Is the proteomic composition of the salivary pellicle dependent on the substrate material?

The use of dental restorative materials is a routine task in clinical dentistry. Upon exposure to the oral cavity, continuous adsorption of salivary proteins and other macromolecules to all surfaces occurs, representing the first step in dental biofilm formation. Different physico‐chemical properties of substrate materials potentially influence the composition of the initial biofilm, termed pellicle. This study aimed at characterizing and comparing the individual proteomic composition of the 3‐min pellicle formed on bovine enamel and six restorative materials.

Under in vivo conditions, all orally exposed surfaces are covered by the acquired salivary pellicle, which represents the base for the development of the oral biofilm [1][2][3]. The continuous formation of the acquired salivary pellicle, termed simply pellicle for the sake of convenience, starts within seconds after oral hygiene and is due to the adsorption of proteins and other biomolecules from the surrounding oral fluids [4,5]. It predominantly consists of salivary proteins and fulfills a protective role by acting as a semipermeable membrane protecting the tooth surface against demineralization processes and mechanical damages [1,[6][7][8].
The proteinaceous composition of the pellicle layer reflects a complex interplay between adsorption and desorption of different proteins, their conformational changes, and enzymatic composition [9]. The very initial step in pellicle formation, taking place within the first few minutes after oral hygiene, is due to the adsorption of salivary proteins to all exposed surfaces [4]. As the pellicle matures, additional salivary biomolecules such as lipids, carbohydrates, nucleic acids, and protein aggregates adsorb, leading to the formation of the so-called pellicle matrix, a meshwork of a densely structured basal layer and an outer layer of globular appearance [10,11]. The continuously growing outer layer provides more and more binding sites for bacterial adhesions, resulting in the formation of a bacterial biofilm. It has been concluded from previous in vitro and in vivo studies that pellicle formation is a highly selective process [4,12,13], whereby the physico-chemical properties of the surface were supposed to determine the adsorption of salivary proteins to it [2].
Many studies investigated the influences of different substrate materials on the bacterial biofilm formation and the antibiofilm properties of diverse dental materials [14][15][16][17]. However, an important prerequisite for bacterial adhesion and biofilm formation is the presence of the acquired salivary pellicle, which forms immediately after oral hygiene [4].
The work of Fischer and Aparicio summarizes the rather small body of studies investigating the formation and characteristics of the salivary pellicle on dental materials [18].
Although protein adsorption is the first biological response at the interface between a substrate surface and a biological fluid, a systematical elucidation of the effect of different substrate materials on the precise proteomic composition of the in situ acquired shortterm pellicle is still missing [2]. Supplemental Table S1 summarizes the available literature related to this topic, including the respective limitations of the single studies due to the study design, methodology, their in vitro character, or the restricted extent. Among them, there are studies investigating the ultrastructure [3], the amino acid composition [19], or the physico-chemical properties of salivary pellicles formed on different substrate materials [20]. Other works focused on the enzymatic activity or the adsorption of single salivary proteins [21][22][23][24], whereat these results should not be transferred unrestrictedly to the adsorption kinetics of the full salivary proteome. More recent studies investigated the salivary pellicle formed on different substrate materials using sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) [25] or two-dimensional gel electrophoresis   [26,27]. Based on these results, the authors proposed differences in the proteomic composition of the salivary pellicle on the different

Clinical Relevance
The characterization and comparison of the initial pellicle proteome formed on seven different dental materials showed unexpected high similarities among all substrates.
These results suggest a minor relevance of the respective substrate material properties on the proteomic composition of the individual 3-min pellicle. Most likely this can be ascribed to a prompt shielding of the physico-chemical substrate properties upon salivary exposure. The proteomic data of the present study impart the first detailed insight in the process of pellicle formation on different dental materials under oral conditions. Far-sighted, the current results can contribute to a deeper understanding of the process of protein adsorption at any interface between a complex biological fluid and solid surfaces. substrate materials. However, those studies were conducted in vitro and simply resulted in protein patterns on the SDS-PAGE/2-DE-gels or the detection of single proteins by immunoblotting. In several recent studies, liquid chromatography-tandem mass spectrometry (LC-MS/MS) was used as up to date-method to elucidate the pellicle proteome on dental materials in detail [26,[28][29][30][31]]. Yet likewise, these studies were conducted in vitro and focused on the differences in the proteomic pellicle profiles formed on variably pretreated titanium specimens or differences between titanium and feldspar ceramics. To the best of our knowledge, the only studies elucidating the detailed proteomic composition of the salivary pellicle formed on a dental material in situ with a contemporary method are Delius et al. 2017 and Trautmann et al. 2019 [12,32]. Therein, the individual 3-min in situ-pellicle formed on ceramics was characterized using nano-liquid chromatography-high resolution-tandem mass spectrometry (nano-LC-HR-MS/MS). Till this day, a comprehensive analysis and comparison of the proteomic composition of the in situ pellicle formed on different dental materials with an up to date methodology enabling reliable protein identifications of its entire proteome is missing.
As the initial pellicle represents the base for all subsequent adsorption processes and bacterial adherence, its proteomic composition is likely to influence those subsequent steps. Therefore, the aim of the current study was to comprehensively elucidate the individual proteomic composition of the 3-min in situ pellicle formed on seven different dental materials. In a second step, substrate material-specific adsorption patterns were analyzed by comparing the proteomic profiles of the 3-min pellicle to the profile of the corresponding saliva.

MATERIALS AND METHODS
The current study follows up on previous work described in Trautmann et al. 2020 [13] and uses essentially the same methodology. For completeness, it is summarized here again. All experiments started at 9.00 am to avoid circadian effects on the salivary composition. Subjects refrained from any food or beverages 2.5 h before starting and conducted individual oral hygiene with the respective standard hygiene products to guarantee individual day-today conditions. To avoid influences caused by any ingredients of the hygiene products, subjects conducted oral hygiene by the use of dental silk and tooth brushing without the use of any tooth paste 30 min before saliva collection and intraoral exposure. Over a fixed time period of maximum 20 min on ice, 10 mL of unstimulated saliva were collected.

Pellicle collection
The enamel slabs were derived from the labial surfaces of bovine  and air-dried after removal from the oral cavity. Single slabs were incubated successively in Triton X-100 (1% v/v Triton X-100 in TRIS-HCl-buffer (0.02 M TRIS, 0.15 M NaCl, pH 7.5)) on ice followed by ultrasonication in RIPA-buffer (Cell Signaling Technology) at 4 • C to elute the adsorbed pellicle components. Both solutions contained EDTA-free protease inhibitor mix (Complete, Roche). Probes were frequently vortexed to increase protein relief during the elution-steps.

Precipitation of pellicle proteins
The pools of eluted proteins from the different dental substrate materials and enamel specimens of each subject were precipitated with trichloroacetic acid, washed twice with acetone, and air dried. The protein pellets were denatured for subsequent gel electrophoresis.

Saliva collection
Over a fixed time period of maximum 20 min, 10 mL of unstimulated saliva was collected on ice, sterile-filtered (5 µm filter unit, Whatman), and centrifuged. Salivary flow rates of the subjects were in the norm of healthy individuals (0.5 mL/min for unstimulated saliva). Ninety micrograms of each sample were denatured for subsequent gel electrophoresis.

Nano-LC-HR-MS/MS
Six mocroliters of the digested peptides were automatically transferred to a nanoflow liquid chromatography system (Ultimate

Quantitative analysis
The exponentially modified protein abundance index (emPAI)-values of each protein were used to calculate the protein content, that is, the percentage of molar amount of substance, using the formula in order to obtain a value comparable between the samples [39].
The fold change (fc; here: measure of degree of change in protein content from the saliva to the pellicle) was considered to be enriched when fc > 2 and depleted when fc < 0.5 as described in Delius et al.
2017 [12]. Distribution patterns of molecular weight (MW) and isoelectric point were compared using the one-sided Mann-Whitney U rank test. Isoelectric points were taken from the isoelectric point database [40].

Individual proteomic composition of the saliva and 3-min pellicle formed on different dental materials
In this study, the initial pellicle formed on different dental restorative materials, namely ceramics, gold, titan, PMMA, composite, and polytetrafluoroethylene (PTFE/Teflon) as most hydrophobic control, was characterized and compared to the one formed on bovine enamel.
Within the initial phase of pellicle formation, taking place within a matter of seconds up to a few minutes after oral hygiene, mainly salivary proteins adsorb to the substrate surface [4,47]. These adsorptions represent the very first interactions between the exposed surfaces and the components of the surrounding fluids. If there are any compositional differences dependent on the substrate material, these differences are most likely present in the initial pellicle. Up to date, there is no extensive information about the proteomic composition of the short-term pellicle based on the lack of practicable techniques for the analysis of this very thin layer. As the composition of this initial pellicle is likely to influence all subsequent interactions/adsorption processes between the surface and the biomolecules as well as microorganisms present in the oral liquid phase, the current study focused on the identification of the 3-min pellicle proteome. In order to identify potential substrate material-specific differences in this basal layer resulting in composi-  I  II  III  IV  V  Diversity  Overlapping proteins   Saliva  1009  875  855  1032  760  1435  525   3-min pellicle  Enamel  490  244  488  479  296  772  147   Ceramics  262  187  286  491  234  613  82   Composite  400  345  384  517  343  706  181   Gold  251  282  298  391  214  565  95   Titan  327  269  276  479  307  624  pellicle on bovine enamel were taken from Trautmann et al. 2020 [13]). To ensure a high identification quality of the individual samples, a chemical elution followed by an elaborate nano-mass spectrometric protocol already described in Trautmann et al. 2019 [32] was applied.
A virtually complete elution of the 3-min pellicles from the different substrate materials was verified by transmission electron microscopy analyses conducted before and after elution (Supplemental Figure S1).  Table S2).
When forming the overlap of all substrate materials, 105 proteins were identified in the 3-min pellicle of all subjects on all substrate materials, including all previously stated proteins of the pelliclebase proteome [32]. Intersecting these 105 proteins with the saliva gave 60 proteins that were identified in the 3-min pellicle on all substrate materials as well as in the saliva of all subjects (exclusively present proteins on single substrate materials listed in Supplemental Table S3).
The current results represent the so far largest number of 1348 uniquely identified pellicle proteins. This is a significant increase over the 772 proteins identified earlier in a previous analysis that was included in the current study [13]. There, the individual 3-min pellicle of the same five subjects formed on bovine enamel was characterized and compared to saliva in terms of qualitative and quantitative aspects.   [13], where the individual 3-min pellicle formed on bovine enamel was characterized and compared to saliva. Therein, mostly small MW proteins were identified to adsorb to the initial pellicle. This study gave very similar results for all substrate materials with the largest part of the adsorbed proteins possessing MWs between 10 and 20 kDa. When comparing the pellicle proteomes to the salivary proteome, the proteins adsorbed to PTFE, PMMA, and gold had significantly higher MWs. However, the effect size of the differences between the salivary proteome and the proteins adsorbed to PTFE, PMMA, and gold was rather small (rank-biserial correlation = 0.002-0.106). One may speculate that large proteins may be able to undergo larger conformational changes than smaller, more stable ones. In the case of PTFE and PMMA, such large proteins would undergo conformational changes to expose their hydrophobic groups and interact with these hydrophobic surfaces. In the case of gold, thiol bonds could be formed between protein side chains having a thiol group and the surface material. Likewise, the proteins may undergo conformational changes to expose the thiol groups and form the bridges [48]. Those interactions would be stronger than the ones formed by small proteins, and therefore the larger proteins might repel those loosely bound smaller proteins from the respective substrate Subsequently, the isoelectric points (pIs) of the identified proteins were analyzed. The proteins were assorted to three pI-ranges: with a pH below 6.3 (negatively charged in the oral cavity), in the physiological pH of the oral cavity ranging from 6.3 to 7.6 (neutral) and with a pH higher than 7.6 (positively charged in the oral cavity) [50]. The qualitative analysis disclosed a broad coverage of pIs between pH 4 and 11 on all substrate materials and in the saliva. The quantitative evaluation showed the distribution patterns of the salivary proteins to possess a broad spectrum of isoelectric points with mostly rather low probabilities of presence (Figure 2/Supplemental Figure S4 for individual distribution patterns). Therein, the largest part (47%) was located in the pH range below 6.3. Only 17% of the salivary proteins possessed isoelectric points in the pH range above 7.6, exhibiting positive net charges. The pellicle proteins found on the different substrate materials had highly similar distribution patterns of the isoelectric points.

Bioinformatics analyses of the identified proteins
For the most part, these patterns were merely differing in the probabilities of presence of the single isoelectric points. Compared to the saliva with only 17%, a clearly increased fraction of proteins had isoelectric points above 7.6 with 24% on PTFE up to 40% on composite. Overall, these analyses revealed similar distribution patterns of the isoelectric points on all substrate materials, relying most likely on the adsorption of the same salivary proteins. The comparison of the distribution F I G U R E 1 Quantitative molecular weight distribution patterns of proteins identified in the saliva and 3-min pellicle on seven substrate materials, averaged over five subjects, based on exponentially modified protein abundance index (emPAI)-values. Occurrence probabilities are plotted against the molecular weight of the identified proteins. Outer histograms depict the distribution patterns on the whole molecular weight spectrum of all proteins identified. Inner histograms depict a magnification of the low to medium molecular weight range from 0 to 100 kDa. Data of saliva and enamel were taken from Trautmann et al. 2020 [13], copyright permission obtained patterns from saliva and the 3-min pellicles revealed a favored adsorption of positively charged proteins on all substrate materials.
Taken together, the current results regarding the physico-chemical properties of the identified proteins point to a favored adsorption of proteins with low MWs and preferentially positively charged proteins during the initial stage of pellicle formation on all substrate materials.
These results are in line with the findings of a previous study on the 3min pellicle on ceramics. There, the adsorbed proteins that were significantly enriched in the 3-min pellicle had on average higher isoelectric points than the ones in the liquid saliva phase [12].
The current results point to a similar composition of the 3-min pellicle proteome independent of the respective substrate material. Those findings deviate from previous literature that reported clear differences in the pellicle composition between dental enamel and the dental materials titanium and PMMA [27]. Those earlier analyses were based on a 2D gelelectrophoresis of the residual saliva remaining after a 2 hexposure to the different substrate materials, depicting the respective protein spots within the different pI-and kDa-ranges. The authors identified different protein patterns on the 2D gels and suggested an influence of the underlying substrate material on the salivary film F I G U R E 2 Quantitative isoelectric point distribution patterns of proteins identified in the saliva and 3-min pellicle on seven substrate materials, averaged over five subjects. Occurrence probabilities based on exponentially modified protein abundance index (emPAI)-values are plotted against the isoelectric points of the identified proteins. Green areas depict the physiological pH-range of 6.3−7.6 in the oral cavity. Percentages of proteins possessing pIs below, within, or higher than the physiological oral pH-range are shown in boxes. Basic data of saliva and enamel were taken from Trautmann et al. 2020 [13], copyright permission obtained

Enrichment analysis of identified molecular functions
To extract biological meaning from the list of annotated MFs, an enrichment analysis against the full genomic background was performed.
Hereby, an enriched MF reflects a significantly elevated presence of such proteins in the adsorbed proteome. Figure 4

Selective adsorption of distinct salivary proteins
In the next step, the selective adsorption of salivary proteins in the 3min pellicle was analyzed on the basis of the mole fractions (mole%) of the identified proteins for the different subjects and substrate materials relative to the saliva proteome (Supplemental Figure S6). A protein was considered to be enriched on a substrate material for an fc > 2.0 and depleted for an fc < 0.5 [12]. In this analysis, we focused on the proteins that showed the same trend (enriched or depleted) for one of the substrate materials tested on all subjects. We found that 127 proteins were commonly enriched or depleted in the 3-min pellicle of all subjects on at least one up to seven substrate materials. The averaged distribution patterns of these 127 proteins are depicted in Figure 5. F I G U R E 5 Fold changes of proteins being commonly enriched or depleted in all subjects in 3-min pellicle on one up to seven substrate materials. Enriched proteins are shown in red shades, depleted proteins are shown in blue shades, proteins present in 3-min pellicle without significant fold change compared to saliva are shown in light gray. Different substrate materials are plotted against the accession numbers (and gene names) of the proteins Virtually all of the enriched or depleted proteins detected in the 3min pellicle showed similar fcs on at least two or even more substrate materials. Only the two proteins nucleobindin-2 (NUCB2) and Thioredoxin (THIO) were found to be enriched on one substrate material and depleted on another. Thereby, the calcium-binding protein NUCB2 was

CONCLUDING REMARKS
Taken together, the overall interpretation of the current results suggests a rather minor relevance of the respective substrate material properties on the proteomic composition of the individual 3-min pellicle. Furthermore, the data underpins the hypothesis of a pellicle layer which is physiologically functional even after a few minutes of formation time [4,12]. Presumably, the surface of the substrate materials is promptly covered with salivary proteins adsorbing upon exposure to the oral cavity, thereby shielding in part its respective physico-chemical properties and forming a uniform basal pellicle layer. This scenario is in line with the highly meaningful pellicle formation-phenomenon in acting as physiological mediator by instantly covering all orally exposed particles and masking the physico-chemical surface properties of different substrate materials.
The proteomic data of the present study not only contribute to a much more detailed explanation of the process of pellicle formation on dental materials under oral conditions. The current results are also of high relevance for a deeper understanding of the process of protein adsorption at any interface between a complex biological fluid and solid surfaces.

ACKNOWLEDGMENTS
This study was funded by the German Research Foundation (DFG SFB1027, projects B3 and C3). Claudia Fecher-Trost was supported by HOMFOR and by DFG SFB512 to Veit Flockerzi.

CONFLICT OF INTEREST
The authors declare no conflict of interest.