JOURNAL OF BACTERIOLOGY, Nov. 2002, p. 6333–6342
Copyright © 2002, American Society for Microbiology. All Rights Reserved.
Vol. 184, No. 22
Novel Two-Component Regulatory System Involved in Biofilm
Formation and Acid Resistance in Streptococcus mutans
Yung-Hua Li,1Peter C. Y. Lau,1Nan Tang,1Gunnel Svensa ¨ter,2Richard P. Ellen,1and
Dennis G. Cvitkovitch1*
Dental Research Institute, University of Toronto, Toronto, Ontario, Canada M5G 1G6,1and Department of
Oral Microbiology, Malmo ¨ University, S-21421 Malmo ¨, Sweden2
Received 22 April 2002 /Accepted 31 July 2002
The abilities of Streptococcus mutans to form biofilms and to survive acidic pH are regarded as two important
virulence determinants in the pathogenesis of dental caries. Environmental stimuli are thought to regulate the
expression of several genes associated with virulence factors through the activity of two-component signal
transduction systems. Yet, little is known of the involvement of these systems in the physiology and pathoge-
nicity of S. mutans. In this study, we describe a two-component regulatory system and its involvement in biofilm
formation and acid resistance in S. mutans. By searching the S. mutans genome database with tblastn with the
HK03 and RR03 protein sequences from S. pneumoniae as queries, we identified two genes, designated hk11 and
rr11, that encode a putative histidine kinase and its cognate response regulator. To gain insight into their
function, a PCR-mediated allelic-exchange mutagenesis strategy was used to create the hk11 (Emr) and rr11
(Emr) deletion mutants from S. mutans wild-type NG8 named SMHK11 and SMRR11, respectively. The
mutants were examined for their growth rates, genetic competence, ability to form biofilms, and resistance to
low-pH challenge. The results showed that deletion of hk11 or rr11 resulted in defects in biofilm formation and
resistance to acidic pH. Both mutants formed biofilms with reduced biomass (50 to 70% of the density of the
parent strain). Scanning electron microscopy revealed that the biofilms formed by the mutants had sponge-like
architecture with what appeared to be large gaps that resembled water channel-like structures. The mutant
biofilms were composed of longer chains of cells than those of the parent biofilm. Deletion of hk11 also resulted
in greatly diminished resistance to low pH, although we did not observe the same effect when rr11 was deleted.
Genetic competence was not affected in either mutant. The results suggested that the gene product of hk11 in
S. mutans might act as a pH sensor that could cross talk with one or more response regulators. We conclude
that the two-component signal transduction system encoded by hk11 and rr11 represents a new regulatory
system involved in biofilm formation and acid resistance in S. mutans.
Two-component signal transduction systems (TCSTSs) func-
tion in bacterial adaptation, survival, and virulence by sensing
changes in the environment and modulating gene expression in
response to a variety of stimuli (12). A typical two-component
regulatory system consists of a membrane-associated, histidine
kinase sensor protein, which senses a specific environmental
condition, and a cytoplasmic response regulator, which enables
the cell to respond via regulation of gene expression when this
condition varies (29). Upon stimulation by a specific ligand or
a signal the histidine kinase sensor protein undergoes auto-
phosphorylation at a conserved histidine residue. The phos-
phoryl group is then transferred to the cognate response reg-
ulator, which can, in turn, activate or repress transcription of
target genes. Two-component regulatory systems have been
shown elsewhere to regulate diverse metabolic processes, the
bacterial cell cycle, cell-cell communication, and virulence fac-
tors in a wide range of bacterial species (8). Because of their
importance in the regulation of cellular physiology, adaptation
to environments, and virulence expression, two-component
regulatory systems have been used as targets to develop anti-
microbial agents (14, 24).
Streptococcus mutans is a bacterium that has evolved a bio-
film lifestyle for survival and persistence in its natural ecosys-
tem, dental plaque (11). Under appropriate environmental
conditions, S. mutans can produce sufficient amounts of acid
from dietary fermentable carbohydrate and cause an imbal-
ance in the demineralization-remineralization process of tooth
enamel, leading to dental caries (23). The ability of S. mutans
to initiate dental caries depends on several virulence-associ-
ated traits, including (i) initiation of biofilm formation by
adherence and accumulation on the tooth surface that is pro-
moted by its synthesis of insoluble, extracellular polysaccha-
rides; (ii) high efficiency in catabolizing carbohydrates and
producing acids; and (iii) the ability to grow and continue to
metabolize carbohydrates at low pH (25). Environmental fac-
tors play important roles in the regulation of these virulence-
associated traits in S. mutans (22). Despite many studies dem-
onstrating the importance of environmental stimuli in the
regulation of physiology and virulence traits of S. mutans, little
is known of the molecular mechanisms by which S. mutans
regulates the expression of these virulence traits in response to
fluctuations in its environment.
We have recently characterized a quorum sensing signaling
system consisting of a two-component regulatory system
(ComDE) in S. mutans. This system responds to its native
signal peptide pheromone and activates transcription of a
number of genes essential for induction of genetic competence,
* Corresponding author. Mailing address: Rm. 449A, Dental Re-
search Institute, University of Toronto, 124 Edward St., Toronto, On-
tario, Canada M5G 1G6. Phone: (416) 979-4917, ext. 4592. Fax: (416)
979-4936. E-mail: firstname.lastname@example.org.
resulting in natural transformation (19). Our previous work
has demonstrated that this system appears to play a global
regulatory role in genetic competence, biofilm formation, and
acid tolerance response (ATR) in S. mutans (18, 20). In this
study, we described a novel two-component regulatory system
and began to evaluate the role of this system in biofilm forma-
tion and acid resistance in S. mutans.
MATERIALS AND METHODS
Bacterial strains, media, and chemicals. The strains used in this study and
their relevant characteristics are listed in Table 1. S. mutans wild-type (wt) strain
NG8 was subcultured routinely on Todd-Hewitt yeast extract (THYE) agar
plates (BBL Becton Dickinson, Cockeysville, Md.), whereas the mutants were
maintained on THYE agar plus 10 ?g of erythromycin/ml. THYE liquid medium
was routinely used to grow the strains unless otherwise specified. To grow
biofilms, a semidefined minimal (SDM) medium was prepared by a modification
of the method described previously (21). The medium contained 58 mM
K2HPO4, 15 mM KH2PO4, 10 mM (NH4)2SO4, 35 mM NaCl, and 2 mM
MgSO2? 7H2O and was supplemented with filter-sterilized vitamins (0.04 mM
nicotinic acid, 0.1 mM pyridoxine HCl, 0.01 mM pantothenic acid, 1 ?M ribo-
flavin, 0.3 ?M thiamine HCl, and 0.05 ?M D-biotin), amino acids (4 mM L-
glutamic acid, 1 mM L-arginine HCl, 1.3 mM L-cysteine HCl, and 0.1 mM
L-tryptophan), 0.2% (wt/vol) Casamino Acids, and 20 mM glucose. Biofilms of all
strains were developed on polystyrene microtiter plates in SDM medium at 37°C
with 5% CO2for 16 h before quantification and microscopic examination.
Construction of the hk11 and rr11 deletion mutants. We initiated a search of
the S. mutans genome database at the University of Oklahoma OU-ACGT
website (http://www.genome.ou.edu/smutans.html) (P. C. Y. Lam and D. G.
Cvitkovitch, abstract, J. Dent. Res. 81:2246, 2002) for homologs of the 13 TC-
STSs identified in S. pneumoniae (14). A tblastn search using the HK03 and
RR03 protein sequences from Streptococcus pneumoniae (14) as queries identi-
fied two genes that shared homology with the hk3 and rr3 genes in S. pneumoniae.
These two genes, designated hk11 and rr11, respectively encoded a putative
histidine kinase and its cognate response regulator in S. mutans. This study
focused on the evaluation of the function of this TCSTS, designated HK/RR11,
in biofilm formation and acid resistance of S. mutans. We constructed individual
deletion mutants of the hk11 and rr11 genes in S. mutans wt strain NG8 by a rapid
PCR-based deletion strategy involving restriction-ligation and allelic replace-
ment as described previously (15). The primers used to construct and confirm the
gene deletion are listed in Table 2. To construct the hk11 mutant, for example,
a 763-bp fragment 5? from the hk11 start codon (HK11-up) was amplified from
S. mutans NG8 genomic DNA by using primers HK11-P1 and HK11-P2 (con-
taining an AscI site at its 5? end). Another amplicon, designated HK11-dw, was
666 bp 3? from hk11 and was amplified with HK11-P3 (with an FseI site at the 5?
end) and HK11-P4 primers. An erythromycin resistance marker, PcEm (860 bp),
from a synthetic Emrcassette (4) was amplified by using Em cst-P1 and Em
cst-P2 primers with AscI and FseI sites engineered into their 5? ends, respectively.
These amplicons were subjected to restriction enzyme digestion and subsequent
ligation to produce an HK11-up::PcEm::HK11-dw fragment. The ligated product
was directly used for transformation of S. mutans wt strain NG8 with the aid of
a synthetic competence-stimulating peptide (CSP) (19). Following double-cross-
over homologous recombination, the internal region of the hk11 gene was com-
pletely replaced by the erythromycin cassette (PcEm). A similar strategy was
used to construct the rr11 deletion mutant.
The integration sites of the PCR constructs in the mutants were confirmed by
PCR. Briefly, genomic DNA was prepared from transformants selected on
THYE-erythromycin (10 ?g/ml) agar plates by a method described previously
(6). The mutant and wt genomic DNAs were then used as templates in PCR with
three combinations of primers (P1 and Em cst-P2, P4 and Em cst-P1, and P1 and
P4) to verify correct recombination of the construct into the mutant genome
based on the predicted size of the products. The wt (NG8) genomic DNA was
used as a negative control.
Growth rates. Strains were grown in both SDM medium and a tryptone-yeast
extract (TYE) medium supplemented with 20 mM glucose to assay their growth
kinetics with a Bioscreen microbiology reader (Bioscreen C Labsystems, Hel-
sinki, Finland) with multiwell disposable microtiter plates. The Bioscreen was
equipped with software that allowed recording and conversion of turbidity read-
ings into growth curves. An aliquot (4 ?l) of cell suspension of the same turbidity
was inoculated into each well containing 400 ?l of fresh medium. Turbidity of the
culture was recorded after brief shaking every 15 min for a total of 20 h. Each
TABLE 1. Bacterial strains, amplicons, and plasmid used in the study
StrainsRelevant characteristics Source or reference
S. mutans strains
NG8 ?hk11::PcEm Emr
NG8 ?rr11::PcEm Emr
A. S. Bleiweis, University of Florida
Emrmarker with a synthetic promoter amplified from ermAM cassette
HK11-up::PcEm::HK11-dw for allelic replacement of hk11
RR11-up::PcEm::RR11-dw for allelic replacement of rr11
pDL289E. coli-Streptococcus shuttle vector; Kmr
TABLE 2. Primers used to construct the hk11 and rr11 deletion mutants by PCR restriction-ligation mutagenesis
Primer Nucleotide sequence (5? 3 3?)a
aAscI sites are in boldface; FseI sites are underlined.
6334LI ET AL.J. BACTERIOL.
sample was assayed in triplicate, and three wells without cells were used as blank
Genetic transformation. To determine if inactivation of hk11 or rr11 had any
impact on the development of genetic competence, the mutants were assayed for
genetic transformation by using a protocol as described previously (19). Briefly,
overnight cultures were diluted with 2 ml of prewarmed, fresh THYE broth
supplemented with 5% horse serum to generate 1:20 and 1:40 dilutions. The
cultures were incubated at 37°C with 5% CO2for 2 h to allow turbidities to reach
1.5 to 2.0 units of optical density at 600 nm. Each sample was then divided into
two aliquots: one containing 1 ?g of transforming plasmid DNA (pDL289,
Kmr)/ml (2) and another containing the same concentration of transforming
plasmid DNA and freshly made CSP (19) at a final concentration of 500 ng/ml.
The cultures were incubated for 2 to 3 h and gently sonicated for 10 s to disperse
the streptococcal chains, and an aliquot (100 ?l) of cell suspension was spread on
THYE plates containing kanamycin (700 ng/ml). An aliquot of the cell suspen-
sion, after appropriate dilution, was also spread on THYE plates without anti-
biotics to determine the total recipient cell number. Transformation of the
parent strain NG8 was used as a positive control. Transformation frequency was
expressed as the number of transformants divided by the total recipient cells per
milliliter of cell suspension.
Biofilm formation and quantification. All strains were assayed for biofilm
formation on a polystyrene surface by the method described previously (20, 21).
To facilitate quantification and microscopy, both 96- and 24-well polystyrene
microtiter plates were used to develop biofilms. The growth of biofilms was
initiated by inoculating 5 ?l of suspended cells from an overnight culture into 300
?l of SDM medium in individual wells of a 96-well microtiter plate or 25 ?l of
cell suspension into 2 ml of SDM medium in 24-well plates. The microtiter plates
were then incubated at 37°C with 5% CO2for 16 h without agitation. After
incubation, liquid medium was removed and wells were rinsed once with sterile
distilled water. The plates (96 wells) were then air dried and stained with 0.1%
(wt/vol) safranin for 10 min. After staining, the plates were rinsed with distilled
water to remove excess dye and air dried for 3 h. Biofilms were quantified by
measuring the absorbance of stained biofilms at 490 nm with an enzyme-linked
immunosorbent assay microplate reader (model 3550; Bio-Rad Laboratories,
Richmond, Calif.). Each assay was performed in triplicate, and wells without
biofilms were used as blank controls after safranin staining. Biofilms formed in
24-well plates were photographed immediately after removal of planktonic cells
Adherence assay. The strains were assayed for their ability to attach to a
mucin-coated polystyrene surface to determine the effect of inactivation of in-
dividual genes on initial adherence. The surface of the polystyrene microtiter
plates was first conditioned with 2 ml of 1% (wt/vol) hog gastric mucin (type III;
Sigma) in an adherence buffer (10 mM KPO4, 50 mM KCl, 1 mM CaCl2, 0.1 mM
MgCl2, pH 7.0) (17). The plates were incubated at room temperature for 2 h with
gentle shaking and air dried after removal of excess mucin solution. Adherence
was then initiated by addition of 2 ml of a previously prepared resting cell
suspension at a density of 108cells/ml. The resting cells were prepared by
centrifugation of overnight cultures, washed twice, and resuspended in adher-
ence buffer. The plates were incubated at 37°C with gentle shaking for 2 h. After
incubation, unattached cells were removed and adherent cells were dissociated
into 2 ml of the buffer by gentle sonication. Viable colony counts of both
adherent and nonadherent cells were performed to determine percentages of
Acid tolerance assays. The effect of pH on the growth of the hk11 and rr11
deletion mutants was first evaluated by assessment of growth on THYE agar
plates at pH 5.0 and 7.0. Both the mutants and the parent strains were grown in
THYE broth (pH 7.0) overnight. One volume of overnight culture was trans-
ferred into 9 volumes of fresh medium, and incubation continued for 2 h at 37°C
in an atmosphere of 5% CO2. The cultures were gently sonicated for 15 s to
disperse the chains of cells prior to serial dilution with 10 mM KPO4buffer (pH
7.2). An aliquot (20 ?l) of cell suspension from each strain was inoculated onto
THYE agar plates at both pH 5.0 and pH 7.0. The plates were then incubated at
37°C in an atmosphere of 5% CO2for 40 h before assessment of acid sensitivity.
Sensitivity to low pH was determined by comparison of the growth of parent and
mutants on THYE plates at pH 5.0 following a 40-h incubation.
The cultures were also grown in broth to assay the inducible ATR by a method
described previously (18). All experiments for ATR were carried out in TYE
medium supplemented with 20 mM glucose (TYEG) at pH 7.5, 5.5, and 3.5
prepared with 40 mM phosphate-citrate buffer. Briefly, mid-log-phase cells were
prepared by transferring 1 volume of overnight culture into 9 volumes (1:10) of
fresh TYEG (pH 7.5) and incubated at 37°C in an atmosphere of 5% CO2for 2 h.
These cells were collected by centrifugation at 10,000 ? g for 10 min and
resuspended in 2 ml of fresh TYEG (pH 5.5) at a turbidity of 0.6 (A600). The cells
were induced for acid adaptation by incubation at 37°C with 5% CO2for 2 h. The
adapted log-phase cells were then exposed to the killing pH of 3.5, which was
predetermined by incubating unadapted, mid-log-phase cells in TYEG at pH
values from 6.0 to 2.0 for 3 h (18). An aliquot of cell suspension was taken
immediately from each sample to determine total viable cell number at zero
time, and the cultures were incubated at 37°C with 5% CO2for 3 h. After
incubation an aliquot of the cell suspensions was taken to determine the per-
centage of survivors by viable cell counts. The ATR was expressed as the per-
centage of cells to survive the killing pH for 3 h.
14C labeling of cells during acid adaptation. Changes in protein expression of
NG8 and SMHK11 during acid adaptation were assessed by exposing cells to
14C-labeled amino acids followed by protein extraction and separation by two-
dimensional (2D) gel electrophoresis. Three independent cultures of NG8 and
SMHK11 were grown in a minimal medium comprised of six amino acids (glu-
tamate, serine, cysteine, valine, leucine, and asparagine), 40 mM phosphate-
citrate buffer (pH 7.5), and 20 mM glucose (1). Cells were grown to the middle
of exponential growth phase (optical density at 600 nm ? 0.7), washed twice in
glucose- and buffer-free medium, and resuspended to 2 ? 108cells ml?1in 2.5
ml of fresh minimal medium. The triplicate cultures of each strain were divided
into two portions, where one was exposed to pH 5.5 and the other was kept at pH
7.5 in the presence of 150 ?Ci of a14C-amino acid mixture. The incubation was
carried out for 30 min at 37°C, and protein synthesis was stopped by adding 2 mg
of chloramphenicol to each tube. Cells were centrifuged (15,000 ? g for 10 min)
and washed in 10 mM Tris-HCl, pH 6.8, with 1 mM EDTA and 5 mM MgSO4.
The cells were stored at ?20°C, and cell protein extracts for 2D gel electro-
phoresis were prepared by using ultrasonication in the presence of glass beads as
previously described (30).
2D gel electrophoresis and image analysis of protein patterns. 2D gel elec-
trophoresis and image analysis of autoradiograms were performed by previously
described methods (30). Isoelectric focusing in the first dimension was carried
out on linear pH 4 to 7 18-cm immobilized pH gradient gel strips (Pharmacia
Biotechnology, Uppsala, Sweden) loaded with 106cpm, corresponding to 150 ?g
of cellular protein. The second dimensional separation was performed with 14%
polyacrylamide gradient gels (185 by 200 by 1.0 mm), and the dried gels were
exposed to X-ray film (Hyperfilm ?-max; Amersham, Oakville, Ontario, Canada)
for 14 days. Proteins visualized on the autoradiograms were analyzed with the
Bio Image software (version 6.1) on a Sun Sparc station. A protein spot was
classified as being differently expressed if the relative integrated optical intensity
was changed more than twofold in the acid-exposed cells (pH 5.5) compared to
the control cells (pH 7.5). Three independent experiments were performed; for
each spot, a coefficient of variation was calculated; and those proteins that
exhibited a high inherent variation in expression were excluded as being acid
Microscopy. To examine the spatial distribution and architecture of biofilms by
scanning electron microscopy (SEM), biofilms formed on the surface of polysty-
rene microtiter plates were washed once with 10 mM phosphate-buffered saline,
fixed by adding 2 ml of 3.7% formaldehyde in 10 mM phosphate-buffered saline,
and incubated at room temperature for 24 h. The samples were then dehydrated
through a series of ethanol rinses (30, 50, 70, 95, and 100%) and critical point
dried with liquid CO2. The bottom surface of the well was cut off, mounted, and
sputter coated with gold. The samples were then examined by SEM (model
S-2500; Hitachi Instruments, San Jose, Calif.).
We previously observed that the CSP encoded by the comC gene in S. mutans
activated an uncharacterized second pathway that appeared to be related to cell
separation or chain formation (20). The second pathway activated by the CSP
remains to be identified. To test if the TCSTS encoded by hk/rr11 was potentially
the second pathway, we compared the length of chains formed by the hk/rr11
mutants grown in biofilms with or without addition of CSP, by using light
microscopy. Briefly, the mutant biofilms were developed in microtiter plates by
the same method as described previously (20), with the exception that each well
contained a sterilized coverslip as a substrate and cultures were supplemented
with 1.0 ?g of fresh CSP/ml. Biofilms were grown in the SDM medium for 16 h,
and liquid was removed. The biofilms were then stained with 0.1% crystal violet
for 1 min before being placed on a microscope slide. Biofilms were then viewed
and qualitatively assessed for cell chain length by light microscopy (Olympus
CH30RF100; Tokyo, Japan).
Genetic confirmation of the hk11 and rr11 deletion mutants.
The genetic locus surrounding the hk/rr11 region was anno-
tated by comparing the deduced amino acid sequences of the
VOL. 184, 2002TWO-COMPONENT SYSTEM INVOLVED IN BIOFILM FORMATION6335
adjacent open reading frames (ORFs) to the GenBank data-
base by using the blastP algorithm. A map and description of
the locus are shown in Fig. 1. The hk11 gene was located at bp
455345 to 456349 in the S. mutans genome database, encoding
a hypothetical protein of 334 amino acids with a predicted
molecular mass of 38,113 Da. The hk11 ORF shared highest
similarity to a putative two-component sensor histidine kinase
from Streptococcus pyogenes (accession no. AAK34394) (blast
similarity score ? 352); a putative histidine kinase, HK03, from
S. pneumoniae (CAB54570) (blast score ? 244); and a pu-
tative histidine kinase, BH1199, from Bacillus halodurans
(BAB04918) (blast score ? 177). The ORF encoding rr11 was
located at bp 456336 to 456983, encoding a hypothetical pro-
tein of 215 amino acids having a predicted molecular mass of
24,067 Da. Interestingly the genes overlapped by 12 nucleo-
tides and the response regulator-encoding gene had a promot-
er-like structure located 5? from its putative start codon. In this
5?-proximal region the ?18 to ?10 sequence TACCAACT was
very similar to the com-box consensus of S. pneumoniae (16) by
a single base pair (TACGAACT). These com-box genes form
part of the CSP-mediated regulon. The hk11 gene had putative
?10 (?12, TAATGA) and ?35 (TGTTATGGA) promoter
sequences as well. It did not, however, appear to have a com-
box in this vicinity. Similarly to the hk11 gene, rr11 had the
FIG. 1. The arrangement of the hk/rr11 genetic locus. Neighboring genes were assigned putative functions based on high blast homology scores
with genes for the indicated proteins: 1, primosomal replication factor V (AAK34400); 2, methionyl tRNA transferase Fmt (AAK34399); 3, RNA
binding protein SunL (rRNA methyltransferase RsmB) (AAK34398); 4, phosphoprotein Ser/Thr phosphatase PppL (AAK34397); 5, Ser/Thr
protein kinase PknB (AAK34396); 6, conserved hypothetical protein (AAK34395); 7, peptidyl-prolyl cis-trans isomerase PpiB (AAK75626); 8,
polyribonucleotide nucleotidyltransferase (general stress protein GSP13) (AAK34392); 9, pyruvate formate lyase-activating enzyme PflC
(AAK74424); 10, transcriptional regulator RdrA (AAK34719); 11, pyruvate formate lyase 2 PflD (AAK34714); 12, transaldolase-like protein MipB
(AAK34713); and 13, glycerol dehydrogenase GldA (AAK74432).
FIG. 2. Growth curves of the parent strain S. mutans NG8 and hk/rr11 mutants SMHK11 and SMRR11 grown in TYEG medium at pH 7.0 and
6336 LI ET AL.J. BACTERIOL.
highest homology to the respective response regulator gene
from S. pyogenes (SPy1621) and the R03 gene from S. pneu-
moniae but had a higher similarity to hypothetical response
regulator gene yvqC from Bacillus subtilis rather than B. halo-
durans. No putative substrate or signal has been assigned to
any of these systems in these organisms.
PCR confirmation demonstrated that the target genes of the
hk11 and rr11 mutants were correctly replaced by the Erm
cassette in their respective mutants (data not shown). The
mutants confirmed by PCR were designated SMHK11 and
SMRR11, respectively. Growth kinetics showed that both mu-
tants had a decrease in growth rate (increased doubling time
[Td]) in SDM medium (Td? 1.47 h for SMHK11 and 1.43 h for
SMRR11) compared to the parent strain (Td? 1.27 h). How-
ever, the final growth yield of the mutants after 12 h of growth
appeared to be the same as that of the parent strain in TYEG
(Fig. 2) or SDM medium (data not shown). Both mutants,
similar to the parent strain, were able to become genetically
competent and were transformed with plasmid DNA with or
without addition of CSP, suggesting that the system encoded
by hk/rr11 did not affect competence development in this or-
Deletion of hk11 or rr11 resulted in defects in biofilm for-
mation. Deletion of the hk11 or rr11 gene resulted in defects in
biofilm formation as illustrated in Fig. 3. Strain SMHK11 had
approximately 50% of the biofilm density and strain SMRR11
had about 75% of the density of the parent strain NG8. It is
unlikely that the reduction in biofilm density observed with the
mutants was caused by their slightly decreased growth rates,
especially since their growth yields at neutral pH were nearly
the same as that of the parent strain in TYEG (Fig. 2) and
SDM medium (data not shown).
A closer examination of the biofilms by SEM revealed that
biofilms formed by the mutants had a very different appear-
ance from the parent biofilm. The mutant biofilms appeared to
have sponge-like architecture with what appeared to be large
intercellular gaps (Fig. 4). We found that such biofilms formed
by both hk11 and rr11 mutants were washed off from the
surface more readily than those formed by the parent strain
during preparation for the biofilm assay. In addition, the rest-
ing cells of the mutants had a reduced ability to attach to the
mucin-coated polystyrene surface (percentage of cells attached
to the surface ? [standard deviation (SD)]: NG8, 12.08 [2.04];
HK11, 6.77 [1.32]; and RR11, 8.36 [1.53]). Taken together, the
apparent defects of the sponge-like architecture and the lower
affinity of the cells for adherence to the surface likely contributed
to the reduced biomass observed with the mutants. SEM also
FIG. 3. Biofilm formation and quantification of S. mutans strains.
The graph represents the turbidity of the biofilms as reflected by their
absorbance after safranin staining. The mean values ? SDs are pre-
FIG. 4. Scanning electron micrographs show spatial distribution and architecture of biofilms formed by S. mutans strains.
VOL. 184, 2002 TWO-COMPONENT SYSTEM INVOLVED IN BIOFILM FORMATION6337
revealed that both SMHK11 and SMRR11 formed very long
chains in comparison to the wt strain when grown as biofilms.
Since this phenotype was suspected of being linked to the
CSP-activated pathway, we examined the impact of CSP on cell
chain formation by the hk/rr11 mutants. Addition of CSP to the
biofilm cultures did not significantly change the length of
chains formed by the SMHK11 and SMRR11 mutants. In an
attempt to quantitate the chain lengths from the SEMs, averages
were obtained from four independent chains selected randomly.
The average numbers of cells per chain (?SD) were as follows:
NG8 (wt), 17 (8.04); HK11, 42 (13.8); and RR11, 38 (11.3).
The hk11 mutant is defective in acid tolerance. Compared to
the parent strain NG8, both SMHK11 and SMRR11 mutants
showed decreased growth rates in liquid culture at pH 5.5.
NG8 had a Tdof 100 ? 1 min. SMHK11 had a Tdof 178 ? 3
min, while SMRR11 doubled every 207 ? 9 min (Fig. 2).
Mutant SMHK11 also had greatly diminished growth on agar
plates at pH 5.0, although it grew as well as the parent strain
did on plates at pH 7.0 (Fig. 5) and nearly as well as the parent
in broth at pH 7.0 (Fig. 2). Interestingly, we were unable to
detect a difference between the growth of the SMRR11 mutant
and that of the parent strain NG8 on the pH 5.0 plates. To
more closely determine if deletion of the hk11 or rr11 gene
affected the inducible ATR, we assayed the log-phase ATR of
the mutants grown in liquid cultures by the method described
previously (18). The results showed that the SMHK11 mutant
had a reduced ATR relative to the parent strain NG8 (Fig. 6).
However, the deletion of rr11 resulted in only a slight decrease
in inducible ATR as observed with strain SMRR11.
The acid stimulon of SMHK11. Changes in protein expres-
sion underlying the ATR of NG8 and SMHK11 were analyzed
by comparative 2D gel electrophoresis of total cellular labeled
proteins of cells exposed to pH 7.5 and 5.5 for 30 min. Of 594
FIG. 5. Effect of pH on the growth of S. mutans strains. The photographs were taken after 40 h of incubation at 37°C with 5% CO2.
FIG. 6. Inducible ATR was assayed in log-phase cells of the mu-
tants. The mean percentages of survivors ? SDs from three indepen-
dent experiments are presented.
6338LI ET AL.J. BACTERIOL.
proteins monitored in these experiments, all those with differ-
ential expression of ?2.0 or ?0.5 upon acid exposure for 30
min were classified as belonging to the acid stimulon of NG8
and SMHK11, respectively. The acid stimulon of NG8 in-
cluded 19 proteins, 12 showing increased and 7 showing de-
creased expression (Fig. 7A and B; Table 3). Interestingly, of
the 12 NG8 proteins exhibiting increased expression following
the acid shock, four proteins were not induced in the mutant
FIG. 7. Autoradiograms of 2D gels obtained with parent strain NG8 (A and B) and mutant HK11 (C and D) of S. mutans exposed to pH 7.5
(A and C) and 5.5 (B and D) for 30 min. Solid and open arrowheads denote proteins that had increased or decreased expression, respectively, in
both NG8 and SMHK11 at the acid shock. Note the presence of four protein spots indicated with arrows in panel D that were induced in NG8
but not in SMHK11. Numbers at the top of each panel are pHs; numbers at the left are molecular masses in kilodaltons.
VOL. 184, 2002 TWO-COMPONENT SYSTEM INVOLVED IN BIOFILM FORMATION6339
strain (Fig. 7C and D; Table 3). These results confirm that
SMHK11 had defects in induction of acid stress proteins rel-
ative to the parent strain NG8 and confirm that hk11 plays an
important role in the inducible ATR.
Two-component regulatory systems are widespread prokary-
otic signal transduction systems that allow regulation of cellu-
lar functions in response to changing environments. Although
increasing information is available regarding identification and
characterization of two-component systems in various species
of bacteria, little is known of these systems in S. mutans, a
primary etiological agent of dental caries. Genome analyses
have revealed many putative TCSTSs in several related gram-
positive organisms such as S. pneumoniae (33), S. pyogenes (9),
Bacillus spp. (13, 32), and others. Using genome analysis, we
have recently described 13 separate TCSTSs in S. mutans and
have constructed 25 individual mutants of the 26 genes encod-
ing these systems (Lau and Cvitkovitch, abstract). Since one of
our major interests was to identify TCSTSs involved in the
expression of virulence factors of S. mutans, we focused our
attention on screening the mutants for phenotypes associated
with biofilm formation, acid tolerance, and other environmen-
tal stresses including ethanol, sodium laurel sulfate (common
in dentifrice), and H2O2.
Previous work in our lab has described a quorum sensing-
signaling system consisting of a two-component regulatory sys-
tem (ComDE) that was demonstrated previously to affect ge-
netic competence (19), biofilm formation (20), and ATR (18)
in S. mutans. In the present study, we present evidence that a
novel two-component regulatory system, HK/RR11, plays an
important role as a determinant of biofilm formation and acid
resistance phenotypes in S. mutans.
Our results clearly demonstrated that deletion of either hk11
or rr11 resulted in the formation of a biofilm with reduced
biomass and a sponge-like architecture (Fig. 3 and 4). One
striking feature observed by SEM was that the mutant biofilms
appeared to consist of many large intercellular channels rela-
tive to the parent biofilm, which had a more confluent appear-
ance. Since water channels in biofilms facilitate exchange of
substrates between a biofilm and bulk liquid phase (7), it is
possible that the hk/rr11 mutant biofilms were impaired in the
transport of a substrate or removal of a metabolic end product.
Another study, by Bhagwhat et al., examined S. mutans mu-
tants defective in the response regulators of six TCSTs (1).
One mutant described in this study was analogous to the RR11
mutant (tcek), and its ability to form biofilms was also assayed.
Yet, the Bhagwhat group did not describe a defect in biofilm
formation by this mutant. Notably, the parent strain and
growth and assay conditions were different from ours. Our
observation that rr11 deletion did not affect genetic compe-
tence was, however, consistent with the observations of Bhag-
wat et al. These investigators did, however, find that inactiva-
tion of tcbR (the comE gene encoding the response regulator
of the ComD/ComE TCSTS) resulted in a 10-fold reduction in
biofilm formation, which was consistent with our previous find-
ings that comD and comE mutants formed defective biofilms
with reduced biomass (20).
Compared to the parent strain, both SMHK11 and SMRR11
biofilms had sponge-like architecture that was composed of
cells organized in very long chains, a feature that we previously
observed with the biofilm formed by a comC mutant unable to
produce the signal peptide pheromone CSP (20). Mutants de-
fective in comD or comE did not, however, have a web- or
sponge-like architecture, suggesting that a separate pathway
was receptive to CSP. To further support the existence of a
second CSP sensor system, we found that exogenous addition
of CSP or complementation of the comC mutant with a wt
comC gene partially restored the wt phenotype of the comCDE
mutant biofilm. Since this mutant was defective in producing
the CSP and its cognate receptor (encoded by comD), we
hypothesized that there was another receptor(s) that recog-
nized CSP and was involved in cell septation or separation,
ultimately affecting biofilm architecture.
Since we suspected that the TCSTS encoded by hk/rr11
might function as the second pathway, we added CSP to the
mutant cultures to assess the effect on chain formation by the
hk11 and rr11 mutant biofilm cells. The results showed that
addition of CSP to the mutant cultures had no observable
impact on the length of chains comprising the mutant biofilms
(data not shown). This result is consistent with HK11 acting as
a CSP receptor but does not provide direct evidence to con-
clusively assign a role to HK11 as a CSP receptor. A closer
examination of the interaction of CSP with the hk/rr11 system
Wen and Burne (35) have recently described a gene, desig-
nated brpA (biofilm regulatory protein), which encodes a 406-
amino-acid protein in S. mutans UA159. Their work also
showed that inactivation of brpA resulted in a strain that pro-
duced an aberrant biofilm, with the mutant forming longer
chains than those of the parent strain. Although the same
phenotype was clearly observed here, there are currently no
data to link the BrpA-mediated effect to the HK/RR11 system.
TABLE 3. Proteins induced or repressed in NG8 and SMHK11
Protein with ?3-fold change,
in expression due to acid,
stress of NG8
Common to acid
stress of SMHK11
4a (pyruvate dehydrogenase).............................................. Yes
26 (unknown)........................................................................ Yes
85 (cysteine synthase).......................................................... Yes
87 (exopolyphosphatase)..................................................... No
155 (unknown)...................................................................... Yes
324 (unknown)...................................................................... Yes
814 (unknown)...................................................................... Yes
858 (unknown)...................................................................... No
1008 (histidine kinase candidate)....................................... No
2009 (unknown).................................................................... Yes
2120 (unknown).................................................................... No
2403 (unknown).................................................................... Yes
94 (unknown)........................................................................ Yes
116 (unknown)...................................................................... Yes
142 (unknown)...................................................................... Yes
603 (unknown)...................................................................... Yes
644 (unknown)...................................................................... Yes
821 (unknown)...................................................................... Yes
822 (unknown)...................................................................... Yes
aIndicated by solid arrowheads in Fig. 7B.
bIndicated by open arrowheads in Fig. 7A.
6340 LI ET AL.J. BACTERIOL.
Future studies will be necessary to examine possible interac-
tions among the HK/RR11 system, brpA, and the comCDE
quorum sensing system. Since we suspected that the hk/rr11
genes may encode a peptide sensing system, we searched the
region for small ORFs that encoded proteins encompassing
potential double GG cleavage sites typical of secreted signal
peptides but were unable to identify any candidate genes. The
putative function of the surrounding genes does not suggest
obvious roles in genetic competence, biofilm formation, or acid
tolerance. Although these neighboring genes do not obviously
appear to have a role in the phenotypes currently associated
with the CSP response, they may aid in optimal existence in a
biofilm or a high-cell-density environment. For example, the
gene encoding pyruvate formate lyase-activating enzyme, pflC,
in S. pneumoniae was recently demonstrated to be activated by
the pneumococcal CSP system (28), the gene for which has no
apparent relationship to genetic competence. The pflC gene
encoding the pyruvate formate lyase-activating enzyme is in
close proximity to hk/rr11 genes; it would be interesting to
investigate a linkage between these genes, since in S. mutans
pyruvate formate lyase is extremely oxygen sensitive and likely
functions optimally at high cell density in anaerobic biofilms
Another interesting observation was that the hk11 mutant
(SMHK11) was significantly impaired in acid tolerance.
SMHK11 was defective both in growth at a low pH and in
resistance to acid killing after adaptation to a signal pH (pH
5.5). This suggested that the membrane-associated protein en-
coded by hk11 might act as a pH sensor involved in activation
of one of the many pathways believed to affect the acid-toler-
ant phenotype of S. mutans. TCSTSs have been shown else-
where to act as pH sensors: most notably actSR of Rhizobium
and lisRK in Listeria monocytogenes (5) are essential for induc-
tion of the adaptive ATR (10). Interestingly, only SMHK11
appeared acid sensitive, since we did not observe the same
phenotypic effect when the rr11 gene was deleted. Intuitively,
one would expect that inactivation or deletion of either of the
genes encoding a TCSTS would generate a similar phenotype,
since a defect in either the histidine kinase receptor or the
cognate response regulator might hinder the input signal from
activating genes and pathways controlled by the response reg-
ulator (29). In our study, however, the observation that
SMHK11 and SMRR11 had different phenotypes suggested
that there may have been cross talk between related receptors
in which the histidine kinase sensor protein of the hk/rr11
system could pass the pH signal to one or more noncognate
response regulators. This phenomenon, called in vivo cross
talk, has been described recently by Verhamme et al. (34), who
demonstrated interaction among four key two-component sys-
tems in Escherichia coli by an in vivo approach. Their results
suggested that a functional histidine phosphoryl-transfer (HPt)
domain of a sensor kinase appears to be the active participant
in physiological cross talk. Further studies will be needed to
identify a putative response regulator(s) involved in cross talk
occurring via the HK11 sensor protein in SMRR11.
A comparison of the 2D profiles of SMHK11 and NG8
revealed that 14 of the acid-inducible (and -repressible) pro-
teins were conserved between the mutant and parent (Table 3).
SMHK11 did, however, have four proteins visible in NG8 that
were not detected in the mutants. One of these proteins, 1008,
was possibly the histidine kinase HK11 itself that migrated at
30 kDa with a pI of 5.5. The values obtained from the deduced
protein sequence of HK11 were as follows: calculated molec-
ular mass of 38,113 Da and an estimated pI of 5.71. Another
interesting protein that was not induced in SMHK11 was spot
number 87, which represents an exopolyphosphatase. Polypho-
sphate metabolism has been linked to biofilm formation in
many bacteria (3, 26, 27) including S. mutans (30). Polyphos-
phate likely provides a rapid source of energy needed to cope
with environmental fluctuations encountered during biofilm
The identification of promoter-like structures 5? from rr11
bearing a striking similarity to the com-box of S. pneumoniae is
intriguing. Expression of rr11 under CSP-limiting and -induc-
ing conditions could help lead to a deciphering of the S. mutans
com-box. We have identified similar structures in proximity to
late-competence orthologs found in the S. mutans genome.
Deduction of the S. mutans com-box could hasten our unrav-
eling of this regulon, as it would allow us to identify candidate
genes by in silico analysis. An understanding of CSP-mediated
and other genes involved in expression of the biofilm pheno-
type will hopefully allow us to discover means to control prob-
We thank Robert Chernecky for the SEMs.
Our work was supported by PHS grant DE 013230 from the National
Institute of Dental and Craniofacial Research and grant MT-15431
from the Canadian Institutes of Health Research and by infrastructure
grants from the Canadian Foundation for Innovation and The Ontario
Innovation Trust. D.G.C. is supported by a Canada Research Chair.
1. Bhagwat, S. P., J. Nary, and R. A. Burne. 2001. Effects of mutating putative
two-component systems on biofilm formation by Streptococcus mutans
UA159. FEMS Microbiol. Lett. 205:225–230.
2. Buckley, N. D., L. N. Lee, and D. J. LeBlanc. 1995. Use of a novel mobilizable
vector to inactivate the scrA gene of Streptococcus sobrinus by allelic replace-
ment. J. Bacteriol. 177:5028–5034.
3. Chen, W., R. J. Palmer, and H. K. Kuramitsu. 2002. Role of polyphosphate
kinase in biofilm formation by Porphyromonas gingivalis. Infect. Immun.
4. Claverys, J. P., A. Dintilhac, E. V. Pestova, B. Martin, and D. A. Morrison.
1995. Construction and evaluation of new drug-resistance cassettes for gene
disruption mutagenesis in Streptococcus pneumoniae, using an ami test plat-
form. Gene 164:123–128.
5. Cotter, P. D., N. Emerson, C. G. Gahan, and C. Hill. 1999. Identification and
disruption of lisRK, a genetic locus encoding a two-component signal trans-
duction system involved in stress tolerance and virulence in Listeria mono-
cytogenes. J. Bacteriol. 181:6840–6843.
6. Cvitkovitch, D. G., D. A. Boyd, T. Thevenot, and I. R. Hamilton. 1995.
Glucose transport by a mutant of Streptococcus mutans unable to accumulate
sugars via the phosphoenolpyruvate phosphotransferase system. J. Bacteriol.
7. Davey, M. E., and G. A. O’Toole. 2000. Microbial biofilms: from ecology to
molecular genetics. Microbiol. Mol. Biol. Rev. 64:847–867.
8. Dziejman, M., and J. J. Mekalanos. 1995. Two-component signal transduc-
tion and its role in the expression of bacterial virulence factors, p. 305–317.
In J. A. Hoch and T. J. Silhavy (ed.), Two-component signal transduction.
ASM Press, Washington, D.C.
9. Ferretti, J. J., W. M. McShan, D. Ajdic, D. J. Savic, G. Savic, K. Lyon, C.
Primeaux, S. Sezate, A. N. Suvorov, S. Kenton, H. S. Lai, S. P. Lin, Y. Qian,
H. G. Jia, F. Z. Najar, Q. Ren, H. Zhu, L. Song, J. White, X. Yuan, S. W.
Clifton, B. A. Roe, and R. McLaughlin. 2001. Complete genome sequence of
an M1 strain of Streptococcus pyogenes. Proc. Natl. Acad. Sci. USA 98:4658–
10. Glenn, A. R., W. G. Reeve, R. P. Tiwari, and M. J. Dilworth. 1999. Acid
tolerance in root nodule bacteria. Novartis Found. Symp. 221:112–126.
11. Hamada, S., and H. D. Slade. 1980. Biology, immunology, and cariogenicity
of Streptococcus mutans. Microbiol. Rev. 44:331–384.
12. Hoch, J. A., and T. J. Silhavy (ed.). 1995. Two-component signal transduc-
tion. ASM Press, Washington, D.C.
VOL. 184, 2002TWO-COMPONENT SYSTEM INVOLVED IN BIOFILM FORMATION 6341
13. Kunst, F., N. Ogasawara, I. Moszer, A. M. Albertini, G. Alloni, V. Azevedo,
M. G. Bertero, P. Bessieres, A. Bolotin, S. Borchert, R. Borriss, L. Boursier,
A. Brans, M. Braun, S. C. Brignell, S. Bron, S. Brouillet, C. V. Bruschi, B.
Caldwell, V. Capuano, N. M. Carter, S. K. Choi, J. J. Codani, I. F. Conner-
ton, A. Danchin, et al. 1997. The complete genome sequence of the gram-
positive bacterium Bacillus subtilis. Nature 390:249–256.
14. Lange, R., C. Wagner, A. de Saizieu, N. Flint, J. Molnos, M. Stieger, P.
Caspers, M. Kamber, W. Keck, and K. E. Amrein. 1999. Domain organiza-
tion and molecular characterization of 13 two-component systems identified
by genome sequencing of Streptococcus pneumoniae. Gene 237:223–234.
15. Lau, P. C., C. K. Sung, J. H. Lee, D. A. Morrison, and D. G. Cvitkovitch.
2002. PCR ligation mutagenesis in transformable streptococci: application
and efficiency. J. Microbiol. Methods 49:193–205.
16. Lee, M. S., and D. A. Morrison. 1999. Identification of a new regulator in
Streptococcus pneumoniae linking quorum sensing to competence for genetic
transformation. J. Bacteriol. 181:5004–5016.
17. Lee, S. F., Y. H. Li, and G. H. Bowden. 1996. Detachment of Streptococcus
mutans biofilm cells by an endogenous enzymatic activity. Infect. Immun.
18. Li, Y. H., M. N. Hanna, G. Svensater, R. P. Ellen, and D. G. Cvitkovitch.
2001. Cell density modulates acid adaptation in Streptococcus mutans: im-
plications for survival in biofilms. J. Bacteriol. 183:6875–6884.
19. Li, Y. H., P. C. Lau, J. H. Lee, R. P. Ellen, and D. G. Cvitkovitch. 2001.
Natural genetic transformation of Streptococcus mutans growing in biofilms.
J. Bacteriol. 183:897–908.
20. Li, Y. H., N. Tan, M. B. Aspiras, P. C. Y. Lau, J. H. Lee, R. P. Ellen, and D. G.
Cvitkovitch. 2002. A quorum-sensing signaling system essential for genetic
competence in Streptococcus mutans is involved in biofilm formation. J.
21. Loo, C. Y., D. A. Corliss, and N. Ganeshkumar. 2000. Streptococcus gordonii
biofilm formation: identification of genes that code for biofilm phenotypes.
J. Bacteriol. 182:1374–1382.
22. Marsh, D. J. 2000. Oral ecology and its impact on oral microbial diversity.
Horizon Scientific Press, Wymondham, Norfolk, United Kingdom.
23. Marsh, P. D. 1994. Microbial ecology of dental plaque and its significance in
health and disease. Adv. Dent. Res. 8:263–271.
24. Matsushita, M., and K. D. Janda. 2002. Histidine kinases as targets for new
antimicrobial agents. Bioorg. Med. Chem. 10:855–867.
25. Quivey, R. G., W. L. Kuhnert, and K. Hahn. 2001. Genetics of acid adapta-
tion in oral streptococci. Crit. Rev. Oral Biol. Med. 12:301–314.
26. Rashid, M. H., and A. Kornberg. 2000. Inorganic polyphosphate is needed
for swimming, swarming, and twitching motilities of Pseudomonas aerugi-
nosa. Proc. Natl. Acad. Sci. USA 97:4885–4890.
27. Rashid, M. H., K. Rumbaugh, L. Passador, D. G. Davies, A. N. Hamood,
B. H. Iglewski, and A. Kornberg. 2000. Polyphosphate kinase is essential for
biofilm development, quorum sensing, and virulence of Pseudomonas aerugi-
nosa. Proc. Natl. Acad. Sci. USA 97:9636–9641.
28. Rimini, R., B. Jansson, G. Feger, T. C. Roberts, M. de Francesco, A. Gozzi,
F. Faggioni, E. Domenici, D. M. Wallace, N. Frandsen, and A. Polissi. 2000.
Global analysis of transcription kinetics during competence development in
Streptococcus pneumoniae using high density DNA arrays. Mol. Microbiol.
29. Stock, A. M., V. L. Robinson, and P. N. Goudreau. 2000. Two-component
signal transduction. Annu. Rev. Biochem. 69:183–215.
30. Svensater, G., J. Welin, J. C. Wilkins, D. Beighton, and I. R. Hamilton. 2001.
Protein expression by planktonic and biofilm cells of Streptococcus mutans.
FEMS Microbiol. Lett. 205:139–146.
31. Takahashi, N., K. Abbe, S. Takahashi-Abbe, and T. Yamada. 1987. Oxygen
sensitivity of sugar metabolism and interconversion of pyruvate formate-
lyase in intact cells of Streptococcus mutans and Streptococcus sanguis. Infect.
32. Takami, H., K. Nakasone, Y. Takaki, G. Maeno, R. Sasaki, N. Masui, F. Fuji,
C. Hirama, Y. Nakamura, N. Ogasawara, S. Kuhara, and K. Horikoshi.
2000. Complete genome sequence of the alkaliphilic bacterium Bacillus
halodurans and genomic sequence comparison with Bacillus subtilis. Nucleic
Acids Res. 28:4317–4331.
33. Tettelin, H., K. E. Nelson, I. T. Paulsen, J. A. Eisen, T. D. Read, S. Peterson,
J. Heidelberg, R. T. DeBoy, D. H. Haft, R. J. Dodson, A. S. Durkin, M.
Gwinn, J. F. Kolonay, W. C. Nelson, J. D. Peterson, L. A. Umayam, O. White,
S. L. Salzberg, M. R. Lewis, D. Radune, E. Holtzapple, H. Khouri, A. M.
Wolf, T. R. Utterback, C. L. Hansen, L. A. McDonald, T. V. Feldblyum, S.
Angiuoli, T. Dickinson, E. K. Hickey, I. E. Holt, B. J. Loftus, F. Yang, H. O.
Smith, J. C. Venter, B. A. Dougherty, D. A. Morrison, S. K. Hollingshead,
and C. M. Fraser. 2001. Complete genome sequence of a virulent isolate of
Streptococcus pneumoniae. Science 293:498–506.
34. Verhamme, D. T., J. C. Arents, P. W. Postma, W. Crielaard, and K. J.
Hellingwerf. 2002. Investigation of in vivo cross-talk between key two-com-
ponent systems of Escherichia coli. Microbiology 148:69–78.
35. Wen, Z. T., and R. A. Burne. 2002. Functional genomics approach to iden-
tifying genes required for biofilm development by Streptococcus mutans.
Appl. Environ. Microbiol. 68:1196–1203.
6342 LI ET AL. J. BACTERIOL.