AMethodforStructure3

CSP-Dependent Transformation Assay

To determine if synthetic peptides activated quorum sensing for induction of genetic competence, we used the comC deletion mutant that was unable to produce, but still responded to CSP, to assay peptide-dependent genetic transformation using the method as described previously (2, 10). A streptococcal suicide vector pVA-gtfA, which harbored a 2.4-kb fragment of the S. mutans gtfA gene and an erythromycin resistance marker, was used as transforming DNA. The wild-type GS5 was used as a positive control, while both ComC (SMdC) and ComE mutants (SMdE) were used as negative controls. The wild-type strain was grown on Todd–Hewitt yeast extract (THYE) medium, whereas the mutants were maintained on THYE plus spectinomycin. When culture reached to early mid-log phase (OD₆₀₀ ≈ 0.25), an aliquot of CSP was added into the culture at a final concentration of 50 nM (AC₅₀ ≈ 0.25 nM). After 20 min of incubation at 37°C, an aliquot of transforming DNA of plasmid pVA-gtfA conferring erythromycin resistance was added at a final concentration of 1 µg mL⁻¹. The culture was incubated for additional 2 h before plated on THYE agar plates plus 10 µg mL⁻¹ erythromycin. Prior to addition of transforming DNA, an aliquot of the cell suspension was plated on THYE plates containing the same antibiotic to monitor spontaneous mutation. Transformation frequency was expressed as percentage of transformants against total viable recipient cells per milliliter cell suspension.

β-Galactosidase Activity Assay

A lacZ reporter assay was performed to monitor quorum-sensing activation using the same method as described previously (10). The newly constructed lacZ reporter strains were used to assay and quantify β-galactosidase activity in response to test peptides. Aliquots of samples were taken to prepare cell lysates at different time points T ₀, T ₁₅, T ₃₀, T ₆₀, and T ₁₂₀ after addition of a test peptide. Protein concentrations of the supernatants were determined by Bio-Rad protein assay (Bio-Rad). Specific β-gal activities (A ₄₂₀ min⁻¹ mL⁻¹ mg⁻¹ protein) were calculated from triplicate samples from two independent experiments. The data from the β-gal activity experiments were then normalized against negative or background controls and plotted as percentages of maximal activation versus log peptide concentrations. The half maximal activation concentration (AC₅₀) was determined from the sigmoidal dose–response curve using Prism 4 (Graphpad, San Diego, CA, USA).

Assay for QS-Controlled Bacteriocin Production

The S. mutans comC mutant SMdC as well as its parent GS5 and wild-type strain UA159 (positive controls) were used to assay for bacteriocin production using a modified agar plate method (7, 23–25). Briefly, an aliquot of mid-log phase cells of the S. mutans wild-type strains and mutant SMdC grown under a condition with or without use of CSP or a test peptide were stabbed onto THYE agar plates. The plates were incubated anaerobically at 37°C for 6 h before carefully overlaid with 100 µL of cell suspension (about 10⁶ CFU/mL⁻¹) of an indicator strain Streptococcus sanguinis SK108. The plates were further incubated anaerobically at 37°C for 24 h before inspection of bacteriocin production. The S. mutans strains showing an inhibitory zone around the stabs were scored as positive for bacteriocin production. The plates were then photographed for records.

Results and Discussion

CD and NMR Structural Analyses of the Signaling Peptides

CD and NMR spectroscopy are non-destructive and non-invasive techniques that have been widely used to determine secondary folding features of proteins or peptides (14–17). In addition, NMR spectroscopy records data from “spin-active” nuclei, such as ¹H, ¹³C, ¹⁵N, or ³¹P. These nuclei are sensitive to their chemical environment, which is determined by the type and number of bonds and the position and proximity of other chemical entities and the solvent. Thus, NMR is a powerful tool that can be used to determine atomic level structural information, dynamics (chemical exchange, motional fluctuations, proton exchange), ligand binding, multimeric states, and effects of competitive receptors (11, 15, 26).

Using CD and NMR spectroscopy, we have determined the three-dimensional structures of two signaling peptides, UA159sp from S. mutans UA159 and a C-terminally truncated peptide TPC3 from JH1005 defective in genetic competence (10). Initial NMR data for both peptides were recorded in aqueous buffer to provide an assessment of the structural nature of the peptides. In the presence of water, both peptides were not soluble enough for extensive NMR studies. In TFE, however, these peptides formed well-defined α-helices from residues Leu4-Gly20 of UA159sp and Ser5-Thr16 of TPC3 (Fig. 2 ). Both UA159sp and TPC3 formed a similar structure in TFE and DPC-d ₃₈micelles, although the C-terminal three residues of TCP3 were truncated. The structural coordinates for two peptides have been deposited in the RCSB Protein Data Bank (http://www.pdb.org/) with RCSB ID code rcsb039055 or PDB ID code 2I2J for UA159sp and RCSB ID code rcsb039053 or PDB ID code 212H for TPC3.

Fig. 2 Computer-simulated three-dimensional structures of UA159sp and TPC3 determined by NMR and CD analyses. Only the α-helical portions of the peptides with designated hydrophobic and hydrophilic residues are presented. a A comparison in three-dimensional structures between UA159sp (CSP) and TPC3 that lacked the proposed C-terminal motif. b UA159sp viewed from different angles indicates the positions of phenylalanine (F) residues on the core of α-helix and C-terminal motif of slightly motional freedom.

An important feature observed from structural analyses was that both UA159sp and TPC3 formed amphipathic α-helices with a well-defined hydrophobic face characterized by a row of four highly hydrophobic phenylalanines (F7, F8, F11, and F15) on one side and hydrophilic residues on another side (Fig. 2 ). The hydrophobic face comprised about 40% of the surface of the helix, whereas hydrophilic and charged residues comprised the remainder of the surface. A major structural difference between UA159sp and TPC3 was that UA159sp showed a clear C-terminal motif of slightly more motional freedom than TPC3. From structures of other similar binding peptides, a well-defined row of hydrophobic residues on an amphipathic α-helix is common and necessary for ligand binding to the receptor protein (12, 19). Our work suggests that the quorum-sensing signaling peptides from S. mutans, probably from other naturally transformable streptococci, fall into the FXXFF motif family, a general protein–protein interaction motif, where the interaction is made through a hydrophobic face formed by the hydrophobic residues packed into a hydrophobic pocket of the receptor (11, 12, 26). The fact that the amphipathic helical structure is reflected in the binding and induction of quorum-sensing suggests that the peptide may be helical within the active site of the receptor. The membrane environment may help to form the helical structure, facilitating binding to the receptor. We have chosen to dissolve the peptides in widely used solvent TFE because it improves structural stability of peptides and is considered to be biologically relevant (27). TFE is also thought to mimic a protein receptor environment by stabilizing helices in regions with intrinsic α-helical propensity that are likely to form helices when binding to their protein partner.

Signaling Peptide–Receptor Activation in Newly Constructed Strains

In previous studies, we exclusively used S. mutans UA159 and its derivatives to assay QS activation in response to signaling peptides (2, 10). However, this genome sequence reference strain is not a typical producer for CSP-induced bacteriocins (7, 24, 25). To construct a comC mutant that allowed us to assay CSP-dependent bacteoricin production, we transformed the previously generated comC deletion construct (20) into S. mutans GS5, a wild-type strain that was demonstrated to produce all known QS-controlled bacteriocins (6, 7, 22). The new mutant, named SMdC, was confirmed to have the internal region of comC replaced by a spectinomycin (Spec^r) resistance cassette. The mutant was confirmed to be defective in quorum sensing for induction of genetic competence (Fig. 3 ) and bacteriocin production (Fig. 4 ), unless CSP or a peptide agonist was added into the culture. Our work confirmed that this mutant provided an excellent negative background, enabling the assay of peptide-dependent QS activation without interference by endogenous CSP.

Fig. 3 Induction of genetic competence of the comC mutant (SMdC) in response to addition of CSP. S. mutans wild-type GS5 was used as a positive control, while the comE deletion mutant (SMdE) was used as a negative control.

Fig. 4 Constructed lacZ reporter strains, SMdC-PnlmAB and SMGS5-PnlmAB, express predicted β-gal activity in response to CSP or agonist SPA-18. a The expression of β-gal activity in SMdC-PnlmAB is CSP-dependent, which is in contrast to strain SMGS5-PnlmAB. b A time-course experiment indicates that SMdC-PnlmAB expresses the highest levels of β-gal activity around 90 min after the addition of CSP or SPA-18 and maintains such high levels of the activity for at least 2 h.