Peptides were synthesized by Genscript (Piscataway Township, NJ, USA) at >80% purity. Peptides P1 and sP1 matched previously published amino acid sequences (Heisig et al., 2014) and the alternative peptide sequences P0, P1–4, P1–7, and P17 were designed as variations of the P1 sequence (Figure 6 and Supplementary Table 2). Lyophilized peptides were diluted in phosphate-buffered saline (PBS) to a stock concentration of 1 mg/ml and used immediately or frozen at -20°C in 100 μl aliquots. Streptococci bacteria (Streptococcus mutans ATCC 25175, Streptococcus oralis ATCC 9811, and Streptococcus salivarius ATCC 13419), Gram-negative bacteria (Alcaligenes faecalis ATCC 8750, Enterobacter cloacae ATCC 23355, and Salmonella typhimurium ATCC 29629), and Gram-positive bacterial strains (Abraham et al., 2017) were obtained from Yale University or Presque Isle Cultures (Erie, PA, USA). Bacterial cultures were maintained on Brain Heart Infusion (BHI) agar or broth (for streptococci), and Tryptic Soy agar or broth (TSA, TSB).

Biofilm assays were performed as previously described (Merritt et al., 2011). Overnight cultures of S. mutans grown in BHI medium added to BHI+1% sucrose (BHI-S) at a concentration of 3–5 × 10⁶ cfu/ml [by either a 1:100 dilution or normalizing to an optical density at 600 nm (OD₆₀₀) of 0.015]. The bacterial suspension was inoculated into quadruplicate wells of a 96-well tissue culture plate and P1 peptide or control sP1 peptide was added at a final concentration of 0.1 mg/ml (46 μM). Bacteria were incubated overnight at 37°C with a CO₂ GasPak (BD, Franklin Lakes, NJ, USA). Crystal violet staining of biofilm plates was performed using standard protocols (Merritt et al., 2011), as follows: Non-adherent cells were washed off by submerging the plate in water, and adherent cells were stained with 0.1% crystal violet, washed twice with water, dried, and resuspended in 33% acetic acid for 10 min. The absorbance of the resuspended solution was measured at 590 nm on a BioTek microplate reader. For the other Gram-negative and Gram-positive bacterial strains, the assay was identical except that the 96-well plates were incubated aerobically, and the biofilm culture media was TSB + 1% glucose for Gram-negatives, and BHI+1% glucose for Gram-positives.

To assess the bactericidal activity of P1 against S. mutans, overnight cultures of S. mutans were inoculated into 12-well or 24-well plates at a 1:100 dilution in BHI-S containing 0.1 mg/ml of P1 peptide, control sP1 peptide, or no peptide, and incubated overnight at 37°C with a CO₂ GasPak. The S. mutans cells were collected in two ways: first, as a total (combined) sample, pooling all cells in both the supernatant (planktonic) and biofilm fraction; secondly, the supernatant (non-adherent) and biofilm fractions were separated and processed separately. For the combined sample, the attached biofilm cells were scraped off the bottom using a sterile 1.8 cm cell scraper, resuspended, and the entire contents of the well were transferred to a microcentrifuge tube. An additional 500 μl of PBS was added to the well and re-scraped to pick up any remaining adherent cells. For the separated fractions, the supernatant (containing planktonic bacteria) was pipetted off and transferred to a 1.5 ml tube, then 500 μl of PBS was added gently to the well to rinse off non-adherent cells, and transferred to the same “supernatant” tube. The biofilm cells were resuspended into 1 ml of PBS using a cell scraper, transferred to a new tube, and an additional 500 μl of PBS was added to the well and re-scraped to collect any remaining adherent cells. All samples (total, supernatant, or biofilm) resuspended in a total volume of 1.5 ml each, were briefly sonicated to break up cell clumps (10 pulses at 30–40% duty cycle), then serially diluted in PBS to 10^-7, spread-plated on BHI agar, and incubated for 48–72 h at 37°C with a CO₂ GasPak. After counting colonies, the number of colony-forming units (cfu/ml) was calculated for each sample.

The sHA disc procedure was carried out according to published protocols (Lemos et al., 2010). We prepared authors’ own clarified saliva for each experiment as follows: ~10 mL of saliva was collected in a Falcon tube, stored on ice, and mixed with an equal volume of Adsorption Buffer (AB; 50 mM KCl, 1 mM potassium phosphate, 1 mM CaCl₂, 0.1 mM MgCl₂, pH 6.5), followed by the addition of 10 μl of 0.1 M phenylmethyl-sulfonyl fluoride (PMSF). The mixture was centrifuged at 5500 × g for 10 min at 4°C. The supernatant (containing clarified whole saliva) was pipetted off carefully and filtered through a 0.22 μm PES low protein-binding filter (Steriflip, EMD Millipore, Billerica, MA, USA) and used immediately or stored at 4°C. Sterile 5-mm diameter hydroxyapatite (HA) discs (3D Biotek, Bridgewater, NJ, USA) were aseptically placed into a sterile 15 ml tube containing 5–10 mL of clarified saliva and rotated at 37°C for 1 h to coat discs with saliva. The sHA discs were then placed vertically in custom-made wire holders in pre-seeded wells of the 96-well culture plates. Triplicate wells were inoculated with a final volume of 200 μl of Streptococcus mutans in BHI-S (diluted to an OD 600 nm of 0.015), with either P1 or control sP1 peptides at 0.1 mg/ml. The plate was incubated at 37°C with a CO₂ GasPak for 24 h. Discs were removed with sterile forceps, dip-washed three times in 0.89% NaCl, and placed in glass tubes containing 1 mL of 0.89% NaCl. The glass tubes were distributed in a custom Styrofoam rack and subjected to a 10-min ultrasonic bath. Following the ultrasonic bath, the detached cell suspension was transferred into microfuge tubes and sonicated (2x 5 pulses, on ice) to break up clumps of cells. Cells were serially diluted in PBS to 10^-5, then plated on BHI agar and incubated at 37°C with a CO₂ GasPak. After 48–72 h of incubation, S. mutans colonies were counted and cfu/disc was calculated.

Approximately 1.5 × 10⁷ cells of S. mutans, from an overnight culture growing in BHI, were pelleted and washed in PBS and resuspended in 100 μL PBS. His-tagged P1 or sP1 peptides (GenScript, Piscataway Township, NJ, USA) were added to the cultures at 0.1 mg/mL. Bacteria-peptide mixtures were incubated for 10 min at 37°C. Cells were pelleted by centrifugation, the supernatant (unbound fraction) was removed, and the pellet-associated (bound) fraction was washed with 0.1% Triton X-100 (Sigma–Aldrich, St. Louis, MO, USA) in PBS. The pellet (Associated) fraction and the supernatant (unbound) fraction were spotted directly onto a nitrocellulose membrane for immunoblot analysis. Bacterial bound (Associated) or unbound (Supernatant) peptide was detected using a monoclonal His-tagged antibody (Clontech, Mountain View, CA, USA). Cell wall bound fraction (Associated) versus unbound (Supernatant) bacterial fractions were distinguished using a polyclonal anti-Wheat Germ Agglutinin (WGA) (Vector Laboratories, Burlingame, CA, USA) antibody that detects bacterial peptidoglycan. Primary antibodies for both His and WGA were used at a 1:2000 dilution. Appropriate IR-conjugated LICOR secondary antibodies were used at a 1:5000 dilution and detected using the LI-COR Odyssey system (LI-COR, Lincoln, NE, USA). To normalize for any differences in reactivity between P1-His and sP1-His with the anti-His antibody, we quantified the bound signal intensity (Associated) and represented it as a percentage of the total signal (Associated + Supernatant):

Overnight cultures of S. mutans were inoculated at a 1:100 dilution in BHI-S with 0.1 mg/ml of P1 peptide or control sP1 peptide. Biofilms were formed either on coverslips in 6-well plates, or directly on the surface of 12-well or 24-well tissue culture plates (Corning, Corning, NY, USA). Bacteria were incubated overnight at 37°C in containers with a CO₂ GasPak. Biofilm cultures and peptides in solution in tissue culture plates were viewed with phase contrast microscopy using a Zeiss Axio Vert.A1 microscope. For eDNA staining, a Miami Orange DNA-binding fluorescent probe (Kerafast, Boston, MA, USA) was used at a 1:200 dilution in the well, and viewed with a TRITC fluorescent filter. Images were captured using either a mounted camera with SPOT imaging software (SPOT Imaging, Sterling Heights, MI, USA) or an iPhone camera (Apple, Cupertino, CA, USA) mounted to the eyepiece of the microscope with a MiPlatform (Scientific Device Laboratory). Scale bars were calibrated using a stage micrometer.

S. mutans was inoculated at a 1:100 dilution into 500 μl of BHI-S in a 12-well tissue culture plate, with glass coverslips or silicon wafers aseptically placed at the bottom to allow for biofilm formation. Peptide P1 at 0.1 mg/ml was added to half of the samples, and the other half contained 0.1 mg/ml control peptide sP1. After 24 h of incubation with CO₂ at 37°C, the coverslips and wafers were gently rinsed twice in 0.12 M phosphate buffer for 10 min. Coverslips/wafers were suspended for 1 h in a fixative solution containing 1.5% glutaraldehyde and 1% formaldehyde in 0.12 M phosphate buffer, followed by two rinses in 0.12 M phosphate buffer. Coverslips and wafers were then fixed in 1% OsO₄ in 0.12 M Phosphate for 30 min, followed by a di H₂O rinse. Next, the coverslips/wafers went through an ethanol dehydration process, suspended in the following concentrations of ethanol for 20 min each: 30, 50, 70, 95, 100% ethanol. Following the 100% ethanol, samples were moved to the critical point dryer in liquid CO₂. After critical point drying, the coverslips/wafers were sputter coated with gold palladium in argon gas. Biofilm samples were imaged at the UConn BEML facility using an FEI Nova NanoSEM 450. Images were photographed at magnifications ranging from 500× to 50,000×, and the 3084× magnification was selected for bacterial chain length measurement. Twenty additional images from the P1 and control sP1 biofilms were taken at 3084×, and two representative images per treatment were used to estimate streptococcal chain length. Chains of cells with a clear beginning and end (30 chains per treatment group) were randomly selected and chain length was measured with ImageJ software (Schneider et al., 2012).

Synthetic peptides P1, P17, and sP1 were diluted to a concentration of 1 and 0.1 mg/ml in PBS. Aliquots of 8 μl per sample were dried as thin films directly on the ATR surface of a Bruker Alpha Platinum-ATR FTIR (Bruker Daltonics, Billerica, MA, USA). To scan for absorption peaks indicating the presence of β-sheets, spectra were recorded in the range of 1800–1500 cm^-1 at a resolution of 2 cm^-1, with 50 scans per sample.

Experiments were repeated independently at least three times, and one representative experiment or pooled data are shown. Data were analyzed using GraphPad Prism version 6.0 software (GraphPad Software, Inc., La Jolla, CA, USA). Statistical significance was determined using Students t-test with Welch’s correction (when comparing two groups) or One-Way ANOVA with Dunnett’s or Sidak’s multiple comparison post-test.

The biofilm microplate assay using crystal violet staining is a rapid and convenient test for static biofilm formation. We cultured Streptococcus mutans cultures in BHI medium containing 1% sucrose, along with peptide P1 or a randomized scrambled peptide, sP1, a previously published negative control (Heisig et al., 2014). Peptides were used at a final concentration of 0.1 mg/ml (46 μM). As shown in Figure 1, when co-incubated with peptide P1 overnight, the biofilm biomass of Streptococcus mutans was reduced by approximately 80% compared to the control biofilms (sP1 peptide or no peptide) (Figure 1A). Likewise, incubation with P1 under the same conditions also reduced biofilm formation by two other oral streptococci, Streptococcus oralis and Streptococcus salivarius (Figure 1A). To determine the minimum effective concentration for biofilm inhibition of S. mutans, we tested decreasing concentrations of peptide P1. Significant anti-biofilm activity was observed at concentrations down to 25 μg/ml (Figure 1B), however, 0.1 mg/ml gave consistent results and was used for all other experiments. Despite inhibiting biofilm formation, peptide P1 was not bactericidal against S. mutans, similar to the effect on S. aureus (Heisig et al., 2014). Viability assays revealed a similar number of total colony-forming units present in control cultures and cultures incubated with P1, after dilution and plating of the entire contents of the tissue culture well (biofilm and non-adherent cells) (Figure 1C). As expected, when cells from the biofilm and non-adherent (supernatant) fractions were processed and quantified separately, there were ~80% fewer cfu/ml in the biofilm-associated fraction of the P1-incubated S. mutans cultures compared to controls, and more cfu/ml in the non-adherent (supernatant) fraction (Supplementary Figure 1).

To investigate the effect of peptide P1 on S. mutans biofilms in an in vitro tooth model, we prepared sHA discs, an established model that mimics tooth enamel coated with a salivary pellicle (Koo et al., 2010). The sHA discs were coated with clarified saliva using a standard method (Lemos et al., 2010) and incubated vertically for 24 h in handmade wire holders in 96-well plates (Figure 1E) with S. mutans in BHI-S with peptide P1 or sP1. After incubation, attached biofilm bacteria were collected from the discs (by washing to remove loosely attached bacteria, then placed in tubes in an ultrasonic bath to detach biofilm bacteria from the discs). The detached biofilm bacterial suspension was sonicated to break up clumps of cells before plating onto BHI agar. When S. mutans was cultivated on the discs in the presence of P1, we observed a 70% decrease in the number of biofilm colony-forming units (cfu) per disc (7.7 × 10⁶ cfu/disc) compared to the sP1 control (2.8 × 10⁷ cfu/disc) (Figure 1D). We did not observe a significant difference in 48-h biofilms of S. mutans (data not shown).

The full-length IAFGP protein, from which P1 was designed, was previously shown to bind to the cell surfaces of a range of bacteria, including both Gram-positive and Gram-negative species (Heisig et al., 2014). Direct binding between both IAFGP and peptide P1 and cell wall peptidoglycan of S. aureus and Enterococcus faecalis was recently discovered (Abraham et al., 2017). To establish whether peptide P1 can bind to S. mutans cells, we performed an immunoblot assay with His-tagged peptide P1 or sP1. The His-tagged peptides were incubated with a culture of S. mutans, then the cells were spun down, washed, and the cell-associated and unbound fractions were probed with anti-His antibody. As shown in Figure 2, there was strong binding detected between His-P1 peptide with the cell-associated S. mutans fraction (anti-His, “Associated”), while only a trace amount of His-sP1 control peptide was detected bound to cells.

Since P1 peptide binds to S. mutans, and IAFGP binds to a range of bacteria (Heisig et al., 2014), we wished to investigate the anti-biofilm effect of P1 against other bacteria. Recently, Abraham et al. (2017) reported that P1 had strong anti-biofilm activity against biofilms of Gram-positive species, and no effect on Escherichia coli and P. aeruginosa biofilm formation (Abraham et al., 2017). We tested the peptide against three additional biofilm-forming Gram-negative strains, A. faecalis, Enterobacter cloacae, and Salmonella typhimurium, and did not detect any statistically significant biofilm inhibition (Figure 3). Supplementary Table 1 summarizes the findings from both studies on peptide P1 against the Gram-positive and Gram-negative bacteria tested thus far. Biofilms of Gram-positive species were reduced by 35–80%, and statistically significant reduction of biofilm by peptide P1 could be detected at concentrations as low as 5 μg/ml against S. oralis, and 10 μg/ml against S. aureus (Supplementary Table 1).

When S. mutans biofilms were allowed to form in the wells of 12- or 24-well tissue culture plates, cultures in the presence of peptide P1 exhibited large aggregates of cells and exopolymeric matrix, that detached readily from the surface (Figure 4 and Supplementary Figure 1). Control sP1-treated biofilms were more uniform and firmly attached, with a continuous network of cells and exopolymeric matrix material (Figure 4 and Supplementary Figure 1). To visualize eDNA in the biofilm matrix, we stained the total contents of the wells (adherent and suspended) with a fluorescent DNA-binding dye, and observed the same pattern of unevenly clumped matrix material with less attached to the surface of the well (Figure 4D).

To further examine the altered biofilm structure of S. mutans treated with peptide P1, we performed SEM to look for differences in biofilm architecture and cell morphology at high resolution. At 500× magnification, there were fewer clumps of attached S. mutans microcolonies in the P1-treated sample (Figure 5A). At 3084× magnification, striking differences in streptococcal chain length were apparent between the samples, with the control cells forming significantly longer chains that were extensively interlinked (Figure 5B). We measured chain lengths using ImageJ software and found that the majority of streptococcal chains in the P1-incubated biofilm were less than 6 μm long, while chains of cells in the control group were more than twice as long, with most measured at 10–16 μm (Figures 5C,D). The representative images in Figure 5B illustrate that cells of control biofilms form a more interconnected network, while cells incubated with P1 exist in discontinuous clumps.

To explore whether altering the sequence of P1 could enhance or eliminate its anti-biofilm activity, we designed modified peptide sequences and tested their anti-biofilm effect on S. mutans and S. aureus. Figure 6 lists the alternative peptide sequences with color-coded amino acid residues. Additional details including peptide molecular weights and net charges can be found in Supplementary Table 2. Peptide P0 directly matches the repeated sequence in the full length IAFGP protein (with ‘PATAAT’ instead of ‘AATAAT’). Peptides P1–4 and P1–7 have fewer (4) or more (7) ‘AAT’ repeats, respectively, than P1. P17 is a different randomized scrambled peptide that lacks the proline residue. As seen in Figure 6, each peptide had similar effects against both S. mutans and S. aureus at 0.1 mg/ml in the microplate crystal violet assay. None of the peptides with alterations in the ‘AAT’ repeats (P0, P1–4, or P1–7) reduced biofilm formation. The only variant peptide that showed some anti-biofilm activity was P17, which is a scrambled version of P1 that lacks the proline residue, and contains the three positively charged amino acids (R,R,K) at the C-terminus. Its effect was slightly less potent than that of P1.

In addition to sharing the 24-aa length and +3 net charge at the terminus, P1 and P17 displayed another common property. These peptides form visible aggregates in BHI-S solution, even without bacteria present. Figure 7 shows microscopic images of each peptide in solution, with and without S. mutans bacteria, at 6 and 24 h of incubation. Figures 7A,C display the conspicuous self-aggregation of peptides P1 and P17 (arrows). After the bacterial biofilm forms, the wells treated with P1 and P17 display floating chunks of matrix material that appear to associate directly with the aggregated peptide.

The ability to self-aggregate indicates that P1 and P17 may have different secondary and tertiary structure than the other peptides, depending on the relative locations of proline and the charged and uncharged amino acids in each sequence (Figure 6). We used Fourier transform infrared spectroscopy (FTIR) to assess the secondary structure of these two peptides compared to the negative control, sP1 (Haris and Chapman, 1995). The assembly of P1 and P17 peptide monomers into higher order β-sheet structures were confirmed by the presence of a peaks at 1621 and 1695 cm^-1 (Figure 8), characteristic for antiparallel β-sheet formation (Cerf et al., 2009).