Sodium pyruvate, L-methionine, L-cysteine, L-cystine, α-D-glucose, sodium trimethylsilyl-[2,2,3,3-²H₄] propionate (TSP) and deuterium oxide (²H₂O) were of the highest possible analytical grade and purchased from the Sigma-Aldrich Chemical Co. (Gillingham, Dorset, UK).

Aqueous solutions containing sodium pyruvate, alpha-Dglucose, L-cysteine and L-methionine (5.00 mM) were prepared in 40.0 mM phosphate buffer (pH 7.00) which was rigorously deoxygenated with O₂-free N₂ gas prior to use. 5.00 ml aliquots of these aqueous solutions were divided into two equivalent portions (2.50 ml). The first of these was treated with gaseous O₃ synthesised by the HealOzone Unit (CurOzone, USA) for a 10 s period (equivalent to a delivery of 4.84 mmol of this reactive oxygen radical species) and then equilibrated at a temperature of 35˚C for a 60 min. period prior to high-resolution ¹H NMR analysis. The second matching group of de-oxygenated (untreated) 2.50 ml volume solutions, which were equilibrated in the same manner, served as essential controls.

Each of the above experiments was repeated three times so that there was a total of n = 4 replicate solutions (each with an initial pre-treatment O₃ scavenger concentration of 5.00 mM) for each biomolecule investigated.

One-dimensional (1D) 600 MHz proton NMR spectra of the chemical model systems described above were acquired on a Bruker AMX-600 spectrometer. Typically, 0.60 ml aliquots of untreated or ozonated biomolecule solutions were treated with a 0.10 ml volume of a 1.47 mM solution of TSP in ²H₂O (the latter providing a field-frequency lock), the mixtures thoroughly rotamixed and then transferred to 5-mm diameter NMR tubes. Typical pulsing conditions were: 64 FIDs using 32,768 data points, 72˚ pulses and a 3 s pulse repetition rate to allow full spin-lattice relaxation of the hydrogen nuclei in the samples investigated. Chemical shifts were referenced to TSP (internal; final concentration 0.21 mM), and exponential line-broadening functions of 0.30 Hz were employed for purposes of processing.

The intensities of the most prominent ¹H NMR resonances of each biomolecule and their corresponding reaction products with O₃ were determined by electronic integration via the spectrometer’s proprietary software (XWIN-NMR), and the concentrations of components detectable were determined by comparisons of their resonance areas with that of the added TSP internal standard (final concentration 0.21 mM). Maintenance of the exact integral regions for each spectrum acquired was ensured.

The (soluble) O₃ concentration of 40.00 mM phosphate buffer solutions treated with O₃ in the same manner as each of the above biomolecule solutions was determined by a modification of the spectrophotometric method of Schechter [15] which involves the oxidation of iodide anion by O₃ to form the chromophoric ion (λ_max = 350 nm, e = 2.32 × 10⁴ M⁻¹·cm⁻¹). Absorbance measurements and electronic absorption spectra were recorded on a Unicam UV-2 spectrophotometer.

Results were reported as the mean ± between replicates (n = 4) standard error of the mean (SEM) percentage consumption of each biomolecule examined in this study (all pre-treatment biomolecule concentrations were equivalent, i.e. 5.00 mM).

5.00 mM aqueous solutions of the α-keto acid anion pyruvate were treated with O₃ as described in the Materials and Methods section in order to investigate the redox reaction occurring between these agents, and proton (¹H) NMR analysis of these solutions demonstrated a marked level of oxidative decarboxylation of this alpha-keto acid anion to acetate and CO₂ (Figure 1), an observation consistent with the reaction depicted in equation (1). Also consistent with this observation, a singlet resonance located at 1.50 ppm and ascribable to pyruvate hydrate [the enol form of this α-keto acid anion ()], of much lower intensity than that of the keto form at 2.388 ppm, was also removed from spectra after O₃ treatment. The decreases observed in the intensities of the pyruvate and pyruvate hydrate signals were, of course, accompanied by corresponding equivalent increases in that of the acetate-CH₃ resonance (δ = 1.92 ppm). The mean ± SEM percentage decrease in the intensities of the combined pyruvateand pyruvate hydrate-CH₃ signals was found to be 93% ± 4% (mean ± SEM), although it should be noted that the effective concentration of O₃ in the system employed here is limited by its solubility in water, together with its rate and level of consumption by the scavenger employed, and also its catalytically-promoted dissociation to dioxygen during the 5 min. treatment period.

¹H NMR analysis of reaction products revealed that interaction of O₃ with the thiomethyl group (-S-CH₃)-containing amino acid L-methionine generated methionine sulphoxide as a major product (Figure 2), a process consistent with equation (2), where R represents H₃N⁺CH ()CH₂CH₂-. Indeed, the -SO-CH₃ protons of this oxidation product has a characteristic singlet resonance located at 2.725 ppm, and it’s concentration is readily monitored by the (relative or normalised) intensity of this signal. Further ¹H NMR signals which were consistent with the generation of methionine sulphoxide were multiplets located at 2.30, 3.02 and 3.85 ppm, which correspond to the β-CH₂, γ-CH₂ and α-CH protons, respecttively, in this oxidation product. No evidence for the production of methionine sulphone (characteristic-SO₂- CH₃ singlet resonance located at δ = 3.09 ppm), which may arise from the oxidation of methionine sulphoxide by further O₃, was provided in these investigations. Under our experimental conditions, the yield of methionine sulphoxide produced was 98% ± 4%, i.e. the reaction was essentially complete and quantitative.

L-cysteine was also chosen for these chemical model system experiments since we have been unable to detect this thiol in human saliva by ¹H NMR analysis in view of its low concentration in the “free” (non-protein-incorporated) state, and also its complex ABX coupling pattern (i.e., no clearly-visible sharp resonances of low multiplicity). ¹H NMR analysis demonstrated that exposure of aqueous solutions of this biomolecule [with its characteristic ABX coupling pattern of resonances centred at 3.02 and 3.10 ppm (AB protons) and 3.97 ppm (X proton)] to O₃ (Section 2) generated its corresponding disulphide, cystine, as a major product (data not shown). Indeed, a reference spectrum acquired on an authentic sample of L-cystine confirmed its identity (clear 4-line

multiplet signals located at 3.20 and 3.41 ppm (AB protons), and a further one at 4.14 ppm (X proton)). Electronic integration of the L-cysteine and cystine resonances demonstrated that 95% ± 6% of the former O₃ scavenger was oxidatively transformed to the latter oxidation product. No evidence for the generation of cysteine’s higher oxidation products (such as cysteic acid) was obtained.

These observations are explicable by the stepwise processes described in equations (3)-(5), i.e. deprotonation of the thiol followed by transfer of an electron from the thiolate anion (RS⁻) to O₃, and then combination of two thiyl radicals (, generated in the reaction depicted in equation (4)) to produce cystine (represented as RSSR, where R = H₃N⁺CH()CH₂-).

However, it should be noted that the aggressivelypowerful oxidant hydroxyl radical () can arise from protonation of the species followed by decomposition of the resulting radical adduct (equations (6) and (7)). The radical generated in this manner can then

also serve to oxidise L-cysteine to cystine via the prior abstraction of an electron from the thiolate anion (equation (8)), a process again forming thiyl radicals which then also combine to yield the corresponding disulphide (equation (5)).

¹H NMR analysis also showed that treatment of phosphate-buffered aqueous solutions of α-D-glucose with O₃ in the manner described in Section 2 generated formate as a major reaction product (identified as singlet ¹H NMR resonance located at δ = 8.46 ppm, Figure 3), an obser-

vation consistent with previous studies conducted on the interactions of ROS (particularly radiolytically-generated radical) with carbohydrates in general [16]. A mean ± SEM formate concentration of 1.21 ± 0.11 mM (representing a 24% ± 2% yield of this oxidation product) was produced from the 5.00 mM glucose substrate solution.

The differences observed in the oxidising capacity of O₃ towards pyruvate, methionine, cysteine and α-glucose can be rationalised in terms of the relative scavenging efficacies of these antioxidants tested (L-methionine ≈ L-cysteine ≈ pyruvate > α-D-glucose). Also of much importance for consideration is their mean salivary (or alternative oral fluid or tissue) concentrations. Indeed, for human saliva, these mean values are in the order pyruvate/pyruvate hydrate (1.67 mM [2]) >> methionine (mean pre-O₃ treatment salivary level 274 µM, a value estimated from the data acquired in this investigation) >> cysteine (mean concentration 32 µM [17]): the glucose concentration of this particular oral fluid is highly variable, but little or none of it is detectable by ¹H NMR analysis of samples collected following an 8 hr. overnight fasting episode [2].

When expressed relative to tryptophan, the rate constants for the reaction of O₃ with methionine, cysteine, and glutathione were found to be 0.77 ± 0.08, 0.88 ± 0.19 and 0.42 ± 0.01 respectively [8] (these relative rate constants correspond to O₃ rather than scavenger consumption). Hence, cysteine reacts with O₃ at a very similar rate to that of methionine, although it is clear that in human saliva, the latter thioether will serve as a much more effective O₃ scavenger than the former thiol in view of its much higher concentration therein (approximately 10-fold greater).

The oxidative decarboxylation of pyruvate by O₃ provides a mechanistic example of the therapeutic actions exertable by the latter therapeutic agent (e.g. its caries-preventative, cariostatic and further therapeutic properties). Indeed, pryruvic acid is an extremely powerful organic acid (proton donor) since it has a pK_a value of 3.20 mM, and therefore is much more powerful than lactic acid (K_a = 0.14 mM) in this context [18]. Indeed, we have previously noted that pyruvic acid may play a significant role in the facilitation of enamel erosion and tooth demineralisation processes [19], and its removal in human saliva, alternative oral fluids, carious dentin and/or plaque by O₃ may serve to inhibit the induction or further development of primary root caries lesions, especially since salivary organic acids in their unionised forms readily diffuse into tooth enamel [20].

The oxidative removal of the amino acids methionine and cysteine (both of which are detectable in human saliva, the former by ¹H NMR analysis) by O₃ represents an observation of some significance in relation to both oral hygiene and clinical periodontology, since hydrogen sulphide (H₂S) and methyl mercaptan (CH₃SH) are produced from these agents via the metabolic pathways of gram-negative bacteria. Indeed, the enzyme cystine reductase reduces cystine to cysteine, and serine sulphydrase gives rise to the desulphydration of cysteine, a process generating H₂S and serine [21,22]. Therefore, data acquired here also provide evidence that O₃ may have the ability to clinically suppress oral malodour via the direct oxidative consumption of VSCs and/or their particular amino acid precursors, especially if applied in the form of an O₃-containing oral rinse system, i.e. ozonated water. In relation to these investigations, in 1992 Grigor and Roberts [17] found that hydrogen peroxide (H₂O₂) significantly diminished salivary thiol levels both in vivo and in vitro.  

However, it is also conceivable that the bacterial enzyme cystine reductase may reverse the O₃-mediated transformation of cysteine to cystine in the oral environment, although it should be noted that O₃ may be present in excess over that of total oral fluid or tissue thiol concentrations, especially when the gaseous concentration is considered along with that present in aqueous solution, the latter level being limited to its solubility in water at physiological temperature, estimates of which are influenced and complicated by that present in the local gas phase and also pH values [23]. However, O₃ determinations in our laboratory performed by a modification of a previously reported method [15] gave a mean concentration of ca. 20 ppm (0.42 mM) when 40.0 mM phosphate buffer solutions (pH 7.00) were treated with this gaseous oxidant under the same experimental conditions as the biomolecule solutions employed for this study (section 2.1), i.e. at ambient temperature and pressure. Hence, a significant fraction of salivary thiols (or those present in alternative oral environments) are expected to be retained in their oxidised (disulphide) forms following reaction with O₃. Furthermore, it is probable that O₃ has the capacity to oxidatively inactivate bacterially-derived cystine reductase, together with a range of further enzymes available at its site of therapeutic application.

It is also of much importance to note that methionine is one of only two amino acids encoded by a single codon (AUG) in the standard genetic code (tryptophan, encoded by UGG, is the other) [24]. The codon AUG also represents the “Commencement” message for a ribosome that signals the initiation of protein translation from mRNA. Consequently, methionine is incorporated into the Nterminal position of all proteins in eukaryotes and archaea during translation, although it is usually removed via post-translational modification. This phenomenon is also clearly of much importance to bacteria; however, in this case, the N-formylmethionine derivative is employed as a primary amino acid adduct [25].

Notwithstanding, it is also important to note that singlet oxygen (¹O₂) is also generated in high yield from the reaction of O₃ with a wide range of biomolecules, including the amino acids methionine and cysteine, and the tripeptide glutathione, together with urate, ascorbate, NADH, NADPH and albumin [26] (that involving methionine, which has a thioether group in its side-chain, is depicted in Equation (9)), and therefore this further ROS may also play an important role in perpetuating damage to biomolecules primarily exerted via the actions of O₃. Indeed, methionine also readily reacts with ¹O₂ to again form its corresponding sulphoxide, a process postulated to involve a highly-reactive persulphoxide intermediate [27] [RS⁺(CH₃)-O-O^– in the case of this substrate, where R represents H₃N⁺CH()CH₂CH₂-].

Interestingly, for cysteine, the mechanism of cystine generation therefrom has been suggested to involve the reaction of an intermediate sulphoxide species (RSOH) with a second molecule of the cysteine substrate [26] (equation (10)).

All of the metabolites investigated here also serve to act as biomolecular protectants against any adverse or toxic effects exertable by the “leakage” of therapeutically-applied O₃ to soft tissue areas in the oral environment which are or may be accessible by oral fluids containing them (in particular, but not limited to, human saliva). In this manner, these biomolecules have the ability to circumvent any deleterious O₃-mediated cell and tissue damage potentially arising from this intra-oral dissemination process.

Indeed, the oxidative modifications of pyruvate, methionine, cysteine and glucose observed here serve as an important examples of the capacity of salivary metabolites to offer protection against any O₃ which diffuses away from its site of application, for example, from primary root carious lesions.

High-resolution, high-field ¹H NMR spectroscopy is a technique which offers many advantages over alternative labour-intensive and onerous analytical methodologies for evaluating the biomolecular fate of O₃ in oral fluids since (1) it permits the rapid, non-invasive and simultaneous examination of a variety of metabolic species and their O₃-induced oxidation status in chemical model systems, and (2) little or no knowledge of sample composition is required prior to analysis. Moreover, chemical shift values, coupling patterns and coupling constants of resonances present in the ¹H NMR spectra of such model  systems provide much valuable information regarding the molecular nature of oxidation products generated from the interaction of biomolecules with O₃.