Peptide and non‐peptide mimetics as potential therapeutics targeting oral bacteria and oral biofilms

Abstract The development of the oral biofilm requires a complex series of interactions between host tissues and the colonizing bacteria as well as numerous interspecies interactions between the organisms themselves. Disruption of normal host–microbe homoeostasis in the oral cavity can lead to a dysbiotic microbial community that contributes to caries or periodontal disease. A variety of approaches have been pursued to develop novel potential therapeutics that are active against the oral biofilm and/or target specific oral bacteria. The structure and function of naturally occurring antimicrobial peptides from oral tissues and secretions as well as external sources such as frog skin secretions have been exploited to develop numerous peptide mimetics and small molecule peptidomimetics that show improved antimicrobial activity, increased stability and other desirable characteristics relative to the parent peptides. In addition, a rational and minimalist approach has been developed to design small artificial peptides with amphipathic α‐helical properties that exhibit potent antibacterial activity. Furthermore, with an increased understanding of the molecular mechanisms of beneficial and/or antagonistic interspecies interactions that contribute to the formation of the oral biofilm, new potential targets for therapeutic intervention have been identified and both peptide‐based and small molecule mimetics have been developed that target these key components. Many of these mimetics have shown promising results in in vitro and pre‐clinical testing and the initial clinical evaluation of several novel compounds has demonstrated their utility in humans.

inflammatory disease that leads to the destruction of periodontium and ultimately results in resorption of the underlying alveolar bone and tooth loss (Darveau, 2010). The incidence of periodontal disease ranges from 30% of the population in developed countries to over 70% of the population in developing countries, with severe disease inflicting 7%-15% of human population worldwide (Eke, Dye, Wei, Thornton-Evans, & Genco, 2012;Gera, 2000;Hugoson, Sjondin, & Norderyd, 2008). In the U.S., the total prevalence of periodontitis in adults aged 30 and older was 47.2%, representing approximately 64.7 million adults (Eke et al., 2012). Periodontal diseases are also associated with many systemic diseases and conditions, including coronary artery disease, rheumatoid arthritis, diabetes, pulmonary diseases, cancers and Alzheimer's disease (Dominy et. al., 2019;Hajishengallis, 2015;Maddi & Scannapieco, 2013;Whitmore & Lamont, 2014). Treatment of periodontal disease involves the physical removal of dental plaque from the subgingival pocket by scaling and root planning and gingival surgery to reduce pocket depth if necessary.
Although the oral cavity carries a high microbial load, incidental tissue damage or tissue damage resulting from minor surgical procedures rarely results in infection, suggesting that potent antimicrobial defence mechanisms exist in the oral cavity. An important aspect of innate immunity that may contribute to these defence mechanisms involves the secretion of antimicrobial peptides and proteins by oral epithelial cells, neutrophils, salivary secretions and gingival crevicular fluid. More than 45 types of antimicrobial peptides have been identified in saliva (Gorr, 2009). Many of these antimicrobial components exhibit a broad spectrum of activity in vitro, suggesting that they may represent promising therapeutics against oral bacteria. However, therapeutic application of antimicrobial peptides has been limited since the natural concentration of many oral antimicrobial peptides in saliva and gingival crevicular fluid is well below the minimal inhibitory concentration for many organisms and in higher doses, these peptides have shown haemolytic activity damaging mammalian cells (Aoki, Kuroda, & Ueda, 2012). In addition, antimicrobial peptides may be susceptible to proteolytic degradation and many of the organisms that are associated with periodontal disease are highly proteolytic. These limitations have spurred research to develop antimicrobial peptide mimetics in order to increase their activity and stability.
The oral cavity is colonized by a diverse group of microorganisms that can comprise over 700 species of bacteria, viruses and fungi (Dewhirst et al., 2010;Paster, Olsen, Aas, & Dewhirst, 2006). Development of this community is a dynamic process involving attachment of bacteria to oral surfaces, cohesion and communication among constituent organisms, and adaptation to the biofilm environment through direct contact, by intra-and interspecies signalling via soluble mediators and/or by nutrient transfer (Kolenbrander et al., 2002;Kolenbrander, Palmer, Periasamy, & Jakubovics, 2010;Kuboniwa et al., 2017). Within the oral ecosystem, bacteria are able to collectively regulate activities and functional specialization is present. With an increased understanding of the complex interspecies and host-microbe interactions that occur and contribute to the formation of the oral biofilm, new potential therapeutic targets to disrupt biofilm development have been identified. This in turn has led to the development of peptides and small molecule peptidomimetics that target specific oral bacteria. This review will summarize some of the efforts to exploit host antimicrobial peptides and the molecular processes contributing to biofilm growth to develop peptide mimetics and small molecule compounds that target oral bacteria and/or the formation of the oral biofilm.

| Histatin 5 mimetics Dhvar4 and Nal-P-113
The histatins are a family of 12 small multifunctional proteins produced by the parotid and submandibular salivary glands (Tsai & Bobek, 1998). The histatins are generated by proteolytic cleavage of two gene products, histatin 1 and histatin 3 and function primarily as antifungal agents although they have broader antimicrobial activities as well. Histatin 5 is a cationic peptide of 24 amino acids that is generated from cleavage of histatin 3 and represents a platform that has been used to synthesize peptide mimetics with improved activity and stability. For example, dhvar4 and dhvar4a are derivatives of the C-terminal 12 amino acids of histatin 5 (Helmerhorst et al., 1999) in which the His residues have been substituted with by hydrophobic amino acids (Phe or Leu) and the penultimate Gly replaced with Lys (Table 1). In dhvar4a, the C-terminus was further modified by amidation. Mimetic dhvar4 (and dhvar4a) exhibits strongly enhanced antibacterial activity compared with histatin 5 against pure cultures of a broad range of oral pathogens. Viable cell counts were reduced by >3 logs for oral streptococci (S. mutans, S. sanguis, S. salivarius), Actinomyces naeslundii and Fusobacterium nucleatum, 2 logs for Veillonella parvula and 0.92 logs for Prevotella intermedia (Helmerhorst et al., 1999). In contrast, histatin 5 had little effect on viable cell counts (log reduction 0.0-0.3). Dhvar4 (100 µg/ml) was also tested against a seven species biofilm grown on hydroxyapatite disks in continuous culture using the organisms referenced above, as well as multispecies biofilms formed by inoculating disks with human saliva or dental plaque samples. Dhvar4 was significantly less effective in reducing viable cell counts of bacteria in the seven species biofilm (0.47-0.76 log reduction). Interestingly, dhvar4 was more effective against obligate anaerobic organisms (F. nucleatum, V. parvula and P. intermedia) in the plaque and saliva biofilms. For all biofilm experiments, chlorhexidine was used as a positive control and significantly reduced viable cell counts (2.5-3.0 log reductions). These results suggest that dhvar4 is more potent against planktonic cells and is significantly less active against biofilms.
By screening a library of 25 peptides that represented various portions of histatin 5, Rothstein et al. (2001) identified peptide P-113 (Table 1), encompassing residues 4-15, as the smallest fragment of histatin 5 that retains the full anticandidal activity of the parent protein. Sajjan et al. (2001) subsequently showed that P-113 also retained the antibacterial activity of histatin 5. Amidation of the P-113 C-terminus increased antimicrobial activity of the mimetic by twofold but substitution of two or more of the five basic residues in P-113 with neutral amino acids resulted in a significant reduction in anticandidal activity, indicating that the cationic character of histatin 5 is essential for activity. Given the potent anticandidal activity of P-113, a clinical study was undertaken using a cohort of HIV patients and showed that treatment with an oral rinse containing P-113 was effective in reducing oral candidiasis (Helmerhorst, Oppenheim, Choi, Cheng, & Reiner, 2007). However, broad application of P-113 therapy was limited since its activity in vitro and in vivo was significantly reduced under conditions of high osmotic strength (Helmerhorst, Hof, Veerman, Simoons-smit, & Amerongen, 1997) and furthermore, P-113 was susceptible to proteolytic degradation when suspended in biological fluids such as serum, saliva and sputum (Sajjan et al., 2001). One approach to increase P-113 stability was to synthesize a derivative of P-113 that comprised D-amino acids (designated P-113D) (Sajjan et al., 2001). P-113D retained the antimicrobial activity of the parent peptide and was stable for 7 days in sputum obtained from cystic fibrosis patients, whereas the half-life of the parent P-113 in sputum was 2.8 min (Sajjan et al., 2001). A second approach intended to increase activity at high ionic strength involved replacing the three His residues in P-113 with bulky hydrophobic amino acids such as Phe, β-naphthylalanine (Nal), β-diphenylalanine (Dip) or β-(4,4'-biphenyl)alanine (Bip) (Yu et al., 2011). Nal-P-113 and Bip-P-113 were shown to retain antibacterial activity at salt concentrations up to 200 mM whereas the parent P-113 peptide lost activity at a salt concentration below 50 mM (Yu et al., 2011). Nal-P-113 also exhibited antimicrobial activity against oral organisms including P. gingivalis (MIC = 20 µg/ml), F. nucleatum (MIC = 40 µg/ml) and S. gordonii (MIC = 80 µg/ml) but consistent with the previous anti-biofilm activity reported for the dhvar4 mimetics, Nal-P-113 was less potent against single or three species oral biofilms (MIC = 320-640 µg/ml and 1,280 µg/ml, respectively) (Wang, Cheng, et al., 2015). Nal-P-113 cytotoxicity against periodontal ligament cells and gingival epithelial cells was minimal at concentrations <320 mM but apoptosis and cell death increased significantly at higher concentrations. More recently, the activity of Nal-P-113 has been examined using in vivo approaches. In a rat model of periodontitis, treatment with 100 µg/ml and 400 µg/ml Nal-P-113 resulted in reduced P. gingivalis-mediated alveolar bone loss and a reduction in P. gingivalis levels and in total oral microbial load (Wang, Lin, et al., 2017). In addition, a recent clinical study showed that treatment of chronic periodontitis patients with 20 µg/ml Nal-P-113 reduced the levels of P. gingivalis, S. gordonii, F. nucleatum and T. denticola in subgingival plaque, however, significant reductions in pocket depth, clinical attachment loss or bleeding index were not detected (Wang, Ai, et al., 2018). A possible explanation for this observation is the short duration of the study (7 days). Finally, incorporating bulky hydrophobic amino acids into P-113 was also recently shown to enhance the LPS neutralizing activity of the peptides (Chih et al., 2019).

| IDR-1018, DJK and mimetics derived from bactenecin
Peptide IDR-1018 is a 12 residue peptide derived from bovine bactenecin that was initially characterized as an innate defence regulator (IDR) peptide that functioned to modulate innate immunity, enhance the induction of chemokines and suppress harmful inflammatory responses (Rivas-Santiago et al., 2013). The peptide exhibited only weak antimicrobial activity against Mycobacterium tuberculosis but significantly reduced microbial load and lung inflammation in infected animals. However, in a subsequent screen to identify peptides with anti-biofilm activity, IDR-1018 was shown to specifically target and kill biofilm cells more potently that other known antibiofilm peptides (de la Fuente-Nunez, Reffuveille, Haney, Straus, & Hancock, 2014). IDR-1018 was shown to interact with and degrade the stringent response mediator ppGpp and the peptide exhibited broad spectrum activity that both inhibited the formation of biofilms and eradicated mature existing biofilms formed by Pseudomonas aeruginosa, Escherichia coli, Acinetobacter baumannii, Klebsiella pneumoniae and Staphylococcus aureus at concentrations that had no effect on planktonic cells. A series of peptide analogues highlighted several functionally important residues of IDR-1018 (de la Fuente-Nunez et al., 2014). Interestingly, reversing the sequence of IDR-1018 in peptide HE1 (Table 1) had little effect on anti-biofilm activity. In contrast, altering the hydrophobic LIVAV sequence between residues 3 and 7 of the peptide by substituting R for V 5 (peptide HE2, Table 1) or deleting VA 5,6 (peptide HE10, Table 1) reduced activity against TA B L E 1 Antimicrobial peptide-based mimetics Nal-P−113 Ac-AKR-Nal-Nal-GYKRKF-Nal-CONH2 Gram-negative organisms but increased anti-biofilm activity against Gram-positive organisms for HE10. de la  subsequently synthesized and tested a panel of D-enantiomer peptides based on the known properties of IDR-1018. Two D-peptides, DJK5 and DJK6 (  (Freitas et al., 2017).
The MIC for planktonic C. albicans and CI cultures was 32 µg/ml and 64 µg/ml, respectively. In contrast, the minimum biofilm inhibitory concentration (MBIC) was 4 µg/ml and 8 µg/ml, respectively. To assess the potential therapeutic utility of IDR-1018, the activity of the peptide was assessed using a mouse candidemia model. Animals were infected with 10 6 C. albicans cells and after 24 hr, mice were treated daily with IDR-1018 at 10 mg/kg body weight. Treated mice showed a significant increase in survival, with 50% of the animals surviving after 8 days. These animals also had reduced candidal burden in the kidneys relative to control mice. In contrast, all of the control animals treated with buffered saline died after 6 days (Freitas et al., 2017).
Recently, IDR-1018 has been further modified in order to specifically target the peptide to enamel (Yang et al., 2019). To accomplish this, IR-1018 was synthesized containing either a high affinity or a medium affinity hydroxyapatite-binding heptapeptide (Gungormus et al., 2008) at its N-terminus to generate peptides SHABP and MHABP respectively (see Table 1). Consistent with the prior activity profile of IDR-1018, neither SHABP nor MHABP inhibited the growth of planktonic supragingival cultures. Both peptides significantly reduced biofilm biovolume and increased killing of biofilm cells upon both short-(3-9 min) and long-term (1-3 days) exposure (Yang et al., 2019), however, SHABP was more effective than MHABP or the control IDR-2018 peptides. Finally, to determine if anti-biofilm activity occurred with biofilms growing on other material surfaces that may be present in the oral cavity, the activity of DJK5 and IDR-1018 peptides against biofilms grown on hydroxyapatite and titanium surfaces was compared. In both cases, DJK5 was more effective and killed more bacteria than IDR-1018, however, there was no significant difference in the extent of killing by either peptide on hydroxyapatite versus titanium surface (Wang, Haapasalo, Gao, Ma, & Shen, 2018). In summary, these results suggest that IDR-1018 and some of its derivative peptide mimetics, especially DJK5 and DJK6, may represent novel therapeutics for treating chronic biofilm infections at a variety of tissue sites and in the oral cavity.
Peptide Bac8c is also a mimetic that was derived from bactenecin (Table 1) and is broadly active against a range of oral organisms including mutans streptococci, various commensal streptococcal species, Actinomyces sp. and Lactobacilli (Ding et al., 2014). MICs ranged from 16-32 µg/ml for oral streptococci and 8-16 µg/ml for

| Dermaseptin K 4 -S4 (1-15)a
Dermaseptin S4 is a member of a large family of antimicrobial peptides ranging in size from 28 to 34 amino acids that are expressed in amphibian skin (Navon-Venezia, Feder, Gaidukov, Carmeli, & Mor, 2002). The peptide displays broad spectrum activity against a range of bacteria, fungi and parasites but is also highly toxic towards human erythrocytes. To maintain the antimicrobial properties of the peptide and limit its toxicity, several truncated peptide mimetics have been synthesized and evaluated (Feder, Dagan, & Mor, 2000). Dermaseptin K4-S4(1-15)a (see Table 1) has been tested against planktonic cultures of a variety of oral organisms (Altman et al., 2006), including several strains of S. mutans, S. sobrinus, Lactobacilli, A. viscosus, A. actinomycetemcomitans, F. nucleatum and P. gingivalis. The MICs for these organisms ranged from 5 to 20 µg/ ml and K4-S4(1-15)a was significantly more active against these organisms than the human antimicrobial peptide LL37 (Altman et al., 2006). However, P. gingivalis was resistant to K4-S4(1-15)a. Although K4-S4(1-15)a was susceptible to gingipain-mediated cleavage, a Denantiomer analogue that was resistant to cleavage retained activity against S. mutans and other organisms at the level of the L-peptide but was still inactive against P. gingivalis (Bachrach et al., 2008). K4-
To further improve antimicrobial activity and reduce potential toxicity, several peptide analogues were synthesized in which N-terminal amino acids of K4-S4(1-15)a were replaced with short fatty acids (Porat, Marynka, Tam, Steinberg, & Mor, 2006). Deletion of 1-3 N-terminal amino acids resulted in a progressive loss of antibacterial activity and a reduction in α-helical character of the peptide, however, replacing the deleted residues with fatty acids resulted in recovery of antimicrobial activity, faster bacteriocidal kinetics and restoration of α-helical structure. The most potent analogues possessed heptanoic (C 7 -S4(1-13) or aminododecanoic acid (NC 12 -S4(1-13) moieties replacing the N-terminal Ala-Leu residues of K4- Table 1). These analogues exhibited MICs between 2 and 9 µM for B. cereus, S. aureus, P. aeruginosa and E. coli and between 8 and 32 µM for oral organisms S. mutans, A. viscosus and A. actinomycetemcomitans. In addition, treatment of the polymicrobial flora present in human saliva with 100 µM C 7 -S4(1-13) resulted in a decrease in CFUs of >3 logs within 2 min .
Finally, mature single species S. mutans or A. viscosus biofilms were eradicated upon 18 hr exposure to 100 µM C 7 -S4(1-13). These results suggest the combined N-terminal truncation and acylation may functionally improve K4-S4(1-15)a and related cationic antimicrobial peptides and their derivatives.

| L-K6 peptide
Temporin-1CEb is a naturally occurring 12-residue peptide that was isolated from skin secretions Rana chensinensis. The peptide exhibited bactericidal activity against Gram-positive organisms but was weakly active against Gram-negative bacteria and like many natural antimicrobial peptides, temporin-1CEb exhibited significant haemolytic activity. To counter these limitations, Shang, Li, et al. (2012) and  designed several temporin-1CEb analogues with increased cationic character and decreased hydrophobicity by substituting Lys residues for neutral and/or non-polar amino acids of the peptide. The most potent analogue, L-K6 (Table 1)

| ZXR-2 peptide
Peptide ZXR-1 encompasses the N-terminal 16 residues of mauriporin, a 48 amino acid peptide isolated from the venom gland of the Moroccan scorpion Androctonus mauritanicus and was initially synthesized as an anti-cancer peptide that induced apoptosis in a variety of cancer cell lines (Zhou et al., 2016). ZXR-1 forms an amphipathic α-helix with clearly defined hydrophilic and hydrophobic surfaces, with the exception that Lys 14 resides on the hydrophobic face of the peptide. In an attempt to improve its activity, Zhou et al. (2016) synthesized a new peptide, ZXR-2, in which Lys 14 was replaced with Leu (Table 1). Interestingly, while this single change increased overall cytotoxic activity, it also altered the mechanism of action of the peptide. ZXR-1 functioned as a pro-apoptotic peptide whereas ZXR-2 disrupted the cell membrane leading to lysis. As a re- poorly penetrates the S. mutans biofilm and thus may be relatively ineffective for eliminating or reducing dental plaque.

| Designer synthetic α-helical peptides KSL and GH12
Naturally occurring antimicrobial peptides often have several limitations for use as potential therapeutic agents in that they are often easily degraded, the peptides in their active form can exhibit sig- to >512 µg/ml) and GH16 was more effective against Lactobacilli (MIC = 8 µg/ml) but was poorly active against the other organisms (MICs from 32 to >512 µg/ml). In contrast, GH12 (Table 1)  vicR, liaR and comDE . Subsequently, Jiang et al. (2018) showed that treatment of a three species biofilm comprising S. mutans, S. gordonii and S. sanguinis with 8 µg/ml GH12 reduced the abundance of S. mutans and reduced the level of water-insoluble glucan . Interestingly, as the abundance of S. mutans decreased in the treated biofilms, the level of S. gordonii increased and became the dominant organism in the biofilm by 48 hr.
Recently, GH12 was fused with peptide TD7 (TKREEVD) derived from the hydrophilic C-terminal region of amelogenin which functions to promote hydroxyapatite nucleation and the growth of enamel crystals (Li et al., 2014) in order to create a bifunctional peptide with both antibacterial and mineral promoting activity . The initial fusion protein exhibited poor antibacterial activity but substituting Gln for Glu 4,5 and Val for Asp 7 in TD7 to generate peptide TVH19 (Table 1) significantly increased activity against S.
mutans. The TVH19 fusion peptide was stable in saliva for 12 hr and was non-toxic towards human oral keratinocytes. To examine the ability of TVH19 to promote remineralization, enamel slices derived from bovine incisors were first incubated in an acidic buffer and then transferred to remineralization buffer containing 1 × MBC TVH19 or 1,000 ppm NaF. The TVH19-treated and NaF-treated groups exhibited significantly higher surface hardness than the deionized water-treated control group. In addition, the radiopaque remineralizing zone of the enamel slice was thicker on samples treated with TVH19 or NaF but did not change significantly on the water-treated control slices. Significantly greater mineral deposition and shallower lesions were observed in the TVH19-treated group or NaF-treated group than in the water-treated control . These results suggest that bifunctional fusion peptides may both target S. mutans and promote remineralization of initial caries and have utility in the treatment of dental caries.

| TARG E TED PEP TIDE MIME TI C S
Although the oral microbiome may contain >700 bacterial species, a relatively small number of species have been associated with the common oral biofilm-mediated diseases, caries and periodontitis. In addition, some of the organisms that are associated with periodontal disease have been suggested to induce dysbiosis in the gut microbiome (Olsen & Yamazaki, 2019). The remaining oral organisms are benign commensals although some species may play a beneficial role in the host (Walker et al., 2018). Caries and periodontal disease are generally treated using procedures that reduce both pathogens and commensals. In some cases, periodontitis can be treated using antibiotics, most of which are also broad spectrum indiscriminate antimicrobial agents. The non-specific nature of these treatment regimens has several potential undesirable side effects, including increased chance for secondary opportunistic infections and/or the development of drug resistance. This has stimulated research to develop novel pathogen-targeted, highly specific antimicrobials that function to reduce or eliminate organisms that are associated with disease without significantly affecting commensal populations.

| STAMPs
Some early attempts to create targeted therapeutics involved the use of monoclonal antibodies as a targeting moiety, or generated fusion proteins where an antimicrobial peptide was linked to a bacterial recognition domain (Qiu et al., 2003;Qiu, Zhang, Wang, & Wu, 2005  Similar results were found after exposing single species biofilms for 1 min to C16G2 or controls (CSP or G2 peptides); biofilm growth of the commensal species was indistinguishable from an untreated biofilm, whereas biofilm growth of S. mutans was significantly inhibited after exposure to C16G2 but not the control peptides. Furthermore, short-term treatment with C16G2 but not control peptides reduced S. mutans levels in a mixed species community in saliva (Eckert et al., 2006) and recolonization of C16G2-treated saliva biofilms by S. mutans was inhibited (Li et al., 2010). These results indicate that short-term exposure to C16G2 is effective in reducing S. mutans populations in both pure culture and a mixed species community.
Finally, C16G2 was shown to selectively kill S. mutans in a human saliva-derived multispecies community (Guo et al., 2015). Interestingly, re-growth of the treated microbial community showed that species with metabolic dependency with S. mutans (e.g., Veillonella atypica) were drastically reduced in abundance whereas commensal streptococcal species and other species associated with health became dominant (Guo et al., 2015). This result suggested that C16G2-me- The basic synthetic scheme used to generate C16G2 was also applied to optimize each of the domains contributing to activity  and to generate other novel STAMP analogues (He, Anderson, Shi, & Eckert, 2009  . This approach generated a diverse number of STAMP sequences and identified mechanisms that may limit STAMP effectiveness as well as strategies to overcome these limitations. In addition, by modifying the linker domain, multiple targeting sequences that are specific for different organisms were synthesized (He et al., 2009). This opens the possibility of using a single STAMP to simultaneously eliminate several species from multispecies cultures or biofilms. Finally, these general approaches are not only limited to the development of potential therapeutics that target S. mutans, but can be generally applied to any organisms for which a suitable targeting moiety has been identified.

| Streptococcal-derived Anti-P. gingivalis peptide (SAPP)
The interaction of P. gingivalis with S. cristatus has been reported to reduce the expression of fimA and reduce P. gingivalis biofilm formation without affecting its planktonic growth (Christopher, Arndt, Cugini, & Davey, 2010;Xie, Lin, Wang, Wu, & Lamont, 2007). This interaction has been subsequently shown to be mediated by a protein-protein interaction between the S. cristatus protein ArcA encoding arginine deiminase and P. gingivalis RagB and the product of the PGN_08056 gene (Ho, Lamont, & Xie, 2017a). Interestingly, repression of fimA expression does not require enzymatic activity.
These results highlight the potential of ArcA to target P. gingivalis and attenuate virulence. By using a peptide array generated from the ArcA sequence, Ho et al. (2017a) identified five ArcA-derived peptides that directly interact with P. gingivalis, the most potent of which was NIFKKNVGFKK representing residues 249 -259 of ArcA.
A synthetic peptide comprising this sequence, designated SAPP ( with P. gingivalis reduced the formation of monospecies biofilm on a saliva-coated surface by up to 70% after 48 hr (Ho, Lamont, & Xie, 2017b). In addition, the formation of dual species biofilms between P. gingivalis and S. gordonii was reduced in the presence of SAPP and furthermore, treatment of a pre-existing in vitro dual species biofilm resulted in a significant reduction in adhered P. gingivalis. This phenotype is consistent with the repression of mfa and fimA since both of these fimbrial proteins are involved in the interaction of P. gingivalis with S. gordonii. A similar reduction in P. gingivalis abundance was observed when a multi-species plaque-derived biofilm was incubated with SAPP (Ho et al., 2018). Interestingly, the abundance of other potential periodontal pathogens, including T. forsythia, F. nucleatum and T. denticola were also reduced by SAPP, however, the levels of A. actinomycetemcomitans and streptococci were unchanged. Finally, SAPP affected other fimbrial-and gingipain-mediated phenotypes of P. gingivalis including inhibiting the invasion of epithelial cells, reducing cell surface gingipain activity, and inhibiting the ability of P. gingivalis to selectively impair the production of IL-8 (Ho et al., 2017b). SAPP is stable in PBS for >24 hr but was gradually degraded over 48 hr in saliva. The peptide also exhibits minimal toxicity towards human oral keratinocytes or periodontal ligament cells (Ho et al., 2018). These characteristics suggest that SAPP may be compatible for formulations for oral applications in humans.
gingivalis to commensal streptococci and that adherence is mediated by a protein-protein interaction between the minor fimbrial antigen (Mfa) of P. gingivalis and antigen I/II (AgI/II) of streptococci (Chung, Demuth, & Lamont, 2000;Lamont et al., 2002;Park et al., 2005). Structure function studies also identified a discrete motif in AgI/II that was essential for P. gingivalis adherence to streptococci.
To further define the structural constraints of NITVK that are important for interaction with P. gingivalis, a combinatorial approach was undertaken and three peptide libraries were synthesized in which one of the functionally active residues of NITVK (N 1182 , T 1184 , or V 1185 ) was replaced by each of the 19 common amino acids (cysteine excluded). The remaining two amino acid positions were randomized by using equimolar mixtures of the 19 common amino acids during synthesis. Thus, each combinatorial library contained 19 distinct peptide mixtures that differed only in the amino acid residue that occupied the "z" position (see Table 2B) and each of the 19 mixtures in a given library contained 361 different peptide sequences, arising from randomization of the two amino acids occupying the "x" positions. As shown in Table 2C, substitutions of positively charged residues for N 1182 and hydrophobic residues for V 1185 improved adherence. Substitutions for T 1184 showed little increase in P. gingivalis adherence (Daep et al., 2006). These observations were subsequently confirmed with a synthetic peptide containing N/R 1182 and V/I 1185 substitutions which exhibited 2.5-fold greater inhibitory activity (IC 50 = 0.5 µM) than BAR (Daep et al., 2008).
Substitution of Pro or Gly for N 1182 or V 1185 reduced P. gingivalis adherence (Table 2C), suggesting that the α-helical charac-  Table 2B).
Unexpectedly, this peptide was 13-fold less effective than the parent BAR peptide (Daep at al., 2008). It was subsequently determined that introducing Cys for L 1177 disrupted a second motif (VQDLL) that was essential for P. gingivalis adherence (Daep et al., 2008). Indeed, the VQDLL motif, flanked by Lys residues, resembled the core LXXLL consensus sequence of the nuclear receptor (NR) box, a eukaryotic protein-protein interaction domain. Thus, BAR may represent the prokaryotic counterpart of the eukaryotic NR box (Daep et al., 2008). A second constrained peptide (CR2-BAR, Table 2B), in which both VXXLL and NITVK remain intact, exhibited > 2-fold increased activity relative to BAR, confirming the role of the VXXLL motif and suggesting that secondary structure of BAR is important for activity.
Potential obstacles for developing potential peptide therapeutics for oral applications include the high cost of synthesis and their susceptibility to proteolytic degradation. This is especially a concern for peptides that target periodontal organisms, many of which are highly proteolytic. Indeed, BAR contains five Lys residues and is likely to be a substrate for the Lys-gingipain. One approach to address this potential limitation involved the synthesis of an analogue of BAR in which D-Lys residues replaced the five L-Lys in the peptide. This peptide was stable when incubated with P. gingivalis or with purified Kgp and exhibited a twofold increased specific inhibitory activity (IC 50 = 0.7 µM) relative to the parent peptide (Daep, Novak, Lamont, & Demuth, 2010). However, although this peptide exhibited increased stability and activity, it was expensive to synthesize. These limitations for BAR and other peptide active agents have led several groups to explore the development of small molecule peptidomimetics as stable and cost-effective alternatives.

| Meta-phenylene ethylene (mPE)
Meta-phenylene ethylene (mPE) is a novel non-peptide mimetic whose design was based on the amphiphilic structure of magainin, an antimicrobial peptide that was initially isolated from the skin of the African frog Xenopus laevis (Zasloff, 1987). mPE exhibits antimicrobial activity against several oral pathogens, including P. gingivalis, S. mutans and A. actinomycetemcomitans, exhibiting MICs within the range of 0.4 µg/ml for A. actinomycetemcomitans to 2.5 µg/ml for P.
gingivalis (Beckloff et al., 2007). Incubation of planktonic S. mutans with 10 x MIC mPE (~5 µg/ml) resulted in a 2 log reduction of viable organisms after 0.5 min exposure and no viable cells remained after 1 hr. In addition, the growth of in vitro S. mutans biofilms formed in the presence of sucrose was significantly inhibited by 2.5 µg/ml mPE and the incubation of established S. mutans biofilms with 50 µg/ml mPE resulted in a 3 log reduction of viable biofilm cells. mPE remains active in the presence of human saliva and its activity also extends to fungal pathogens, as mPE was shown to inhibit the growth of several Candida species in the concentration range of 0.5 to 1 µg/ml.
More recently, the activity of mPE was tested using a biofilm model against two bacteria associated with periodontitis, A. actinomycetemcomitans (Kaplan, Meyerhofer, & Fine, 2003) and P. gingivalis (Davey, 2006). Metabolic activity and microbial biomass measurements demonstrated that mPE rapidly killed A. actinomycetemcomitans, reducing overall metabolic activity by 60% after 2 hr exposure to mPE (Hua, Scott, & Diamond, 2010). Similarly, a decrease in both metabolic activity and viable biofilm bacteria was observed upon exposure of P. gingivalis biofilms to 4-8 µg/ml of mPE (Hua et al., 2010).
Interestingly, 2 µg/ml of mPE also inhibited interleukin 1β-induced secretion of IL-8 in gingival epithelial cells in a dose-dependent manner and functioned as an anti-inflammatory agent in infected tissues, similar to the functions of the L-K6 mimetic. The antibacterial activity of mPE against both planktonic and biofilm cultures along with its ability to suppress cytokine production in gingival tissue and lack of cytotoxicity suggests that mPE may represent a novel broad spectrum potential therapeutic against a variety of oral pathogens.

| Peptidomimetics of BAR peptide
An alternative approach to generate stable inhibitors of P. gingivalis/ streptococcal adherence in a cost-effective manner involved the design and synthesis of small molecule mimetics of BAR peptide. Patil, Tan, Demuth, and Luzzio (2016) selected the 2,4,5-trisubstituted oxazole scaffold as the peptidomimetic model of the NITVK motif in BAR peptide. The oxazole ring comprised a central torus where two aromatic rings, located at the 4 and 5 positions of the heterocycle, bear hydrophobic residues in the form of halogens, alkoxy groups or alkyl groups to mimic the essential hydrophobic character of this functional motif (Daep et al., 2006). The NITVK mimetics also contained an azide moiety that facilitated coupling to the VXXLL mimetic compounds. The initial VXXLL mimetics comprised a simple aromatic ring to mimic the hydrophobic character of the VXXLL amphipathic helix. These compounds also contained an alkynyl group to facilitate coupling to the NITVK azide group via the click reaction (Takayama, Kusamori, & Nishikawa, 2019) to produce the first generation 1,2,3-triazole-based inhibitors of P. gingivalis adherence.
A total of 50 first generation compounds were tested for the inhibition of P. gingivalis adherence and formation of dual species biofilms and four compounds exhibited potent activity (IC 50 = 5-8 µM) (Patil et al., 2016). These compounds were subsequently shown to inhibit adherence of P. gingivalis to streptococci in the presence of F. nucleatum, a bridging organism that can independently interact with both P. gingivalis and oral streptococci. In addition, the most potent first generation compound was capable of disrupting a preexisting three species biofilm (Tan, Patil, Luzzio, & Demuth, 2018).
All four of the active first generation mimetics were subsequently shown to inhibit P. gingivalis virulence in vivo and exhibited minimal levels of toxicity towards various human and mouse cell lines (Tan et al., 2018). Recently, Patil, Tan, Demuth, and Luzzio (2019) reported the synthesis of second generation 1,2,3-triazole-based inhibitors which utilized a 1,3,5-trisubstituted-2,4,6-triazine scaffold to mimic the VXXLL motif. This scaffold was based on a structure that was used to mimic the core LXXLL motif in the eukaryotic NR box. The most potent second generation peptidomimetics showed improved activity relative to first generation compounds (IC 50 = 2-4 µM), most likely because the 1,3,5-trisubstituted-2,4,6-triazine scaffold more closely mimics the VXXLL motif than the aromatic ring used in first generation compounds. These second generation mimetics exhibited minimal toxicity towards human gingival keratinocytes (Patil et al., 2019) and significantly reduced P. gingivalis-mediated alveolar bone loss in infected mice (D. R. Demuth, unpublished). Thus, stable small molecule mimetics of BAR peptide have been developed and represent cost-effective potential novel therapeutics that may limit P. gingivalis colonization of the oral cavity.

| Structure-based and functional screening of small molecule libraries
The 20%. Treatment with 2A4 was shown to reduce the expression of three biofilm-associated genes in S. mutans and reduced cell surface expression of AgI/II and GTF but had no effect on their analogues in commensal streptococci. Subsequently, Pan, Fan, Wu, Melander, and Liu (2015) demonstrated that treatment with a 2-aminoimidazole/triazole conjugate prevented S. mutans biofilm accumulation in a mouse biofilm model whereas significant S. mutans accumulation occurred in control animals.
The same small molecule library was screened by Wright, Wu, Melander, Melander, and Lamont (2014) to identify inhibitors of P.
gingivalis/S. gordonii biofilm formation. Interestingly, two of the 2aminoimidazole or 2-aminobenzamidazole compounds identified by Liu, Worthington, et al. (2011a) also appeared to inhibit the formation of P. gingivalis/S. gordonii biofilms, the most potent of which was 2A4 (IC 50 = 3.4 µM). The inhibitors identified had no effect on planktonic growth of P. gingivalis but did repress the expression of fimA and mfa encoding the major and minor fimbrial subunit proteins. Thus, 2-aminoimidazoles or 2-aminobenzamidazoles may represent broad spectrum anti-biofilm compounds that function in part by altering the expression of genes that are essential for biofilm formation.
A structure-based screen has also identified small molecules inhibited S. mutans biofilms by 85% at 12.5 µM. G43 significantly reduced glucan production and inhibited both GtfB and GtfC by 80%. The compound had no effect on planktonic growth or viability of S. mutans, commensal streptococci, A. actinomycetemcomitans or A. viscosus and did not affect biofilm formation of S. sanguinis or S. gordonii. In contrast, G43 significantly reduced the abundance of S. mutans in mixed species biofilms. Finally, treatment with G43 significantly reduced caries scores in a rat model of dental caries and did not exhibit overt toxicity in the treated animals. Thus, both functional and structure-based small molecule screens identified targeted, highly potent compounds that represent potential novel anti-S. mutans therapeutics.
The screens described above identified anti-biofilm compounds that had little or no effect on planktonic cell growth, however, small molecule libraries can also be useful for identifying bactericidal compounds that target essential gene products. For example, Xu et al.
(2011) developed a system to predict essential genes in S. sanguinis and subsequently applied this system to P. gingivalis W83, resulting in the identification of 212 essential genes. One of these genes encoded meso-diaminopimelate dehydrogenase (m-Ddh) that functions in the lysine biosynthetic pathway of P. gingivalis and is essential for protein and cell wall biosynthesis. Experimental inactivation of the gene encoding this enzyme (PG0806) was subsequently shown to be lethal (Stone et al., 2015). Thus, m-Ddh represents an ideal target for therapeutic intervention. Stone et al. (2015) used a structure-based screen to identify potential inhibitors of m-Ddh by screening the ZINC 3D database (Irwin & Schoichet, 2005) and initially identified 11 commercially available potential inhibitors. Of the 11 compounds, four resulted in >90% inhibition of m-Ddh when incubated with purified enzyme at a concentration of 3 mM and exhibited IC 50 values for the inhibition of enzymatic activity between 157 µM and 1.1 mM.
In the three most active compounds (4, 5 and 6), each contained a sulfonamide scaffold and were subsequently tested for inhibition of P. gingivalis planktonic growth. MIC/MBC for compounds 4, 5 and 6 were 250/374 µM, 167/254 µM and 2821/ND µM respectively. At 5 × MIC, compound 4 reduced P. gingivalis CFU by 2 logs after incubation for 6 hr, whereas no viable cells remained (5 log reduction) after 2 hr incubation with compound 5. Thus, while these compounds specifically targeted m-Ddh, they possess only moderate bactericidal activity, suggesting that analogous structures and/or optimization of the sulfonamide scaffold might be required to improve whole cell inhibition.

| FUTURE D IREC TI ON S AND CON CLUDING REMARK S
A variety of approaches have been pursued to develop novel potential therapeutics that are active against and/or target oral bacteria. The structure and function of naturally occurring antimicrobial peptides from various sources have been exploited to develop numerous peptide mimetics that show increased activity, stability and other desirable characteristics relative to the parent peptides. In addition, as our understanding of the molecular mechanisms that contribute to the formation of the dental biofilm has increased, new potential targets for therapeutic intervention have been identified and both peptide and small molecule mimetics have been developed that target these key components. Many of these mimetics have demonstrated potent activity in pre-clinical laboratory testing and in various animal models of oral disease, however, for the most part their activity against the more complex oral microbial communities that occur in vivo have not been thoroughly investigated. That said, the current hypothesis that oral disease arises from the disruption of normal host-microbe homoeostasis suggests that targeting specific organisms that initiate or promote the formation of dysbiotic communities may represent a viable and effective therapeutic strategy. In addition, peptide mimetics or small molecule agents may be useful administered after a professional prophylaxis to direct the re-establishment of a microbial community that is associated with health. Future studies to further develop these active agents will likely begin to focus on developing suitable formulations for oral delivery, additional toxicity testing and evaluating stability and pharmacokinetics in the oral environment. The C16G2 STAMP has already been formulated in an oral gel, in a dental varnish and in oral strips. The initial clinical evaluation of these formulations indicated that a single varnish application outperformed multiple gel applications and resulted in a significant reduction in S. mutans populations.
Further Phase II clinical testing is currently ongoing. It is likely that additional peptide and/or small molecule mimetics will undergo clinical evaluation in the future.

ACK N OWLED G EM ENTS
Parts of this work were supported by grants R01DE014605 and R01DE023206 from the National Institute of Dental and Craniofacial Research (to DRD). The authors declare no conflict of interest.