APPLIED AND ENVIRONMENTAL MICROBIOLOGY, May 2009, p. 2629–2637 Vol. 75, No. 90099-2240/09/$08.00�0 doi:10.1128/AEM.02145-08Copyright © 2009, American Society for Microbiology. All Rights Reserved.

AguR Is Required for Induction of the Streptococcus mutansAgmatine Deiminase System by Low pH and Agmatine�

Yaling Liu, Lin Zeng, and Robert A. Burne*Department of Oral Biology, University of Florida, Gainesville, Florida

Received 16 September 2008/Accepted 22 February 2009

Acidic conditions and the presence of exogenous agmatine are required to achieve maximal expression of theagmatine deiminase system (AgDS) of Streptococcus mutans. Here we demonstrate that the transcriptional activatorof the AgDS, AguR, is required for the responses to agmatine and to low pH. Linker scanning mutagenesis was usedto create a panel of mutated aguR genes that were utilized to complement an aguR deletion mutant of S. mutans. Thelevel of production of the mutant proteins was shown to be comparable to that of the wild-type AguR protein.Mutations in the predicted DNA binding domain of AguR eliminated activation of the agu operon. Insertions intothe region connecting the DNA binding domain to the predicted extracellular and transmembrane domains werewell tolerated. In contrast, a variety of mutants were isolated that had a diminished capacity to respond to low pHbut retained the ability to activate AgDS gene expression in response to agmatine, and vice versa. Also, a numberof mutants were unable to respond to either agmatine or low pH. AguD, which is a predicted agmatine-putrescineantiporter, was found to be a negative regulator of AgDS gene expression in the absence of exogenous agmatine butwas not required for low-pH induction of the AgDS genes. This study reveals that the control of AgDS geneexpression by both agmatine and low pH is coordinated through the AguR protein and begins to identify domainsof the protein involved in sensing and signaling.

The agmatine deiminase system (AgDS) of Streptococcusmutans is encoded by the aguBDAC operon (11). Agmatineenters the cell via an agmatine-putrescine antiporter (AguD),where it is hydrolyzed to N-carbamoylputrescine and ammoniaby agmatine deiminase (AgD; EC 3.5.3.12), encoded by aguA.Putrescine carbamoyltransferase (EC 2.1.3.6), encoded byaguB, mediates the phosphorolysis of N-carbamoylputrescine,yielding putrescine and carbamoylphosphate. Finally, a phos-phate group is transferred from carbamoylphosphate to ADPby carbamate kinase (EC 2.7.2.2), the product of the aguCgene, to generate ATP, CO2, and NH3. Putrescine is thenexchanged for agmatine via the antiporter.

Streptococcus mutans is the etiological agent of dental caries(17). A key virulence attribute of this organism is acid toler-ance, since catabolism of dietary carbohydrates lowers the pHof dental plaque to values below 4.0 (3). S. mutans has multiplestrategies to cope with low pH and gains a competitive advan-tage over acid-sensitive bacteria under acidic conditions (19,21). Although a variety of oral streptococci utilize ammoniageneration from urea or from amino acids as a fundamentalacid tolerance mechanism (19), alkali production was not con-sidered a mechanism employed by S. mutans to cope withacidification until the AgDS was characterized for S. mutansUA159 (11) and, subsequently, for other mutans streptococci(13). The primary roles of the AgDS in S. mutans and othermutans streptococci appear to be to catabolize agmatine, toproduce ATP, and to augment acid tolerance (11, 12).

The aguR gene, which encodes a trans-acting factor requiredfor efficient AgDS expression, is located 239 bp upstream of

aguB and is transcribed in the opposite direction (12). AguR isa putative LuxR-like transcriptional regulator belonging to theFixJ-NarL superfamily, which includes two-component re-sponse regulators involved in quorum sensing (10). AguR is318 amino acids long, with a predicted molecular mass of37,611 Da, and has several noteworthy structural features (Fig.1). Computer algorithms at http://smart.embl-heidelberg.de/and http://bp.nuap.nagoya-u.ac.jp/sosui/ predict an N-terminalsignal sequence and as many as four membrane-spanning do-mains that may allow for a substantial portion of the protein tobe exposed to the environment. A predicted helix-turn-helix(HTH) motif of the LuxR family is located in the C terminusof the protein and likely mediates DNA binding. It was re-ported previously that induction of the AgDS of S. mutans isoptimal at low pH when agmatine is present (12). In this study,we tested the hypothesis that agmatine and low-pH inductionof the AgDS of S. mutans is mediated through the AguRtranscriptional regulator.

MATERIALS AND METHODS

Bacterial strains, growth conditions, and reagents. S. mutans UA159 and itsderivatives were maintained in brain heart infusion broth (BHI; Difco Labora-tories, Detroit, MI) at 37°C in 5% CO2 and 95% air, with antibiotics added at thefollowing concentrations when necessary: kanamycin (Km), 1 mg ml�1; erythro-mycin (Em), 5 �g ml�1; and spectinomycin (Sp), 1 mg ml�1. To monitor AgDSexpression, batch cultures of S. mutans were grown in a tryptone-vitamin (TV)base medium (5) supplemented with 25 mM galactose, with or without 10 mMagmatine, to an optical density at 600 nm (OD600) of 0.5. Escherichia coli strainswere grown in Luria-Bertani (LB) medium supplemented with antibiotics at thefollowing concentrations: ampicillin (Ap), 100 �g ml�1; Km, 50 �g ml�1; Em,500 �g ml�1; and Sp, 50 �g ml�1. Chemical reagents and antibiotics wereobtained from Sigma (St. Louis, MO).

For studies on pH- and agmatine-dependent regulation of AgDS expression, S.mutans strains were grown in continuous culture in a Biostat i Twin Controllerchemostat (Braun Biotech, Inc., Allentown, PA) in a tryptone-yeast extract (TY)medium (29) supplemented with 25 mM galactose at a dilution rate (D) of 0.3h�1. Cultures were maintained at pH 5.5 or pH 7.0 by the addition of 2 M KOH.

* Corresponding author. Mailing address: Department of Oral Biology,University of Florida, P.O. Box 100424, Gainesville, FL 32610-0424.Phone: (352) 392-4370. Fax: (352) 392-7357. E-mail: rburne@dental.ufl.edu.

� Published ahead of print on 6 March 2009.

2629

on Septem

ber 9, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

Samples were harvested at steady state, which was defined as cultivation undergiven conditions for at least 10 generations. Where indicated, cultures werepulsed with 10 mM agmatine for 1 h prior to sampling.

DNA manipulation. Genomic DNA was isolated from S. mutans UA159 aspreviously described (7). Plasmid DNA used in sequencing reactions was pre-pared from E. coli by use of a miniprep kit (Qiagen, Valencia, CA). Restrictionand DNA-modifying enzymes were purchased from Life Technologies Inc.(Rockville, MD) or New England Biolabs (Beverly, MA).

Construction of reporter gene fusions. A 157-bp fragment upstream of theaguB start codon (PaguB) was amplified using primers PaguB-S and PaguB-AS

(Table 1), with inserted SacI and BamHI restriction sites to facilitate cloning.The promoter region was fused to a promoterless lacZ gene derived fromStreptococcus salivarius on pMC195 (6). After confirming the correct sequence ofthe promoter fusion by DNA sequencing, the PaguB-lacZ construct was clonedinto plasmid pMC340B (30), which allows for integration of the gene fusion in S.mutans as a single copy at the mtlA-phnA locus, and the construct was trans-formed into S. mutans UA159 to create strain WTZ. Transformants were se-lected on BHI agar with Km and screened by PCR to confirm that the desiredintegration had occurred. Strain WTZ was transformed with plasmid pMSP3535(4), which allows for nisin-inducible expression from the nisA promoter, to

FIG. 1. Predicted topology and linker insertions for AguR. The four-transmembrane-domain model is based on computer algorithms from http://smart.embl-heidelberg.de/ and http://bp.nuap.nagoya-u.ac.jp/sosui/. Transmembrane domains are indicated by boxes and numbered in the top leftcorners. The circles indicate that insertions adjacent to these residues resulted in altered responses to pH and agmatine. The square indicates thatinsertions at this residue resulted only in altered responses to agmatine. The diamond indicates that insertions at this point in AguR altered only pHresponsiveness. The hexagons indicate that mutagenesis at these residues resulted in no significant changes in the response to stimuli.

TABLE 1. Primers used in this study

Primer Sequence (5�–3�)a Applicationb

PaguB-S GCATCCAACGAGCTCGGAGCTCCTTT PaguB amplificationPaguB-AS GAGGAGGGATCCATGATGAAAAAAACAGA PaguB amplificationaguR1-S AAAGATCTAGAAAAGGAGGGTTCTTCCATG Deletion of aguRaguR1-AS CTTGTCATGTACAGAATATTTCCAATTTA Deletion of aguRaguR2-S AAAGAGGATCCAAAGGAGGGTTCTTCCATG Cloning of aguR into pMSP3535aguR2-AS CTTGTCACTCGAGGAATATTTCCAATTTA Cloning of aguR into pMSP3535aguR3-S AGTGTTGCTATTTTAATTGGATCCACTTTTGTTATT Recombinant AguR expressionaguR3-AS CTTGTCACTCGAGGAATATTTCCAATTTA Recombinant AguR expressionaguD-S GTGTTGAACGTCACCAAAGTG Deletion of aguDaguD-SmaI-AS CTTTCCTTCCATACCCGGGTTCCTTTCTT Deletion of aguDaguD-SmaI-S TAGGTGAACTTATTATTATCATCCCGGGAAT Deletion of aguDaguD-AS CGTCCCTTGGCATCTGTCTG Deletion of aguDpMSMP3535F GATTCAGGTGCATCACCACGCATT DNA sequencing for recombinant plasmidpMSMP3535R GGCCGATTCATTAATGCAGGTTAACCT DNA sequencing for recombinant plasmidaguR-RT-S CACTACCATCACTGCCCAATAC For RT-PCR or real-time PCR to amplify aguR geneaguR-RT-AS TAGCAACACTGGCAATCATAGC For RT-PCR or real-time PCR to amplify aguR gene

a Sequences in bold indicate engineered restriction sites.b RT-PCR, reverse transcription-PCR.

2630 LIU ET AL. APPL. ENVIRON. MICROBIOL.

on Septem

ber 9, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

generate WTZ/V, or with a nisR/nisK-deficient derivative of pMSP3535 (pMSP-�nisRK) to generate WTZ/V-RK. pMSP-�nisRK was constructed by removal ofan EcoRV fragment from pMSP3535 to delete both the nisR and nisK genes.Plasmid transformants were selected on BHI agar containing Em.

Deletion of aguR. Primers aguR1-S and aguR1-AS (Table 1) were used toamplify a 2.2-kbp aguR-containing DNA fragment. The PCR product was thendigested with BsrGI and HindIII and ligated with an Sp resistance cassette. Theligation mixture was introduced into strains WTZ, WTZ/V, and WTZ/V-RK(Table 2) by natural transformation to generate R�, R�/V, and R�/V-RK,respectively, and transformants were selected on BHI agar with Sp. The correctmutations were confirmed by DNA sequencing.

Construction of aguR-complemented strains. The expression vectorpMSP3535, which allows for controlled expression of inserted genes by additionof subinhibitory concentrations of nisin to the growth medium (4, 20), was usedfor complementation studies with the aguR gene of S. mutans. To construct anaguR-complemented strain, a 1.1-kbp fragment of aguR, containing the aguRcoding sequence and the native ribosome binding site, was amplified with prim-ers aguR2-S and aguR2-AS to introduce BamHI and XhoI restriction sites forsubsequent cloning (Table 1). The PCR product was then digested with BamHIand XhoI and cloned into pMSP3535, ensuring that transcription of aguR was inthe same orientation as that of the nisin promoter, resulting in pMSP-aguR.Faithful amplification of the gene was confirmed by DNA sequencing withprimers pMSP3535F and pMSP3535R (Table 1). The complemented strain(R�/R) was constructed by introducing pMSP-aguR into the AguR-deficientmutant of S. mutans (R�) by natural transformation, and transformants wereselected on BHI agar with Em.

Deletion of aguD. Primers aguD-S and aguD-SmaI-AS (Table 1) were used toamplify a 0.6-kbp region upstream of aguD. Primers aguD-SmaI-S and aguD-AS(Table 1) were used to amplify a 0.6-kbp region downstream of aguD. The PCRproducts were digested with SmaI and ligated to a nonpolar Km resistancecassette from pALH124 (2). The ligation mixture was introduced into S. mutansUA159 by natural transformation to generate an aguD mutant, and bacteria wereplated on BHI agar with Km. Double-crossover mutants were confirmed by PCR.

Linker scanning mutagenesis. Mutagenesis was performed with the GPS-LSlinker scanning system (New England Biolabs) according to the supplier’s in-structions. The GPS-LS system allows for random insertion of 15 nucleotides intothe gene of interest, with only two of the six possible reading frames creating stopcodons. In the other four reading frames, the insertion introduces five aminoacids into the protein of interest. A transposon (Tp) donor plasmid (pGPS5; Kmresistant) and the TnsABC transposase were used to mutagenize plasmid pMSP-aguR in vitro. The mutagenized plasmids were electroporated into E. coli byselection on LB agar containing Km, and the colonies were pooled for isolationof plasmids. The Tp-containing XbaI-SphI fragments harboring the mutatedaguR genes from pools of targeted plasmids were cloned back into pMSP3535 toconstruct pMSP-aguR::Tp plasmids. One hundred individual pMSP-aguR::Tpplasmids were each digested with PmeI to release the Tp and then were religated.The resulting clones were selected on BHI agar containing Em and screened forKm sensitivity. Insertions were located by DNA sequencing with the primerspMSP3535F and pMSP3535R (Table 1). Selected pMSP-aguR::Tp plasmids wereused to transform the R� strain, and transformants were selected on BHI agarcontaining Em.

�-Galactosidase assay. Measurements of �-galactosidase activity were ob-tained using a modification of the Miller protocol (31). Briefly, cells were har-vested by centrifugation, washed once with Z buffer (60 mM Na2HPO4, 40 mMNaH2PO4, 10 mM KCl, 1 mM MgSO4, and 5 mM �-mercaptoethanol, pH 7.0),

and resuspended in 1/10 of the original culture volume in the same buffer. Onemilliliter of the suspension was used to determine the OD600. Another 500 �l ofthe sample was vortexed with 25 �l of a toluene-acetone mix (1:9) for 2 min andthen kept at 37°C. A reaction was initiated by adding 100 �l of ONPG(o-nitrophenyl-�-D-galactopyranoside) solution (4 mg/ml) and terminated bythe addition of 500 �l of 1 M Na2CO3. Samples were centrifuged at maximumspeed for 2 min using a tabletop centrifuge, and the OD of the supernatant fluidwas measured at 420 and 550 nm. Activity was expressed in Miller units (31).

Expression and purification of recombinant AguR in E. coli. A 324-bp frag-ment encoding the C-terminal one-third (amino acids 210 to 318) of AguR wasamplified by PCR using primers aguR3-S and aguR3-AS, with BamHI and PstIsites added (Table 1). The PCR product was then digested with BamHI and PstIand inserted in frame behind the malE gene on pMAL-p2X (New EnglandBiolabs) to create a maltose binding protein fusion. Expression of the recombi-nant AguR protein was induced with 1 mM IPTG (isopropyl-�-D-thiogalactopy-ranoside) for 3 h, and protein lysates were prepared by homogenization with aBeadBeater (Biospec, Bartlesville, OK). The recombinant protein was purifiedusing an amylose column as recommended by the supplier. Polyclonal AguRantibody was elicited in rabbits with the purified recombinant AguR protein atLampire Biological Laboratories (Pipersville, PA).

Western blot assays of AguR protein. Mid-exponential-phase (OD600 � 0.5 to0.6) cells were centrifuged and washed twice with Tris-buffered saline (10 mMTris, 0.9% NaCl, pH 7.4). Whole-cell lysates were obtained by homogenizationin sodium dodecyl sulfate (SDS) boiling buffer (60 mM Tris, pH 6.8, 10%glycerol, and 5% SDS) in the presence of glass beads, followed by centrifugationat 2,000 � g for 10 min (8). Protein samples were separated by SDS-polyacryl-amide gel electrophoresis (SDS-PAGE), blotted onto polyvinylidene difluoridemembranes, and incubated with affinity-purified (23) AguR antibody, followedby peroxidase-conjugated goat anti-rabbit immunoglobulin G (KPL, Gaithers-burg, MD). Signals were developed using a SuperSignal West Femto chemilu-minescence kit (Thermo, Waltham, MA).

Real-time PCR. Extraction of RNA, reverse transcription-PCR, real-timePCR, and copy number normalizations were performed as previously described(1). The primers used for reverse transcription-PCR and real-time PCR for theaguA gene are described elsewhere (12). The aguR-specific primers (aguR-RT-Sand aguR-RT-AS) were designed using Beacon Designer 4.0 software (PremierBiosoft International, Palo Alto, CA) and are detailed in Table 1.

RESULTS

Involvement of AguR in induction by pH and agmatine. In aprevious study using batch cultivation of cells, agmatine deimi-nase activity was optimal in cells growing in acidified mediumin the presence of agmatine (12). To assess whether AguRexerted transcriptional control over the agu operon in responseto pH and agmatine, the wild-type strain and strain R� carry-ing the aguB promoter (PaguB) fused to a Streptococcus saliva-rius lacZ gene were examined. The strains were cultured batch-wise in TV medium containing 25 mM galactose, with orwithout 10 mM agmatine. Cells were grown to mid-exponentialphase, and LacZ activity was measured. The expression ofPaguB was induced 2.5-fold by growth in the presence of 10 mMagmatine in the wild-type background, whereas the R� straindisplayed a basal level of PaguB expression and no agmatineinduction was evident (Fig. 2). WTZ and R� mutants were alsogrown to mid-exponential phase in TY medium that was acid-ified using HCl to pH 5.5 or buffered at pH 7.0 with 50 mMpotassium phosphate buffer, and LacZ activity was measured.In the wild-type background, cells expressed twofold higherLacZ activity from PaguB at pH 5.5 than at pH 7.0, whereas nosignificant difference in PaguB expression was detected in theR� strain at pH 5.5 and pH 7.0 (Fig. 3).

There are a number of limitations to the use of batch cul-tures for studies of the effect of pH, including the fact thatgrowth rates can be substantially different in cells growing atpH 5.5 and at pH 7.0, as can the amount of available carbo-

TABLE 2. S. mutans strains used in this study

Strain Relevantcharacteristic(s) Source or reference

UA159 Wild type University of Alabama,Birmingham

WTZ UA159/PaguB-lacZ This studyWTZ/V WTZ/pMSP3535 This studyWTZ/V-RK WTZ /pMSP-�nisRK This studyR� �aguR/PaguB-lacZ This studyR�/V R�/pMSP3535 This studyR�/V-RK R�/pMSP-�nisRK This studyR�/R R�/pMSP-aguR This studyAguD �aguD This study

VOL. 75, 2009 AGMATINE DEIMINASE REGULATION 2631

on Septem

ber 9, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

hydrate, which may affect agu gene expression through catab-olite repression (12). In contrast, continuous chemostat cultureallows for tight control over growth rates and the availability oflimiting nutrients. To test if there was indeed a specific effect ofpH that was independent of carbon catabolite repression orgrowth rate, WTZ and R� mutants were cultured in a chemo-stat at pH 7.0 and pH 5.5 to steady state, followed by a 10 mMagmatine pulse for 60 min. The results, shown in Table 3,revealed that the effects of agmatine and low pH on PaguB

expression were similar to those observed during batch culti-vation (Fig. 2 and 3), reinforcing the observation that AguR isrequired for activation of the agu operon by low pH andagmatine.

Generation of linker scanning mutants of AguR. To begin todefine structure-function relationships in AguR, linker scan-ning mutagenesis was used to introduce insertions of fiveamino acids randomly throughout the protein (Table 4). Atotal of 34 unique mutants were characterized, with 12 inser-tions located in the predicted transmembrane domains, 15 in

predicted extracellular domains, and 7 in predicted intracellu-lar domains, including 2 within the HTH motif (Fig. 1). Themutants (RLS) were named according to the amino acid num-ber preceding the site of insertion (Table 4). In addition, nineclones were characterized that contained nonsense codons thatwere predicted to lead to production of a truncated AguRprotein (Table 4). All of the mutant proteins were expressed ina host with a complete deletion of the aguR gene, as detailedin Materials and Methods.

To test whether the linker scanner insertions caused decreasesin the levels of the AguR derivatives expressed in the R� strain,possibly through decreased translation or stability of the alteredproteins, clarified cell lysates from the wild-type and mutantstrains were subjected to SDS-PAGE and Western blot analysiswith an affinity-purified polyclonal antibody raised against theC-terminal third of AguR (Fig. 4). A 37-kDa band, which matchesthe predicted molecular mass of AguR, was detected in the WTZstrain. A lower-molecular-mass species, which we believe may bea cleavage product of AguR, was consistently observed in the

FIG. 2. LacZ specific activities in AguR variants of S. mutans carrying PaguB-lacZ gene fusions. Cells were grown in TV medium containing 25mM galactose, with or without 10 mM agmatine, to mid-exponential phase, and LacZ specific activities were measured. The values shown representthe means and standard deviations for three independent experiments.

FIG. 3. LacZ specific activities of AguR variants of S. mutans carrying PaguB-lacZ. Cells were grown to mid-exponential phase in TY brothcontaining 25 mM galactose with 10 mM agmatine that had been acidified to pH 5.5 with HCl (TY-HCl) or in TY broth buffered at pH 7.0(TY-KPB), and the LacZ specific activities were measured. The values shown represent the means and standard deviations for three independentexperiments.

2632 LIU ET AL. APPL. ENVIRON. MICROBIOL.

on Septem

ber 9, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

Western blots. As expected, higher levels of AguR protein wereobserved in the R�/R strain (Fig. 4), in which aguR was carried onpMSMP3535, a multicopy plasmid that can drive gene expressionin gram-positive bacteria (4, 9, 14). The AguR protein could be

detected in all RLS mutants at levels similar to that in the R�/Rstrain (Fig. 4), which indicated that the linker scanning insertionsdid not have any major effects on AguR biogenesis or stability. Itwas also noteworthy that when protein samples from the WTZstrain were loaded in the same amount as that from the R�/Rstrain and RLS mutants, only a very faint band could be detected,which is consistent with our observations that AguR is expressedat extremely low levels in the wild-type strain. By adding fivefoldmore protein from the WTZ strain, AguR protein and the lower-molecular-weight species could be detected (Fig. 4C).

Regions required for agmatine induction. To begin to iden-tify domains in AguR that may be required for agmatine in-duction, strains were cultured to mid-exponential phase in TVmedium containing 25 mM galactose, with or without 10 mMagmatine, and LacZ activity was measured. As shown in Fig. 1and Table 4, the RLS 46, 52, 53, 60, 65, 66, 67, 81, 88, 104, 105,110, 168, 175, 209, 287, and 297 mutants did not respond to thepresence of agmatine with changes in expression from the aguBpromoter, similar to the strain lacking AguR (R�/V). The RLS41, 121, 158, 161, and 201 mutants showed around a 2.5-foldincrease in PaguB expression when they were grown with agma-tine, which was similar to the level of induction seen in thecomplemented (R�/R) strain (Fig. 1 and 2). Also, an increasein PaguB expression of about fivefold in response to growth inagmatine was observed for the RLS 73, 103, and 106 mutants(Fig. 1 and 2).

Regions of AguR required for low-pH induction. To deter-mine whether mutagenesis of different domains of AguRwould affect pH-dependent expression from the aguB pro-moter, the various strains were grown in TY broth containing10 mM agmatine that was acidified to pH 5.5 or buffered at pH7.0 (Fig. 1 and 3). Cells were harvested in mid-exponentialphase, and LacZ activity was measured. Unlike the strain car-rying a wild-type copy of the AguR protein, the RLS 41, 46, 52,53, 73, 81, 88, 168, and 175 mutants showed no difference inPaguB transcription at pH 5.5 or pH 7.0. An increase of about1.3-fold in PaguB transcription at low pH was detected in theRLS 60, 65, 66, and 67 mutants, compared with the 2-foldinduction seen at pH 5.5 in the R�/R strain. A twofold increasein PaguB expression at low pH was observed for the RLS 103,104, 105, 106, 110, 121, 158, 161, 201, and 209 mutants (Fig. 1and 3).

TABLE 3. LacZ activity expressed from PaguB in S. mutans strains growing in continuous chemostat culturec

Strain

LacZ activity (nmol/OD600/min)

Before agmatine pulse After agmatine pulse

pH 7.0 pH 5.5 pH 7.0 pH 5.5

WTZ/V 50.75 � 1.37 128.41 � 10.21b 99.05 � 9.17a 187.41 � 12.87a,b

R�/V 45.75 � 2.10 40.75 � 0.61 42.15 � 3.07 40.09 � 2.56R�/R 65.98 � 1.74 150.55 � 16.99b 117.75 � 10.37a 217.41 � 1.98a,b

RLS 41 61.04 � 8.61 82.91 � 19.52 169.21 � 11.52a 180.40 � 20.18a

RLS 46 58.64 � 4.65 51.91 � 11.52 59.51 � 4.65 76.74 � 17.90RLS 175 44.80 � 4.31 60.25 � 1.91b 40.56 � 2.92 62.09 � 1.96b

RLS 110 117.09 � 1.96 114.45 � 10.28 261.72 � 30.56a 212.60 � 4.26a

a The LacZ activities of the mutants cultured before the agmatine pulse differed significantly (P 0.05) from those of mutants cultured after agmatine pulsing atthe same environmental pH.

b The LacZ activities of the mutants cultured at pH 5.5 differed significantly (P 0.05) from those of mutants cultured at pH 7.0 at the same agmatine concentrations.c The WTZ/V, R�/V, and R�/R strains and various RLS strains of S. mutans were cultured in a chemostat in TY medium containing 25 mM galactose at a dilution

rate (D) of 0.3 h�1. Cultures were maintained at pH 5.5 or pH 7.0 and pulsed with 10 mM agmatine for 1 h. Cultures were sampled before or after agmatine pulsing,and LacZ activities were examined.

TABLE 4. Linker mutation locations, inserted residues, andresponsiveness of mutants to pH and agmatine in agu

operon induction

RLSmutanta

Residuesinsertedb Location

Responsiveness toc:

pH Agmatine

41 CLNTS Transmembrane domain 1 �� �46 CLNIF Transmembrane domain 1 �� ��52 CLNNS Transmembrane domain 1 �� ��53 CLNTV Transmembrane domain 1 �� ��60 CLNII Extracellular domain � ��65 LFKHN Extracellular domain � ��66 MFKHN Extracellular domain � ��67 LFKHN Extracellular domain � ��73 CLNTS Extracellular domain �� �81 FLFKH Extracellular domain �� ��88 LFKHF Extracellular domain �� ��102 V*TWW Extracellular domain �� ��103 CLNTW Extracellular domain � �104 VFKQW Extracellular domain � ��105 CLNRE Extracellular domain � ��106 CLNKY Extracellular domain � �110 LFKHL Transmembrane domain 2 � ��121 CLNNY Transmembrane domain 2 � �158 VFKQN Intracellular domain � �161 LFKHK Intracellular domain � �162 V*TLG Truncated protein �� ��168 CLNKC Transmembrane domain 3 �� ��174 V*TAM Truncated protein �� ��175 MFKHM Transmembrane domain 3 �� ��176 V*TMT Truncated protein �� ��194 V*TYS Truncated protein �� ��201 CLNKI Extracellular domain � �205 V*TNV Truncated protein �� ��209 VFKQE Transmembrane domain 4 � ��213 V*TII Truncated protein �� ��230 V*TLL Truncated protein �� ��286 V*TTV Truncated protein �� ��287 CLNNS HTH motif �� ��297 TNLET HTH motif �� ��

a The mutant number indicates the last amino acid number (relative to theinitiating amino acid of AguR) preceding the site of insertion.

b �, stop codon in the linker scanning sequence.c �, positive response; �, negative response.

VOL. 75, 2009 AGMATINE DEIMINASE REGULATION 2633

on Septem

ber 9, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

Chemostat studies with AguR variants. Since the effects ofthe mutated versions of aguR on aguB promoter expressionwere studied in batch culture, we examined whether a subset ofthe mutants did in fact have an aberrant response to low pHand agmatine in continuous culture. LacZ activity was mea-sured in cells that were cultivated to steady state at pH 7.0 orpH 5.5 and then pulsed with 10 mM agmatine for 60 min. TheRLS 110 strain expressed approximately threefold higher LacZactivity in response to agmatine, but no induction at pH 5.5 wasnoted (Table 3). Conversely, a twofold increase in PaguB ex-pression of RLS 41 was stimulated by the addition of 10 mMagmatine, but no difference in PaguB expression in this strainwas seen at pH 7.0 and pH 5.5 (Table 3). The RLS 46 and 175mutants, in which responses to low pH and agmatine wereeliminated, showed a basal level of PaguB expression, and noinduction by agmatine or low pH was detected (Table 3).Therefore, the results using chemostat-grown cells confirmedthat AguR is required for the response to low pH and agmatineand that mutations in different regions of the predicted mem-brane-spanning or extracellular domains can affect pH andagmatine induction independently or can eliminate both re-sponses.

Mutations in the HTH motif. AguR is a LuxR-like transcrip-tional regulator in which the C-terminal HTH domain isimplicated in DNA binding (28). However, the conserved acy-lated homoserine lactone binding region, typical of LuxR,could not be found in AguR. To confirm that the C-terminalHTH motif of AguR was involved in AguR-DNA binding, theRLS 102, 162, 174, 205, and 286 mutants, which contain trun-cated AguR proteins with termination codons before the HTHmotif (Table 2; Fig. 1), were analyzed. First, real-time PCR wasused to confirm that the truncated aguR genes could be ex-pressed at a level comparable to that of the parental copy, andthe data showed that the truncated aguR derivatives RLS 102,

162, 174, 205, and 286 were transcribed at a level similar to thatof wild-type aguR in the R�/R strain (data not shown). TheRLS 286 mutant was the only mutant in which the truncatedAguR protein had an overlapping region to the C-terminalone-third of AguR that was used to elicit the AguR antibody.By use of Western blot analysis, the truncated AguR protein inRLS 286 could be detected at levels comparable to that ofwild-type AguR in the R�/R strain (Fig. 4B). All of thesemutants showed a basal level of PaguB transcription, and noinduction by agmatine or low pH could be observed (data notshown), suggesting that the agu operon could not be activatedby inducing signals through AguR derivatives lacking the HTHmotif.

Involvement of AguD in regulation of the AgDS. AguD is apredicted agmatine-putrescine antiporter. Recently, CadC,which activates the transcription of the lysine decarboxylasegenes of E. coli in the presence of low pH and lysine, wasshown to sense the presence of extracellular lysine indirectlythrough the LysP permease. S. mutans UA159 and an AguD-deficient derivative were cultured in TV medium containing 25mM galactose, with or without 10 mM agmatine. Expression ofthe AgDS was monitored by real-time PCR with the aguA(AgD) gene as a target. The data (Fig. 5) showed that in theabsence of exogenous agmatine, aguA mRNA was about 100-fold more abundant in the aguD mutant than in the wild-typestrain. When the strains were grown in the presence of 10 mMagmatine, a slight but statistically significant increase (1.5-fold) in aguA expression was seen in the AguD-deficient strain(Fig. 5A). Thus, in the absence of exogenous agmatine, AguDacts as a negative regulator of agu gene expression.

To test if AguD could also play a role in induction of agugene transcription at low pH, the cells were cultured in TY-KPB or TY-HCl medium (pH 5.5) containing 25 mM galactoseand 10 mM agmatine. The aguA gene was expressed at similar

FIG. 4. Western blot analysis of wild-type, WTZ, R�/V, R�/R, and RLS mutant strains of S. mutans. Following SDS-PAGE, proteins weretransferred to a nitrocellulose membrane and subjected to Western blotting using an affinity-purified AguR antibody at a dilution of 1:100. Lanes aremarked with the strains examined. (A) Western analysis of various RLS insertion mutants and control strains (see Table 2). (B) The truncated derivativeRLS 286 (lane 3) expresses a protein of the correct predicted molecular mass at a level comparable to that of the intact AguR protein (lane 1). Lane2 contains an equivalent amount of protein from an aguR deletion mutant. (C) Western blot showing the presence of full-length AguR (37 kDa) andthe potential processed form of AguR expressed as a single copy in the wild-type background, except that fivefold more protein was loaded in that lane(5�WTZ) (lane 1). Lane 2 harbors a control showing no reactive bands for an AguR-deficient mutant. Lanes 3 and 4 harbor one-fifth the amount ofprotein in lanes 1 and 2, but extracts were prepared from strains expressing plasmid-borne derivatives of mutant aguR genes (see Table 4).

2634 LIU ET AL. APPL. ENVIRON. MICROBIOL.

on Septem

ber 9, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

levels in the AguD-deficient and wild-type strains growing ateither pH 7.0 or pH 5.5 (Fig. 5B), suggesting that AguD is notrequired for low-pH induction of the AgDS.

DISCUSSION

Induction of the AgDS at low pH and in the presence ofagmatine is believed to contribute to the competitive fitness ofS. mutans in complex oral biofilms (11). Here we have shownthat the AguR protein is required for responsiveness of agugene transcription to low pH and to agmatine. The predictedextracellular domain (Fig. 1) between transmembrane regions1 and 2 and the transmembrane domains themselves appear tobe essential for the response to both pH and agmatine. Notsurprisingly, the C-terminal HTH motif of AguR was abso-lutely required for activation of agu gene transcription. Inmany ways, the regulation of the AgDS of S. mutans is remi-niscent of that of the lysine decarboxylase (cadAB) operon of

E. coli (25), whose primary role appears to be to protect theorganism from acid damage (26). Induction of cad gene ex-pression occurs only in the presence of the inducer lysineunder acidic conditions and requires the transcriptional acti-vator CadC. CadC has a single transmembrane domain, and asubstantial portion of the protein resides outside the cytoplas-mic membrane. Conformational changes induced by low pHand binding of lysine are transduced to the N-terminal cyto-plasmic portion of the protein, which contains the HTH do-main (25). Recently, though, it was reported that sensing oflysine requires the LysP permease, which appears to interferewith the ability of CadC to activate transcription when externallysine concentrations are low (16).

The mutational analysis conducted here demonstrates that itis possible to isolate derivatives of AguR that lose the ability toactivate transcription in response to pH but not agmatine, andvice versa (Fig. 2 and 3). Likewise, we were able to isolatemutant forms of AguR that lost the ability to respond to eithersignal. The introduced mutations likely either disrupt a bindingsite or cause conformational changes that inhibit normal AguRfunction, since our data show that diminished production orstability of the proteins is not responsible for the observedeffects. The simplest interpretation of these findings is thatAguR can bind to agmatine, which stimulates AguR binding toits target sequence, and that acidic conditions favor a confor-mation for AguR that makes signal transduction to the DNAbinding domain more efficient. Such a model would be reason-able were it not for the fact that deletion of the aguD gene,which encodes the predicted agmatine-putrescine antiporter,dramatically altered agu gene expression in the absence ofagmatine. It seems, therefore, that the mechanism for regula-tion of agu gene expression by agmatine may be similar to whatwas recently reported for regulation of cadAB expression bylysine, CadC, and LysP. Specifically, the lack of exogenousagmatine may result in an association between AguD andAguR that interferes with the interaction of AguR with itstarget. The addition of exogenous agmatine could destabilizethis association. An alternative explanation is that internalizedagmatine is the actual inducing signal and that the loss of theantiporter inhibits accumulation of agmatine in cells. However,if the predicted topology and orientation of AguR are correct,then induction by intracellular agmatine seems unlikely sincethe mutations that affect induction by agmatine do not map tointracellular domains. Further analyses of the topology ofAguR and its ability to directly interact with agmatine areneeded to reach a firm conclusion. Notably, we have attemptedto produce a full-length AguR protein to explore the latterpossibility but have not been able to express the N-terminaltwo-thirds of the protein in E. coli. It is also noteworthy thatAguR is produced in S. mutans at extremely low levels, so ananalysis of the binding characteristics of AguR will require abreakthrough in expression of this protein in sufficient quan-tities to conduct the necessary biochemical studies.

A recent report suggested that the regulation of cadAB geneexpression by CadC may be even more complex than previ-ously appreciated. In particular, site-specific proteolysis of theperiplasmic domain of CadC, which is generated in cells ex-posed to acidic conditions, causes the release of a biologicallyactive cytoplasmic form of the N-terminal DNA binding do-main of CadC that can stimulate target gene activation (18). It

FIG. 5. Real-time PCR monitoring of aguA mRNA. Reverse tran-scription was initiated from 1 �g of total RNA from the wild-type(WT) and AguD-deficient strains of S. mutans cultured in TV mediumcontaining 25 mM galactose, with or without 10 mM agmatine (A), orin TY-KPB or TY-HCl containing 25 mM galactose and 10 mM ag-matine (B). The amount of aguA cDNA was determined by real-timePCR using SYBR green. The data represent means � standard devi-ations obtained from three different RNA preparations and reversetranscription reactions. Asterisks indicate that aguA gene expression inthe mutant differed significantly from that in the wild-type strain underthe same culture conditions (P 0.05; Student’s t test).

VOL. 75, 2009 AGMATINE DEIMINASE REGULATION 2635

on Septem

ber 9, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

is clear from our Western blots that a smaller species of AguRis commonly observed in strains expressing full-length wild-type or mutated AguR proteins. The observation that thislower-molecular-weight protein is indeed a derivative of AguRis supported by the fact that the protein is absent in strainslacking AguR and that all Western blotting was done with anantibody that was raised against a highly purified recombinantAguR protein that was expressed in E. coli. Furthermore, theantibodies used in the Western blots were affinity purifiedusing the purified recombinant AguR protein. There is oneopen reading frame (SMu0441) in S. mutans with limited ho-mology to the C-terminal DNA binding protein of AguR, butinactivation of the gene for that protein did not alter the resultsof Western blotting (data not shown). We are in the process ofdetermining where the potential cleavage of AguR occurs andwhether the processing is stimulated by growth under particu-lar conditions, e.g., low pH.

An unexpected finding that arose from this study was thataguB promoter activity in the presence of inducing concentra-tions of agmatine was altered in strains carrying only the vectorpMSP3535. Furthermore, deletion of the nisRK two-compo-nent system (TCS) from the vector eliminated this effect onagu expression (Fig. 2). NisRK is a TCS that senses environ-mental nisin and activates nisA promoter expression by using aclassical TCS phosphorelay circuit (15). Notably, we have seenan effect of NisRK on one other unrelated gene in S. mutans(L. Zeng and R. Burne, unpublished data). While this obser-vation may serve primarily as a cautionary note on the use ofpMSP3535 for certain gene regulation studies, the finding mayhave some importance in terms of agu gene expression. Inparticular, the regulation of certain genes by pH, as well as thegeneral property of acid tolerance in S. mutans, has beenshown to be influenced in strains lacking a number of differentTCS constituents, including CovR, CiaRH, VicRK, andComDE (2, 22, 24, 27). The influence of heterologous expres-sion of a TCS in S. mutans, in this case NisRK, that maypotentially cross talk with endogenous TCSs raises the possi-bility that pH-dependent expression of the agu operon couldbe, at least in part, controlled by a TCS. Efforts are ongoing toprobe in more detail the basis for low-pH-dependent enhance-ment of agu gene expression.

Elucidation of the genetics, physiology, and biochemistry ofagmatine catabolism is an important first step in understandingthe role of the agmatine deiminase in oral biofilm ecology anddisease. The AgDS conveys bioenergetic advantages to S. mu-tans and other oral streptococci through enhancement of the�pH and through generation of ATP (12), so it should con-tribute in major ways to the persistence and virulence of theseorganisms. Moreover, the complexity of regulation of theAgDS by substrate, catabolite control, and relevant environ-mental stresses, including pH, along with the relatively broaddistribution of the system in pathogens and commensals,makes this an excellent model for the study of gene regulationand physiology in streptococci.

ACKNOWLEDGMENTS

We thank Ann R. Griswold for providing an AguD-deficient strainof S. mutans.

This work was supported by Public Health Service grant DE10362from the National Institute of Dental and Craniofacial Research.

REFERENCES

1. Ahn, S. J., J. A. Lemos, and R. A. Burne. 2005. Role of HtrA in growthand competence of Streptococcus mutans UA159. J. Bacteriol. 187:3028–3038.

2. Ahn, S. J., Z. T. Wen, and R. A. Burne. 2006. Multilevel control of compe-tence development and stress tolerance in Streptococcus mutans UA159.Infect. Immun. 74:1631–1642.

3. Bender, G. R., S. V. Sutton, and R. E. Marquis. 1986. Acid tolerance, protonpermeabilities, and membrane ATPases of oral streptococci. Infect. Immun.53:331–338.

4. Bryan, E. M., T. Bae, M. Kleerebezem, and G. M. Dunny. 2000. Improvedvectors for nisin-controlled expression in gram-positive bacteria. Plasmid44:183–190.

5. Burne, R. A., Z. T. Wen, Y. Y. Chen, and J. E. Penders. 1999. Regulation ofexpression of the fructan hydrolase gene of Streptococcus mutans GS-5 byinduction and carbon catabolite repression. J. Bacteriol. 181:2863–2871.

6. Chen, Y. Y., M. J. Betzenhauser, and R. A. Burne. 2002. cis-Acting elementsthat regulate the low-pH-inducible urease operon of Streptococcus salivarius.Microbiology 148:3599–3608.

7. Chen, Y. Y., and R. A. Burne. 1996. Analysis of Streptococcus salivariusurease expression using continuous chemostat culture. FEMS Microbiol.Lett. 135:223–229.

8. Chen, Y. Y., C. A. Weaver, D. R. Mendelsohn, and R. A. Burne. 1998.Transcriptional regulation of the Streptococcus salivarius 57.I urease operon.J. Bacteriol. 180:5769–5775.

9. Eichenbaum, Z., M. J. Federle, D. Marra, W. M. de Vos, O. P. Kuipers, M.Kleerebezem, and J. R. Scott. 1998. Use of the lactococcal nisA promoter toregulate gene expression in gram-positive bacteria: comparison of inductionlevel and promoter strength. Appl. Environ. Microbiol. 64:2763–2769.

10. Fuqua, C., M. R. Parsek, and E. P. Greenberg. 2001. Regulation of geneexpression by cell-to-cell communication: acyl-homoserine lactone quorumsensing. Annu. Rev. Genet. 35:439–468.

11. Griswold, A. R., Y. Y. Chen, and R. A. Burne. 2004. Analysis of an agmatinedeiminase gene cluster in Streptococcus mutans UA159. J. Bacteriol. 186:1902–1904.

12. Griswold, A. R., M. Jameson-Lee, and R. A. Burne. 2006. Regulation andphysiologic significance of the agmatine deiminase system of Streptococcusmutans UA159. J. Bacteriol. 188:834–841.

13. Griswold, A. R., M. N. Nascimento, and R. A. Burne. 2009. Distribution,regulation and role of the agmatine deiminase system in mutans streptococci.Oral Microbiol. Immunol. 24:79–82.

14. Kleerebezem, M., M. M. Beerthuyzen, E. E. Vaughan, W. M. de Vos, andO. P. Kuipers. 1997. Controlled gene expression systems for lactic acidbacteria: transferable nisin-inducible expression cassettes for Lactococ-cus, Leuconostoc, and Lactobacillus spp. Appl. Environ. Microbiol. 63:4581–4584.

15. Kuipers, O. P., M. M. Beerthuyzen, P. G. de Ruyter, E. J. Luesink, andW. M. de Vos. 1995. Autoregulation of nisin biosynthesis in Lactococcuslactis by signal transduction. J. Biol. Chem. 270:27299–27304.

16. Kuper, C., and K. Jung. 2005. CadC-mediated activation of the cadBApromoter in Escherichia coli. J. Mol. Microbiol. Biotechnol. 10:26–39.

17. Kuramitsu, H. K. 1993. Virulence factors of mutans streptococci: role ofmolecular genetics. Crit. Rev. Oral Biol. Med. 4:159–176.

18. Lee, Y. H., J. H. Kim, I. S. Bang, and Y. K. Park. 2008. The membrane-boundtranscriptional regulator CadC is activated by proteolytic cleavage in re-sponse to acid stress. J. Bacteriol. 190:5120–5126.

19. Lemos, J. A., J. Abranches, and R. A. Burne. 2005. Responses of cariogenicstreptococci to environmental stresses. Curr. Issues Mol. Biol. 7:95–107.

20. Lemos, J. A., T. A. Brown, Jr., and R. A. Burne. 2004. Effects of RelA on keyvirulence properties of planktonic and biofilm populations of Streptococcusmutans. Infect. Immun. 72:1431–1440.

21. Lemos, J. A., and R. A. Burne. 2002. Regulation and physiological signifi-cance of ClpC and ClpP in Streptococcus mutans. J. Bacteriol. 184:6357–6366.

22. Levesque, C. M., R. W. Mair, J. A. Perry, P. C. Lau, Y. H. Li, and D. G.Cvitkovitch. 2007. Systemic inactivation and phenotypic characterization oftwo-component systems in expression of Streptococcus mutans virulenceproperties. Lett. Appl. Microbiol. 45:398–404.

23. Levin, P. 2002. Light microscopy techniques for bacterial cell biology. Meth-ods Microbiol. 31:115–132.

24. Li, Y. H., P. C. Lau, N. Tang, G. Svensater, R. P. Ellen, and D. G. Cvitko-vitch. 2002. Novel two-component regulatory system involved in biofilmformation and acid resistance in Streptococcus mutans. J. Bacteriol. 184:6333–6342.

25. Neely, M. N., C. L. Dell, and E. R. Olson. 1994. Roles of LysP and CadC inmediating the lysine requirement for acid induction of the Escherichia colicad operon. J. Bacteriol. 176:3278–3285.

26. Park, Y. K., B. Bearson, S. H. Bang, I. S. Bang, and J. W. Foster. 1996.Internal pH crisis, lysine decarboxylase and the acid tolerance response ofSalmonella typhimurium. Mol. Microbiol. 20:605–611.

27. Senadheera, M. D., B. Guggenheim, G. A. Spatafora, Y. C. Huang, J. Choi,

2636 LIU ET AL. APPL. ENVIRON. MICROBIOL.

on Septem

ber 9, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from

D. C. Hung, J. S. Treglown, S. D. Goodman, R. P. Ellen, and D. G. Cvitko-vitch. 2005. A VicRK signal transduction system in Streptococcus mutansaffects gtfBCD, gbpB, and ftf expression, biofilm formation, and geneticcompetence development. J. Bacteriol. 187:4064–4076.

28. Slock, J., D. VanRiet, D. Kolibachuk, and E. P. Greenberg. 1990. Criticalregions of the Vibrio fischeri LuxR protein defined by mutational analysis. J.Bacteriol. 172:3974–3979.

29. Wexler, D. L., M. C. Hudson, and R. A. Burne. 1993. Streptococcus mutans

fructosyltransferase (ftf) and glucosyltransferase (gtfBC) operon fusionstrains in continuous culture. Infect. Immun. 61:1259–1267.

30. Zeng, L., Z. T. Wen, and R. A. Burne. 2006. A novel signal transductionsystem and feedback loop regulate fructan hydrolase gene expression inStreptococcus mutans. Mol. Microbiol. 62:187–200.

31. Zubay, G., D. E. Morse, W. J. Schrenk, and J. H. Miller. 1972. Detection andisolation of the repressor protein for the tryptophan operon of Escherichiacoli. Proc. Natl. Acad. Sci. USA 69:1100–1103.

VOL. 75, 2009 AGMATINE DEIMINASE REGULATION 2637

on Septem

ber 9, 2018 by guesthttp://aem

.asm.org/

Dow

nloaded from