly in the major groove of the target DNA and in a position to interact with nucleotides at positions 7, 16, and 20 of the SRE. Positions 7 and 16 are equivalent in the proposed half-sites shown in Fig. 9, and Ser39 is well positioned to interact with adenines at that position. The adenines at positions 7 and 16 were especially sensitive to mutation in our expression profiling studies, and substitutions at position 7 in particular consistently led to the loss of SloR binding affinity as reported here (Fig. 6). Similarly, Glu40 is directly adjacent to position 20, which is also sensitive to mutation in our assays. Pro36, at the N terminus of the recognition helix, is close to base pairs at positions 5 and 6 (17 and 18 in the equivalent half site) and could be responsible for forming key van der Waals contacts with thymines at those positions. Studies of the impact of mutations at these nucleotide positions on sloABC transcription and SRE binding are under way.

A number of unresolved issues remain with respect to the structural mechanism by which SloR binds DNA. For instance, the current binding model does not fully explain the significance of the adenine residue at position 12u, which is important for SloR binding in light of our real-time semi-qRT-PCR results (Fig. 5). However, according to our homology model, the nucleotide at this position is immediately adjacent to nucleotides potentially in contact with the winged H-T-H motif of SloR. It may be that the A/T base pair that includes adenine 12u contributes to a DNA conformation that is selectively recognized by SloR. It is known that A/T-rich regions of DNA are susceptible to bending in the minor groove, which can be essential to the formation of some protein-DNA complexes (29). Additionally, the 2:1 binding stoichiometry of SloR dimers with the 22-bp and 24-bp duplexes that we used in binding studies is surprising given that only independent half-sites for SloR binding are present in the sequence. Results from the sequential binding model that we use to fit data from fluorescence anisotropy titrations suggest that there may be cooperative interactions between SloR dimers that bind to the region. Finally, the compromised binding of SloR to an SRE variant with a mutation at nucleotide position 13 is at odds with the hyperrepression of sloABC transcription that derives from this variant in real-time semi-qRT-PCR experiments. This might be explained by the positioning of this mutation within the −35 sloABC promoter element which, in turn, affects RNA polymerase binding and transcription. Alternatively, this finding might imply dual roles for SloR, with a role as a facilitator of S. mutans gene expression in addition to its role as a repressor. We are continuing our exploration of these issues.

A crystal structure informs the SloR metal ion-effector interaction.

The SloR-Zn²⁺ structure provides the first view of a DtxR family member with three (primary, ancillary, and secondary) metal ion binding sites per subunit simultaneously occupied. Accumulating evidence supports the idea of involvement of all three binding sites in SloR functionality as a repressor or facilitator of S. mutans gene transcription. Specifically, Haswell et al. (12) reported functional roles for Glu99, Glu102, and His103 in the primary metal ion binding site and for His76, Glu80, Cys123, and His125 in the secondary binding site, and recent work in our laboratory confirmed involvement of His76 and Asp160 in binding metal at the ancillary site (unpublished observations). Several structures of DtxR and its close cognate, IdeR from Mycobacterium tuberculosis (MtIdeR), include occupied primary and ancillary sites (30), but residues that could form the secondary site are not present in those proteins. A recently published structure of IdeR from Thermoplasma acidophilum (TaIdeR) contains two iron ions (the oxidation state of each iron atom is not clear) bound with different geometries in the secondary and ancillary sites, while the primary site remains vacant (31).

SloR belongs to a group of metalloregulators from streptococcal species that share substantial sequence identity: ScaR from Streptococcus gordonii, PsaR from Streptococcus pneumoniae, and MtsR from Streptococcus pyogenes are 62%, 58%, and 57% identical in sequence to SloR, respectively (23, 25, 28). Each of these proteins has been previously characterized, and, in the case of ScaR, X-ray crystal structures of the metal-free and the Zn²⁺- or Cd²⁺-bound forms have been reported (23). MtsR, like SloR, appears to regulate both Fe²⁺ and Mn²⁺ levels in the cell (26), whereas ScaR and PsaR function solely as Mn-sensitive regulators (24, 28). Biochemically, all three appear to be relatively unselective for the activating metal in vitro, though evidence indicates that zinc is a poor activator of both PsaR and ScaR. The structure of Zn²⁺-activated ScaR contains a single occupied site per subunit that is not equivalent to either the primary site or the ancillary site of DtxR, involving residues E80, C123, H125, and D160 (23). Additionally, a set of residues (D7, E99, H102, and E103) at positions equivalent to those in the primary site of DtxR appear capable of forming a metal binding site and evidence indicates that two Mn²⁺ ions are bound per subunit in its active state. In PsaR, two metal binding sites have been identified via spectroscopic work, with a likely zinc-specific site showing similarity to the secondary site of ScaR and a manganese-specific site that has tentatively been identified as equivalent to the primary site of ScaR and DtxR (32, 33). PsaR is poorly activated when both sites are occupied by zinc but has high affinity for its cognate DNA sequence in a mixed-metal state, with one Zn²⁺ ion and one Mn²⁺ ion binding per subunit.

It is likely that the structure of SloR-Zn²⁺ represents a low-activity conformation of the protein. SloR has roughly 10-fold-lower affinity for its cognate DNA sequence in the presence of 10 μM Zn²⁺ than in the presence of 10 μM Mn²⁺. Furthermore, results from work on the close homologs ScaR and PsaR indicate that zinc is a poor activator of other streptococcal members of the DtxR family. The inhibitory effect of Zn²⁺ has been studied in PsaR and may be further understood via inspection of the SloR-Zn²⁺ structure. In PsaR, zinc binds to the secondary site with high affinity in an active conformation that includes Mn²⁺ binding to at least one other site. However, when zinc replaces Mn²⁺ at that additional site, DNA binding affinity is reduced. Our results suggest that the reduction in affinity is at least partially due to substitution of Zn²⁺ for Mn²⁺ at the primary site. The tetrahedral geometry of the primary-site Zn²⁺ in SloR differs from the octahedral geometry that has been observed with active complexes of all other structurally characterized DtxR homologs. The hexacoordinate geometry is achieved via additional ligating interactions from the backbone carbonyl of the residue equivalent to Glu99 and a solvent molecule that is missing in the SloR-Zn²⁺ complex (10, 34). In addition, a more distantly related DtxR family member, the Mn-specific MntR from B. subtilis, whose C site is composed of the same set of residues as the primary site of SloR, binds strongly activating metals with octahedral geometry (32, 35). Mn²⁺ typically prefers octahedral geometry in biological settings (36) and likely incorporates the backbone carbonyl of Glu99 as part of its coordination shell in the primary site of SloR. The lowered activity of SloR when exposed to Zn²⁺ is due to an alteration in primary-site geometry, which could be transmitted to improper positioning of the HTH motifs relative to one another.

A similar mechanism for a selective response to Mn²⁺ in preference to Zn²⁺ has been described in MntR, where the lower coordination number of bound Zn²⁺ leads to a metal ion complex with lower DNA binding affinity, albeit also with lower metal binding stoichiometry (32). To date, there are no confirmed examples of a DtxR family member functioning as a Zn²⁺-responsive regulatory protein. It could be that the octahedral geometry of the primary site, which is common to all known family members, is a functional screen against activation by Zn²⁺. However, results of a recent study of MtIdeR suggested that Zn²⁺ may productively bind to the primary site of that regulator, though the affinity of MtIdeR and DtxR (which share the same set of metal binding residues) for DNA is lowered in the presence of Zn²⁺ relative to iron and other divalent metal ions (11, 37). Additional structures of zinc-bound MtIdeR and Mn²⁺-bound SloR will help illuminate the role of primary-site geometry of activation of DtxR family members.

In summary, we defined at least two cis-acting elements within a 42-bp expanded SloR binding target that result in the formation of two SloR-DNA complexes with different mobilities in EMSAs. Taking those findings together with the results of DNA footprinting studies, we propose a SloR-DNA binding model that accommodates SloR dimers in a 2:1 stoichiometry with a 42-bp core sequence consisting of two hexameric inverted-repeat units that abut the −10 and −35 promoter elements. Hence, we hypothesize that under conditions of feast, SloR is activated by Mn²⁺ to bind to inverted hexamers, thereby blocking transcription of the sloABC metal ion transport system. We suggest that the chief regulatory site for Mn²⁺ binding to SloR is formed by residues Asp7, Glu99, His103, and Glu102 in the primary site. It is unclear whether the ancillary and secondary sites play a structural or regulatory role. The activated SloR-Mn²⁺ binding complex, by remaining bound to its core target sequences, would foster S. mutans survival and persistence in the plaque environment by preventing the excessive uptake of exogenous manganese. As manganese levels drop, such as during the times between meals, however, the primary site is vacated and SloR is released from its promoter-proximal target sequence. The sloABC genes would therefore become transcriptionally active so that S. mutans could scavenge for essential manganese during these extended periods of famine. We hypothesize that S. mutans coordinates this control of metal ion homeostasis with that of its virulence attributes to ensure its success as an oral pathogen. Indeed, we have already demonstrated SloR binding to the promoter regions of multiple virulence genes throughout the S. mutans genome (2, 3; unpublished observations), including those whose products contribute to S. mutans adherence, biofilm formation, and acid and oxidative stress tolerance. The details of the SloR-DNA binding interaction at these loci await further characterization.

DNA footprinting experiments with single- and double-mutant SRE variants, and with variants that harbor mutations in the protected and proposed hexameric repeat units, are under way in our laboratory to see if SloR binding in the sloA promoter region is compromised. Future experiments will focus on resolving a crystal structure of the SloR-Mn²⁺ complex bound to the 42-bp SRE to further elucidate the details of the SloR-SRE binding interaction at the sloABC locus and at other virulence gene loci that are subject to SloR control. Taken together, these studies can further inform a S. mutans SloR SRE binding model and so elucidate strategies for rational drug design aimed at alleviating and/or preventing caries and associated complications.