Here we present an earlier version of the article. This version contains a vast of additional data that had to be ommitted from the final version du to journal restrictions. However this version of the article contain additional information that might be of value.
When food is entering our alimentary canal through the mouth, it first meets saliva excreted by the three mayor salivary glands (Parotid, submandibular and sublingual). Acid containing beverages and food are a menace to teeth as these agents contribute to erosion of tooth surfaces (1,2). Enamel and dentine are composed primarily of a carbonate substituted calcium deficient hydroxyapatite Ca10(PO4)6(OH)2. When hydroxyapatite is in contact with water (saliva), the following reaction takes place (3):
Addition of acid to this solution e.g. while drinking apple juice (pH 3), leads to a shift of the chemical equilibrium to the dissolution side as hydroxyl ions (OH–)are removed from the tooth. Dissolution ends and remineralisation of dental hard tissue occurs when the pH in close proximity of the tooth is rising (2). Saliva that permanently covers the structures forming the oral cavity contains three buffer systems, the carbonate, the phosphate and the protein buffering system(4,5). Together these buffer systems form the first line of defence against acidic or basic challenges, a salivary function of utmost importance.
Inorganic buffers are aqueous solution of acids and their conjugated bases that are resistant to pH changes upon the addition of small amounts of acid or base (6). A buffer, as defined by Van Slyke is "a substance which by its presence in solution increases the amount of acid or alkali that must be added to cause unit change in pH” (7). This attribute of the buffering solution is quantified by the buffering value b which is calculated as the quotient of the differential addition of acid over differential pH change b=-dC/dpH (7).
The primary salivary buffer is composed of carbonic-acid (H2CO3) and hydrogencarbonate also known as bicarbonate (HCO3-). Bicarbonate is excreted by the submaxillary duct system by means of an active transport mechanism (8). A bicarbonate containing solution, like saliva exhibits optimal buffering when its pH is equal to the negative logarithm of the acidic constant (pka) of carbonic acid, which is 6.1 (37°C, (9)). According to the Henderson-Hasselbach equation (10), the solution has a maximal buffer range from 5.1 to 7.1 (pka +/- 1 pH unit, (6)) Beyond this range, no buffering from the carbonate system occurs. The reactions of the carbonate system with acids and bases are (4):
Displacement of carbon dioxide from hydrogencarbonate
The concentration of hydrogencarbonate ranges from 5 mM in resting saliva (4) up to 60 mM in stimulated saliva (8). Upon acidic or basic challenges, the carbonate system forms carbonic acid, which rapidly decays to form water and gaseous carbon dioxide. This feature of the carbonate system is called phase-buffering (4). Buffering will occur as long as no more than 50% of the hydrogencarbonate is transformed into carbonic acid. The amount of acid that is buffered is called the ‘buffer power B’ or ‘molecular buffer bM’ with unit mol/l (7). The buffer power is calculated according to the formula B=c2/2c=0.5c, were c is the concentration in mol/l of the buffer components (7).
The secondary salivary buffer is based on the inorganic ions di-hydrogenphosphate (H2PO4-) and mono-hydrogenphosphate (HPO42-). Since the pH of saliva varies between 5.8 and 8 (11), the second dissociation constant of phosphoric acid pka=7.1 defines the buffering optimum of the phosphate system at pH 7.1 (4,8) and a buffer range from pH 6.1 to 8.1 (Fig. 1). The concentration of di-hydrogenphosphate in resting saliva is 7.8 mM and less than 1 mM in stimulated saliva (8). The concentration of mono-hydrogenphosphate varies from a level below 1 mM in resting saliva to 3 mM in stimulated saliva (8). The total phosphate concentration in resting saliva is 8.4 (+/-) 3 mM (12). The reactions of the phosphate system with acids and bases are (4):
Reaction 9:HPO42-+ H3O+ßà H2PO4- + H2O (addition of acid)
Reaction 10: H2PO4- + OH-ßà HPO42-+ H2O (addition of base)
The buffering levels of the carbonate system and the phosphate system have an overlapping zone from pH 6.1 to 7.1 where both systems synchronously are active. Further information and an in-depth discussion of the carbonate and phosphate buffer systems were published by Bardow and co-workers (4).
As high concentrations of inorganic salts such as hydrogencarbonate and di-hydrogenphosphate might interfere with biological reactions, the human body has found another way to provide additional buffering. Similar to the zwitterionic Good-buffers (13,14), proteins have anionic and cationic sites present as non-interacting carboxylate and ammonium side chains. These types of buffers display good water solubility and have a low interference with biological processes. The protein buffer system which is part of the human salivary proteome, currently, is known to comprise 944 protein species (15-22). There is insufficient evidence to known how many of the 944 components of the salivary proteome substantially contribute to buffering. As proteins have many acidic and basic side chains no discrete pka value exists to estimate their buffer level. The amino acids in a protein that can be ionized are referred to as titratable groups. They are divided into acidic and alcaline residues. Aspartic acid (Asp), Glutamic acid (Glu), Cysteine (Cys), Serine (Ser), Tyrosine (Tyr), Tryptophane (Thr) belong to the acidic residues as well as the C-terminal end of the protein. Alcaline residues are the Histidine (His), Lysine (Lys) and Arginine (Arg) and the N-terminal end.
For the acidic groups glutamic- and aspartic acid (Asp pka = 4,4 and Glu pka =3.9), the following reaction takes place at a pH above pH 4.4.
R stands for the rest of the Amino acid (H2N-CH-COOH).
When the pH is higher than the pka the equilibrium of reaction 9 and 10 is on the left side. When acid is added and the pH reaches a value below the pka the equilibrium of both reactions is shifted to the right side. For the basic groups lysine (pka = 10.5) and arginine (pka = 13.2), the following reactions take place:
At a pH lower than the pka the equilibrium of reactions 13 and 14 is on the right side. As soon as the pH is higher than the pka the equilibrium of reactions 11 and 12 is shifted to the left side.
Being part of protein titrable groups are in a complex environment produced by the three dimensional structure of the protein. This can profoundly affect the pka and the reaction of each acidic or basic group. As the pKa of a titrable group is determined by its micro-environment, it can take on a range of values radically different from those measured for individual amino acids. The estimation of individual pka values of a titrable group in a protein is done for instance by the Poisson-Boltzmann method (23). Based on such data titration curves can be calculated to estimate isoelectric points of proteins (24).
The isoelectric point which is the pH where the sum of all positive and negative charges equals zero, is used to define the pH of optimal buffering. Of 944 proteins, 346 salivary proteins have an isoelectric point beyond the buffer range of the carbonate and phosphate system (74 below pH 5.1 and 272 above pH 8.1). Therefore buffering based on protein can be awaited above pH 8.1 and below pH 5.1.
Carbonic anhydrases catalyze the reversible hydration of carbon dioxide (5). They therefore deserve special consideration here. Carbonic anhydrases are located in the human alimentary tract (25,26) and participate in the maintenance of pH homeostasis, in biological fluids of the human body. Of 11 isoenzymes, carbonic anhydrase II and IV are involved in oral physiology (27). Carbonic anhydrase IV is located in the enamel pellicle and not involved in the regulation of actual salivary pH or buffer capacity (5). Carbonic anhydrase II is thought to produce bicarbonate in saliva but until recently, it had not been determined if, and how much, bicarbonate was produced (5).
With the exception of the total protein concentration from 1.5 to 6.5 g/l (28) and the presence of 944 different protein species in saliva, the information about the protein buffer system is scarce (29-33). Lilienthal reported that proteins derived from human saliva do not possess a major buffer effect (29). This finding was confirmed by Bardow(4).
The aim of these experiments was to quantify the buffer features of: single proteins; a two component protein buffer system as well as a saliva model system including two proteins. Our Hypothesis was that the buffering effect of proteins could be measurable in a model system if the adequate analytical parameters are chosen. To quantify the buffer features of the examined systems we used the buffer power B and the buffer value b. Bufferingfrom protein in saliva is likely to occur as in the human body proteins are the most potent buffering substances (34).
MATERIALS & METHODS
Ten ml of the titrand was placed in a vessel in a water bath and stirred at 37°C. First 5 ml of NaOH 0.01 mol/l were added in steps of 200 ml to enclose the buffer range of di-hydrogenphosphate (pH 6.1-8.1), then 25 ml of HCl 0.01 mol/l were added in steps of 200 ml. The pH was recorded after each addition step (150 pH measurements per titration). The buffer power B [mol/l] was calculated with Excel 2003 according to the formula B=c2/2c=0.5c, where c is the concentration in mol/l of the buffer component(s) (7). Data points were exported and fitted with Sigmaplot V9.0. Buffer values b in [mol/(l*pH)=Slyke] (35) were calculated as b=-DC/DpH (7). To calculate the inflections of the titration curves, the concentration and pH were normalized. Inflections were read out of a plotted of DpH/DC (first derivative) against the pH.
The saliva was collected using a widely accepted procedure (36) under resting conditions, between 9:00 am and 10:00 am from volunteers who refrained from eating, drinking, smoking and performing oral hygiene for 2 hours before the collection. Prior to saliva collection all subjects were given explanation of the procedures involved. Samples of 10 ml unstimulated saliva were collected over 10 minutes by frequent releases into a small vessel and the buffering capacity of their saliva sample was determined using the CRT®buffer (Ivoclar Vivadent, Schaan Lichtenstein) was done according to the manufacturer’s protocol.
Averaging saliva titration curves
Unstimulated saliva samples of 5 male subjects aged 35 to 45 with buffer capacities ranging from low to high according the CRT®buffer test (Ivoclar Vivadent, Schaan Lichtenstein), were titrated as described above. The equally spaced pH values (due to monotone titration), were summed up and divided by 5 (averaged), and the standard deviation was calculated.
Amino acid sequence similarity search and protein parameters
Sequence similarity search was done with the basic local alignment search tool (BLAST) against the UniProt Database at www.expasy.org (37). The query sequence was aligned against every subject sequence in the database. The results were reported in form of a ranked list of individual sequence alignments with the quantity of observed similarity expressed in percent of sequence identity. Protein parameters (Mr, pI, number of AS) were calculated with the ProtParam analysis tool (38) at www.expasy.org.
Solution a: 10 mM NaHCO3 Merck 30.06.10, M=84.01 g/mol. Solution b: 5 mM KH2PO4 z.A. Merck 9662790, M=136.09 g/mol. Solution c was composed of: 10 mM NaHCO3, 5 mM KH2PO4. After adjustment of the pH to 7 the solutions were stored in a gas-proof closed vessel.
Lysozyme from hen egg white: Swissprot P00698, Fluka BioChemika 62970, 147 AS, Mr=14.6 kD, pI=9.4, was used as model for human salivary Lysozyme, Swissprot P61626, 148 AS, Mr=16.5 kD, pI=9.4, c=0.1 g/l.
Amyloglucosidase from Aspergillus niger, Swissprot P69328, Fluka BioChemika 10115640 AS, Mr=98 kD, pI=4.25, was used as model for a-amylase, Swissprot P04745, 511 AS, Mr=57.8 kD, pI=6.4, c=0.38 g/l.
The protein formulation was composed of solution c plus 0.5% (340 mM) lysozyme and 0.1 % (10 mM), amyloglucosidase. After adjustment of the pH to 7 the solutions were stored in a closed gas-proof vessel.
Modelling the primary and secondary buffer systems
Results of titrations from di-hydrogenphosphate and hydrogencarbonate solutions looked similar to a polynomial function curve of the third degree (Fig. 1, Table 1). The buffer power B was quantified at the inflections (Ia to Id) where all buffer molecules were neutralized and compared to the calculated values for B.
For the phosphate system we calculated a buffer power of 25 mmol acid and base. The experimentally assessed di-hydrogenphosphate buffer power was 30 mmol acid (hydrogen ions, H+) and 24 mmol base (hydroxyl ions, OH-). Buffering was optimal at pH 6.7 with 0.004 mol/l*pH (Fig. 1,a). Monohydrogenphosphate was not included in this solution as this molecule is present at a level below 1 mM in resting saliva.
For the carbonate system we calculated a buffer power of 50 mmol acid and base. The system buffered 74 mmol (H+) and 8 mmol (OH-). Optimal buffering was found at pH 6.2 with 0.005 mol/l*pH (Fig. 1,b).
The combined phosphate and carbonate systems were theoretically to buffer 75 mmol acid and base. In this experiment, it was found that they buffered 110 mmol (H+), and 22 mmol (OH-) the same amount of base as the phosphate system alone (Fig. 1,c). The distance between inflections Ia and Ib was 44 (H+), between Ib and Ic 36 mmol (H+) respectively which was more than the calculated distances of 25 mmol (H+). Best buffering was detected at pH 6.7 with 0.008 mol/l*pH.
Search for salivary a-amylase and lysozyme substitutes
As purified genuine or recombinant expressed salivary proteins were not available in the desired purity and quantity substitutes were searched. BLAST searches unveiled lysozyme from hen egg which has a 57% sequence homology to human salivary lysozyme and amyloglucosidase from Aspergillus niger with a 35% similarity to human salivary a-amylase as ideal substituents. Both proteins are water soluble and available in high purity in the gram range at reasonable costs.
Modelling a protein buffer system
0.1% amyloglucosidase disolved in water had a buffer power of 68 mmol acid and 0 mmol base. The buffer range spanned from pH 3.3 to 5.3 with optimal buffering at pH 4.35 and a buffer value of 0.004 mol/l*pH (Fig. 2, b). For 0.5% lysozyme none of these data were measurable (Fig. 2,a) although the protein has 32 titrable groups (24). If both enzymes were dissolved together a buffer power of 84 mmol acid and 0 mmol base was detected. The buffer range was effective from pH 3.3 to 5.3 with optimal buffering at pH 4.5 and a buffer value of 0.007 mol/l*pH. Titration of 0.1% amyloglucosidase and 0.5% lysozyme in a solution of 5 mM di-hydrogenphosphate and 10 mM hydrogencarbonate gave an almost linear titration curve (Fig. 2,d). This protein formulation buffered 158 mmol of acid and 38 mmol of base. After subtraction of hydrogencarbonate and di-hydrogenphosphate buffer power 48 mmol acid and 14 mmol base buffer power remained to the enzymes. The b values measured at pH 4.4 was 0.005 mol/l*pH and 0.01 mol/l*pH at pH 6.5 (Table 1).
Total salivary buffering
Titration of human saliva gave an almost linear curve (Fig. 3A,b). The total acidic buffer power was 168 mmol (H+) and the total alcaline buffer power 42 mmol (OH-). Human saliva had a buffer zone from pH 4 to 8 instead of discrete buffer optima with b values from 0.005 mol/l*pH (pH 4.3) to 0.01 mol/l*pH (pH 6.5) (Table 1).
Comparison of saliva vs. model buffer systems
Titration of human saliva gave a curve (Fig. 3A,a) different to the ones of hydrogencarbonate (Fig.1,b) and di-hydrogenphosphate (Fig.1,a) or a combination of both (Fig.1,c). However the protein supplemented hydrogencarbonate and di-hydrogenphosphate solution showed behaviour comparable to full saliva (Fig. 3A,b). The total acid buffer power was 58 mmol (H+) larger than the combined power of the di-hydrogenphosphate and hydrogencarbonate systems and only 10 mmol (H+) and 0.1 pH units different to the protein formulation. The total basic buffer power was 20 mmol (OH-) higher than hydrogencarbonate and di-hydrogenphosphate together and 4 mmol larger than the protein formulation.
Averaged saliva titrations vs. model buffer systems
The titration data obtained for the protein formulation and the control group were superimposed. Figure 3B shows that the titration curve of the protein formulation runs within one standard deviation of the control group. The curves were almost congruent from 0 to 10 mmol addition of base and from 28 to 50 mmol addition of acid. The buffer power only differed by 4 mmol.
As the protein buffering is inaccessible even by state-of-the-art protein analytics, a new combination of synthetic and analytical approach was developed. This method involved the stepwise formulation of a buffer solution that acts as model buffer system. The procedure allowed quantification of the buffer attributes at each stage of the gradually increasing formulation and a direct comparison to experimental data obtained from human saliva.
The experimentally determined buffer power of 30 mmol acid and 24 mmol base of the phosphate system corresponded to the theoretical predictions of 25 mmol acid and 25 mmol base (Table 1). This was not the case for the carbonate system. The acid buffer power was 48% higher than expected. The fact that an open system, similar to the mouth to was used to assess a phase-buffer process driven by the loss of gaseous carbon dioxide was the reasoning behind for this observation. A second manifestation of the same phenomenon is the broader-than-calculated spacing between the inflections of di-hydrogenphosphate and hydrogencarbonate, as well as, between hydrogencarbonate and the hydrogencarbonate/di-hydrogenphosphate solution (Fig. 1). Above pH 7, the carbonate system rapidly lost its buffering power, which became evident when only 16% of the theoretically possible maximum was detected.
The synchronous action of di-hydrogenphosphate and hydrogencarbonate was calculated to produce a consolidated buffer power. This was what was observed in the acidic pH range. 30 mmol of di-hydrogenphosphate buffer power added to 74 mmol of hydrogencarbonate buffer power roughly equalled 110 mmol of combined buffer power. When the loss of carbon dioxide is neglected, 75 mmol would be expected. If the loss of CO2 was taken in account, the increase from 75 to 110 mmol (H+) by 46% was consistent with our observation for hydrogencarbonate alone (48%). The base-buffering power differed only by 9% to the one of di-hydrogenphosphate alone, which supports the theory that the carbonate system is almost inactive above pH 7.
This investigation showed that the protein formulation had a buffer range from pH 3.3 to 7.5 with optima at pH 4.3 (0.005 mol/l*pH) and pH 6.5 (0.01 mol/l*pH). The protein used in this formulation (0.6%) did not exceed the total protein concentration found in human saliva (28). For human saliva, identical values at pH 4.3 (0.005 mol/l*pH) and pH 6.5 (0.01 mol/l*pH) were detected. Strikingly, in saliva, a zone of increasing buffer values beginning with 0.004 mol/l*pH at pH 3.3 reaching the top of 0.01 mol/l*pH at pH 6.5 was measured. This remained at this level up to pH 8, whereas, in the protein-supplemented hydrogencarbonate and di-hydrogenphosphate solution, a discrete optimum at pH 4.3 (0.005 mol/l*pH) was measured.
The optimum of 0.01 mol/l*pH at pH 6.5 exactly corresponded to the one measured for 5 mM di-hydrogenphosphate and 10 mM hydrogencarbonate with the difference that the buffer value in the hydrogencarbonate and di-hydrogenphosphate solution was only 0.008 mol/l*pH. Hence it can be concluded that the buffer optimum at pH 6.5 in the protein supplemented hydrogencarbonate and di-hydrogenphosphate solution is created 80% by di-hydrogenphosphate and hydrogencarbonate and 20% by amyloglucosidase and lysozyme. As the same buffer value is found in saliva, we have evidence that also here proteins contribute a similar fraction of the buffer value at pH 6.5. This contradicts the hypothesis of Sellmann that proteins buffer only at low pH values (39) and supports the one of Freidin (31) who proposed a protein buffer zone from pH 5.5 to 7.8.
These experimentally-measured buffer values for saliva are in agreement with those published by Bardow (4). Interestingly, there was no difference between those buffer values measured in a closed system (4) or in an open system as in this study although an open system continuously loses carbon dioxide during titration.
The second buffer optimum of the protein formulation was located at pH 4.3 with a buffer value of 0.005 mol/l*pH corresponding to the amyloglucosidase buffer optimum in water. Surprisingly lysozyme, which had itself no measurable buffer attributes at all, influenced the buffer value and buffer power of amyloglucosidase. A 20% increase in the acidic buffer power and a 75% increase in the buffer value were detected, with a shifted buffer optimum from 4.3 to pH 4.5. The mechanisms behind this synergistic effect are not known, and need to be further investigated. In contrast to the hypothesis of Bardow (4) and Lilienthal (29) these results suggest that amylase possesses an important buffer effect that is amplified through interaction with lysozyme. Maybe, the use of hydrochloric acid in a concentration of 1 mol/l in relation to the saliva proteins which are present in millimolar concentrations, explains why Bardow could not detect a major buffer effect (4).
If amyloglucosidase and lysozyme were dissolved in hydrogencarbonate and di-hydrogenphosphate solution the resulting total acidic buffer power was increased by 43%, the buffer value by 25%. Therefore these results support the hypothesis of Bardow (4) that interaction of the protein system with hydrogencarbonate or di-hydrogenphosphate increases the buffer power of the protein system. Together amyloglucosidase and lysozyme contributed 31% of the total acidic buffer power, leaving 19% to di-hydrogenphosphate and 50% to hydrogencarbonate. Above pH 7 amyloglucosidase and lysozyme constituted 35% of the total basic buffer power. These values are in the same range as for human saliva, where 35% of acid and base buffer power for the protein system was measured.
One has to keep in mind that a discrete buffer optimum like the one found in the protein formulation at pH 4.5 is not adequate to explain the existence of the detected buffer zone in saliva. In this context it is worth mentioning the recently discovered human salivary a-amylase subproteom. This proteome consists of 67 amylase subspecies with isoelectric points ranging from pH 3.5 to 7.6 (40). Most of the subspecies are truncated a-amylase variants that have no catalytic function. The reason why the human body produces such a wide variety of non functional enzymes is still a mystery. Maybe the answer to this mystery is the provision of a protein buffer system, covering four pH units.
1. Jaeggi T, Zero D. The role of diet in the aetiology of dental erosion. Caries Res 2004; 38 Suppl 1: 34-44.
2.Zero DT, of enamel erosion – Intrinsic and extrinsic factors. In: Tooth wear and sensitivity: Martin Dunitz Ltd, 2000:121-139.
3.Dawes C. What is the critical pH and why does a tooth dissolve in acid? J Can Dent Assoc 2003; 69: 722-724.
4.Bardow A, Moe D, Nyvad B, Nauntofte B. The buffer capacity and buffer systems of human whole saliva measured without loss of CO2. Arch Oral Biol 2000; 45: 1-12.
5.Lenander-Lumikari M, Loimaranta V. Saliva and dental caries. Adv Dent Res 2000; 14: 40-47.
6.Mortimer CE. Säuren und Basen. In, Das Basiswissen der Chemie. 5 ed. Stuttgard: Georg ThiemeVerlag; 1987:283
7.Van Slyke D. On the measurement of buffer values and the relationship of buffer value to the dissociation constant of the buffer and the concentration and reaction of the buffer solution. J Biol Chem 1922; 52: 525-570.
8.Kreusser W, Hennemann H, Heidland A. [Saliva electrolytes and flow rate in the diagnosis of Bartter's syndrome, pseudo-Bartter's syndrome and Conn's syndrome]. Verh Dtsch Ges Inn Med 1972; 78: 1518-1522.
9.Sigaard-Anderson O. In, The acid-base status of the blood. Copenhagen: Munksgaard; 1963
10.Hasselbach KA. Die Berechnung der Wasserstoffzahl des Blutes aus der freien und gebundenen Kohlensäure desselben, und die Sauerstoffbindung des Blutes als Funktion der Wasserstoffzahl. Biochem Z 1917; lxxvi: 112-143.
11.Larsen MJ, Jensen AF, Madsen DM, Pearce EI. Individual variations of pH, buffer capacity, and concentrations of calcium and phosphate in unstimulated whole saliva. Arch Oral Biol 1999; 44: 111-117.
12.Rehak NN, Cecco SA, Csako G. Biochemical composition and electrolyte balance of "unstimulated" whole human saliva. Clin Chem Lab Med 2000; 38: 335-343.
13.Good NE, Winget GD, Winter W, et al. Hydrogen ion buffers for biological research. Biochemistry 1966; 5: 467-477.
14.Good NE, Izawa S. Hydrogen ion buffers. Methods Enzymol 1972; 24: 53-68.
15.Kojima T, Andersen E, Sanchez JC, et al. Human gingival crevicular fluid contains MRP8 (S100A8) and MRP14 (S100A9), two calcium-binding proteins of the S100 family. J Dent Res 2000; 79: 740-747.
16.Ghafouri B, Tagesson C, Lindahl M. Mapping of proteins in human saliva using two-dimensional gel electrophoresis and peptide mass fingerprinting. Proteomics 2003; 3: 1003-1015.
17.Yao Y, Berg EA, Costello CE, Troxler RF, Oppenheim FG. Identification of protein components in human acquired enamel pellicle and whole saliva using novel proteomics approaches. J Biol Chem 2003; 278: 5300-5308.
18.Huang CM. Comparative proteomic analysis of human whole saliva. Arch Oral Biol 2004; 49: 951-962.
19.Vitorino R, Lobo MJ, Ferrer-Correira AJ, et al. Identification of human whole saliva protein components using proteomics. Proteomics 2004; 4: 1109-1115.
20.Wilmarth PA, Riviere MA, Rustvold DL, et al. Two-dimensional liquid chromatography study of the human whole saliva proteome. J Proteome Res 2004; 3: 1017-1023.
21.Hu S, Xie Y, Ramachandran P, et al. Large-scale identification of proteins in human salivary proteome by liquid chromatography/mass spectrometry and two-dimensional gel electrophoresis-mass spectrometry. Proteomics 2005; 5: 1714-1728.
22.Xie H, Rhodus NL, Griffin RJ, Carlis JV, Griffin TJ. A catalogue of human saliva proteins identified by free flow electrophoresis-based peptide separation and tandem mass spectrometry. Mol Cell Proteomics 2005; 4: 1826-1830.
23.Nielsen JE, Vriend G. Optimizing the hydrogen-bond network in Poisson-Boltzmann equation-based pK(a) calculations. Proteins 2001; 43: 403-412.
24.Kuramitsu S, Hamaguchi K. Analysis of the acid-base titration curve of hen lysozyme. J Biochem (Tokyo) 1980; 87: 1215-1219.
25.Parkkila S, Kaunisto K, Rajaniemi L, et al. Immunohistochemical localization of carbonic anhydrase isoenzymes VI, II, and I in human parotid and submandibular glands. J Histochem Cytochem 1990; 38: 941-947.
26.Parkkila S, Parkkila AK, Juvonen T, Rajaniemi H. Distribution of the carbonic anhydrase isoenzymes I, II, and VI in the human alimentary tract. Gut 1994; 35: 646-650.
27.Kadoya Y, Kuwahara H, Shimazaki M, Ogawa Y, Yagi T. Isolation of a novel carbonic anhydrase from human saliva and immunohistochemical demonstration of its related isozymes in salivary gland. Osaka City Med J 1987; 33: 99-109.
28.Alfonsky D. Saliva and its relation to oral health. Drawer: University of Alabama press, 1961
29.Lilienthal B. An analysis of the buffer systems in saliva. J Dent Res 1955; 34: 516-530.
30.Dawes C. The effects of exercise on protein and electrolyte secretion in parotid saliva. J Physiol 1981; 320: 139-148.
31.Freidin LI, Nikolaev AA, Freidin BL. [Protein buffer in human mixed saliva]. Stomatologiia (Mosk) 1985; 64: 16-17.
32.Van Nieuw Amerongen A, Bolscher JG, Veerman EC. Salivary proteins: protective and diagnostic value in cariology? Caries Res 2004; 38: 247-253.
33.Pedersen AM, Bardow A, Nauntofte B. Salivary changes and dental caries as potential oral markers of autoimmune salivary gland dysfunction in primary Sjogren's syndrome. BMC Clin Pathol 2005; 5: 4.
34.Stoelting KR. Pharmacology & Physiology in Anesthetic Practice. Third Edition ed, 1999:702.
35.Izutsu KT. Theory and Measurement of the buffer value of bicarbonate in saliva. J Theor Biol 1981; 90: 397-403.
36.Nagler RM, Hershkovich O. Relationships between age, drugs, oral sensorial complaints and salivary profile. Arch Oral Biol 2005; 50: 7-16.
37.Altschul SF, Madden TL, Schaffer AA, et al. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res 1997; 25: 3389-3402.
38.Gasteiger E, Hoogland C, Gattiker A, et al. Identification and Analysis Tools on the ExPASy Server. In: Walker JM ed, The Proteomics Protocols Handbook: Humana Press; 2005:571-607
39.Sellman S. The Buffer Value of Saliva and its Relation to Dental Caries. Acta Odotol Sacand 1949; 8: 244-268.
40.Hirtz C, Chevalier F, Centeno D, et al. MS characterization of multiple forms of alpha-amylase in human saliva. Proteomics 2005; 5: 4597-4607.