|Home | About | Journals | Submit | Contact Us | Français|
Microbial contamination of whole human saliva is unwanted for certain in vitro applications, e.g., when utilizing it as a growth substratum for biofilm experiments. The aim of this investigation was to test gamma irradiation for its suitability to sterilize saliva and to investigate the treatment's influence on the composition and integrity of salivary proteins in comparison to filter sterilization. For inhibition of bacterial growth by gamma irradiation, a sterility assurance level of 10−6 was determined to be reached at a dose of 3.5 kGy. At this dose, the integrity of proteins, as measured by fluorescence, circular dichroism, and gel electrophoretic banding pattern, and the enzymatic activities of salivary amylase and lysozyme were virtually unchanged. Filtration reduced the total protein concentration to about half of its original value and decreased lysozyme activity to about 10%. It can be concluded that irradiation is suitable for sterilizing whole saliva in its native form.
Saliva, a complex biological fluid, is essential for the processing of food and maintenance of health in the oral cavity and the upper digestive tract (20). Many of these beneficial functions are mediated by the numerous proteins present in saliva (12, 41) that also modulate the microbial colonization of the mouth (32, 35). However, in vitro functional studies involving human whole mixed saliva, the fluid that actually is present in the mouth in vivo, are often hampered by the presence of bacteria shed from resident oral microbial biofilms (1).
There have been longstanding and continuous efforts to purge saliva of viable microorganisms for various experimental purposes, such as its use as a substrate for bacterial or fungal growth (5, 6, 18, 22, 39), studies of microbial adhesion to tooth or biomaterial surfaces (3, 23-25, 28, 30), bacterial coaggregation (17), or biofilm formation (9, 27), measurement of antimicrobial activities (14), transformability of oral bacteria with DNA (21), evaluation of tissue regeneration or host response in the presence of saliva (11, 13, 45), oral hygiene product testing (19), and even for preparing autologous saliva as a supplement during radiotherapy for head and neck cancer (40).
Methods used to sterilize saliva include filtration, pasteurization, gamma irradiation, and UV irradiation, as well as hydrogen peroxide, ethylene oxide, or chlorhexidine treatment. The relative efficacies of most of these methods have been evaluated by Williams and Kraus (46). Filtration has been deemed to be preferable (16) and is now the most commonly used method (9), even though it has been reported that the amount of total salivary protein as well as enzyme activities can be decreased by filtration (16, 40, 46).
In order to achieve sterilization of saliva with minimum loss of components or alteration of their functional activities, gamma irradiation, which has been increasingly employed for preservation of foodstuffs, medical products, pharmaceuticals, cosmetics, and sterile packaging materials (47), deserves to be evaluated as a means of sterilizing saliva. Cobalt-60 is typically used as a source of gamma rays to ionize chemical bonds. There are two possible mechanisms of action in the destruction of microorganisms by gamma irradiation. One is the irreversible damage to critical biomolecules of the bacterium, most specifically, the DNA. The other way, presumably more effective, is the ionization of water molecules present in and around the bacteria, which produces free radicals that attack macromolecules (47). We hypothesized that irradiation of saliva would be a valid alternative to filter sterilization, provided that the radiation doses could be delineated that kill bacteria but leave salivary proteins unaffected. Thus, in the present investigation, the influence of gamma irradiation sterilization on protein integrity was studied and compared to filter sterilization.
Subjects’ rights were protected by review through the Ethics Committee of the Medical Faculty of the University of Regensburg, and informed consent was granted. Whole saliva was obtained separately from three different individuals by expectoration into sterile 50-ml polypropylene vials (Blue Max, Falcon; Becton Dickinson Labware, Franklin Lakes, NJ) with continuous stimulation of salivary flow through chewing of paraffin-like material (parafilm; Pechiney Plastic Packaging Company, Chicago, IL). Prior to collection of saliva, donors rinsed their mouth with water and refrained from eating or drinking for at least 2 h. For each irradiation and filtration experiment, 20 to 35 ml of saliva was freshly collected from each donor. All saliva samples were kept on ice during collection and throughout all further experimental procedures. To allow for precise pipetting, the viscosity of saliva was broken by repeatedly drawing the samples through a sterile blunt needle (0.9 by 22 mm; Miraject, Duisburg, Germany) fit to a sterile 5-ml syringe (Luer-Lok; Becton Dickinson, Droghede, Ireland). After completion of the experimental treatments, aliquots of saliva for microbial counting were plated onto culture plates, and aliquots for gel electrophoresis were denatured in SDS sample buffer. Samples for analysis of protein integrity were frozen and stored at −80°C. For later use in the enzymatic activity tests, samples were slowly thawed at 4°C overnight. During the assays, samples were kept on ice at all times to preserve the activities of the enzymes. Total protein concentrations were determined with the use of the bicinchoninic acid protein assay reagent (Pierce, Rockford, IL).
The salivary samples from three separate individuals were irradiated by the company Isotron (Allershausen, Germany). Aliquots (1 ml) of each donor's saliva were kept on ice in sterile 2-ml polypropylene screw-cap microcups (Sarstedt & Co., Nümbrecht, Germany) and exposed to different doses of gamma irradiation by use of an industrial-sized cobalt-60 source (Pallet Irradiator type JS 9000; Nordion International Inc., Kanata, Ontario, Canada) with an energy output of 1.17 MeV (99.9%) and 1.33 MeV (100%). Samples were placed at certain distances from the gamma radiation source and were exposed for certain time intervals that had been previously determined as optimal to obtain the calculated doses. Time intervals varied between 5 min (1 kGy) and 1 h (12 kGy). Two film dosimeters were mounted to each set of samples and evaluated later to precisely record the doses actually obtained. Nonirradiated control samples were kept under the same conditions except that they were not exposed to the radiation source. For the evaluation of irradiation to larger volumes of saliva, 20-ml aliquots from each donor were placed in sterile syringes, capped, sealed, and exposed to the radiation source at doses between 3 and 5 kGy as described above.
Filtration was performed under sterile conditions in a laminar flow hood. Saliva from two separate individuals was passed consecutively through sterile syringe tip membrane filters of decreasing pore size (Minisart CE single-use syringe filters, hydrophilic; 5.0, 1.2, 0.8, 0.45, and 0.2 μm; Sartorius AG, Göttingen, Germany) fit to a sterile 5-ml Luer-Lok syringe (Becton-Dickinson).
All microbial cultures grown on plates were incubated for 2 days at 37°C. Deinococcus radiodurans ATCC 13939 was cultured on low-salt Luria-Bertani medium containing 5 g/liter glucose (LSLB-Glc) followed by overnight liquid culture in LSLB-Glc medium at 37°C under constant motion. Escherichia coli ATCC 25922 was grown on MacConkey agar (Merck, Darmstadt, Germany), Geobacillus stearothermophilus ATCC 7953 was grown on CASO agar (casein-peptone soymeal-peptone agar; Merck, Darmstadt, Germany), and Candida albicans ATCC 10231 was grown on Albicans ID2 agar (bioMérieux Deutschland GmbH, Nürtingen, Germany).
Salivary samples were serially diluted in saline solution, and 50-μl aliquots of each sample were plated in triplicate on CASO agar. For spiking experiments, 900 μl of saliva was mixed with 100-μl aliquots of suspensions of D. radiodurans in saline solution to obtain an estimated final concentration of 106 deinococci per ml. For determination of bacterial viability following experimental treatments, samples containing the pure bacterial strains or samples spiked with D. radiodurans were further cultured in their appropriate growth medium as described above. Bacterial colonies were counted manually from plates that contained between 50 and 100 colonies. For sterility tests, 500-μl aliquots of saliva samples were added to sterile blood culture flasks (20 ml; BacT/Alert PF; bioMérieux Deutschland GmbH, Nürtingen, Germany). From the 20-ml volumes of irradiated saliva, 1-ml aliquots were added to triplicate blood culture flasks. After incubation for 48 h at 37°C, blood culture flasks were visually evaluated for color change, following the instructions of the manufacturer.
Salivary proteins were denatured under reducing conditions, separated by SDS-PAGE (0.75 μg per lane) on 4-to-20% gradient gels (Invitrogen, Karlsruhe, Germany) run under a constant voltage of 125 V for 90 min in a vertical gel apparatus (X Cell II minicell; Novex Electrophoresis GmbH, Frankfurt, Germany), and visualized by means of an ammoniacal silver stain kit (SilverXpress; Invitrogen, Carlsbad, CA) or transferred (10 μg per lane) to nitrocellulose with a 0.4-μm pore size (Invitrogen) in a semidry electroblotting apparatus (X Cell II blot module; Novex) as previously described (31). Molecular weight standards, unstained for silver stains and prestained for blotting (Broad Range and Precision Plus; Bio-Rad Laboratories, München, Germany), were included on all gels. Glycoproteins were labeled by a modification of the hydrazide method as previously described (33) using 10 mM sodium meta-periodate (Fisher, Fairlawn, NJ) for oxidation of sugars and biotin-LC-hydrazide (Pierce, Rockford, IL) together with avidin D-horseradish peroxidase (Vector, Burlingame, CA) and 4-chloro-1-naphthol (Bio-Rad Laboratories) for detection and staining, respectively. Gels and blots were scanned with a high-resolution scanner (Sharp color image scanner, model JX-330, fit with a Sharp film scanning unit, model JX-3F6) in the transparent and reflective modes, respectively.
Thirty microliters of saliva was mixed thoroughly with 970 μl of 17.5 mM sodium potassium phosphate buffer (pH 7.0), and fluorescence excitation and emission of 1:33.3-diluted saliva samples were recorded with a spectrofluorometer (Jobin Yvon Spex FluoroMax-2; HORIBA Jobin Yvon GmbH, Unterhaching, Germany). A wavelength of 280 nm was chosen for excitation of the intrinsic fluorophores tryptophan and tyrosine, and 295 nm was used for selective excitation of tryptophan. For evaluation of the emission data, a wavelength of 340 nm was selected. Circular dichroism spectra of 1:5-diluted saliva samples were recorded in a spectropolarimeter (Aviv 62A-DS circular dichroism spectrometer; AVIV Biomedical, Lakewood, NJ). Spectroscopic data were evaluated and interpreted as described previously (7, 8).
Amylase activity was measured according to a modified version of the Wohlgemuth procedure (34). A range of 1:5 to 1:50 dilutions of saliva with 17.5 mM sodium potassium phosphate buffer (pH 7.0) was prepared. Seventy-five microliters was mixed with 300 μl 1% (wt/vol) starch (Merck, Darmstadt, Germany) and 375 μl buffer and kept at a constant temperature of 25°C by using a Dri-Block (DB-3; Techne Inc., Princeton, NJ). The hydrolysis of starch was stopped at 5-min intervals over a period of 25 min by reaction with an iodine-potassium iodide test solution. Absorbance measurements at a wavelength of 578 nm were taken at 25°C by using a spectrophotometer (UV/VIS spectrophotometer V-530; JASCO GmbH Deutschland, Groß-Umstadt, Germany). Standard curves were obtained by using serial dilutions of porcine pancreas α-amylase (Sigma-Aldrich, München, Germany). Enzymatic activity of lysozyme was measured using the method described by Shugar (36). Thirty microliters of saliva either undiluted or diluted 1:3 in 66 mM sodium phosphate buffer containing 17 mM NaCl (phosphate-buffered saline [PBS], pH 7.0) was added to 2,970 μl of a cell suspension of Micrococcus lysodeikticus (Sigma-Aldrich, München, Germany) in PBS (0.2 mg/ml). The decrease in extinction (turbidity) was measured in the spectrophotometer at a wavelength of 450 nm for a continuous time period of 3 min at 25°C. Serial dilutions of chicken egg white lysozyme (Serva, Heidelberg, Germany) were used to obtain standard curves.
Statistical calculations were performed using the software SPSS PC+ version 5.01 (SPSS Inc., Chicago, IL). Mean CFU were calculated from triplicate culture plates. Medians together with the 25% and 75% quantiles for CFU, protein concentrations, and enzyme activities before and after filtration or irradiation were calculated with the help of the SigmaPlot software (version 8.0; SPSS Inc.) for two donors in a total of five independent experiments and for three donors in a total of three independent experiments, respectively. The radiation dose that caused a 10-fold (90%) reduction in viability (D10) and the sterility assurance level (SAL) in saliva, as defined by a reduction to a theoretical value of 10−6 CFU per ml, were calculated by linear regression of values, not including values below the theoretical detection limit.
To determine the functionality and effectiveness of the irradiation procedure, the well-characterized radio-resistant strain D. radiodurans ATCC 13939 was used as a control and compared to suspensions of E. coli strain ATTC 25922 and G. stearothermophilus strain ATCC 7953 (Fig. (Fig.1A).1A). Viabilities of both E. coli and G. stearothermophilus cells were reduced to below detection limits by irradiation with doses of 1 kGy and above. Analogous results were found for C. albicans (data not shown). In contrast, viability of D. radiodurans was reduced by only approximately 10- to 100-fold, even at doses as high as 10 kGy. When D. radiodurans was suspended in saliva instead of saline before irradiation, no significant differences in viability could be observed, thus excluding the possibilities of a radio-protective or radio-sensitizing effect of saliva suspension on bacterial viability. The decline of D. radiodurans viability was found to be almost linearly related to the increase in irradiation dose (r2 = 0.73 and 0.84 in saline and saliva, respectively), showing the reliability of the irradiation method over this wide dose range.
When whole saliva was irradiated, bacterial viability for all samples fell below detection limits at doses of more than 2.5 kGy (Fig. (Fig.1B).1B). Blood cultures from irradiated saliva samples were found to be negative at doses of 3.0 kGy and beyond. The cumulative value for the SAL in saliva was calculated to be reached at a dose of 3.54 kGy (Fig. (Fig.1B,1B, insert). The D10 value was calculated to be an interval of 0.25 kGy in radiation dose. When larger volumes of saliva (20 ml instead of 1 ml) were irradiated, 5.0 kGy was found to be sufficient for inhibiting bacterial growth, as evidenced by negative blood cultures (data not shown).
When saliva was filtered successively through membrane filters of decreasing pore size, a 10- to 100-fold decrease in CFU was observed after filtration with pore diameters ranging from 5.0 to 0.80 μm (Fig. (Fig.1C).1C). A significant drop in CFU to below detection limits was found only after use of a filter with a pore size of 0.45 μm. This was confirmed by the use of blood cultures, which did not detect any residual bacterial growth in samples filtered with pore sizes of 0.45 or 0.20 μm (data not shown).
Salivary proteins were subjected to separation by SDS-PAGE and stained with silver to evaluate a possible effect of irradiation and filtration on protein denaturation and loss of protein components, respectively (Fig. (Fig.2).2). At doses up to 5 kGy, no significant change of the protein subunits was observed. When salivary samples were irradiated with doses higher than 5 kGy, however, protein bands appeared more diffuse, which is indicative of denaturation of proteins (Fig. (Fig.2A).2A). This effect became most pronounced at the highest doses applied (>10 kGy). Analogous results were obtained when nitrocellulose transfers were stained for glycoproteins by the hydrazide method (data not shown). Similarly, filtration changed the pattern of protein bands only slightly. However, at all pore sizes substantial losses of three protein components (approximately 11, 27, and 48 kDa) were observed. The most pronounced effect was visible after initial filtration with a pore size of 5.0 μm, which resulted in the almost complete loss of a low-molecular-mass component of about 11 kDa (Fig. (Fig.2B).2B). The loss of this band was consistently observed with saliva from different donors and at different sampling times. This effect cannot be accounted for by simple size exclusion of such a small molecule, because of the large pore size of this filter membrane. Although the reported molecular mass of human lysozyme is 14.7 kDa, it is still possible that the protein band at the molecular range of 11 kDa represents a salivary form of lysozyme. Notably, few alterations in the relative proportion of large-molecular-size glycoprotein bands were observed on hydrazide-stained transfers (data not shown). Furthermore, measurement of far UV circular dichroism spectra of the salivary probes revealed no changes of the secondary structure (helix content) of the involved proteins, after either irradiation or as a consequence of the filtration process (data not shown).
The concentration of total proteins was not influenced by irradiation (Fig. (Fig.3A),3A), but filtration through the 5.0-μm-pore membrane resulted in a substantial drop in total protein concentration, while subsequent steps with smaller-pore diameters had no further effect on total protein concentration (Fig. (Fig.3B),3B), in agreement with the results from the SDS-PAGE. Increasing doses of radiation caused a decrease in intrinsic fluorescence by about 30% at 10 kGy (Fig. (Fig.3C),3C), which indicates a decrease in the integrity of the fluorophores tryptophan and tyrosine and, thus, suggests that these amino acids become degraded at higher doses of irradiation. Initial filtration through the membrane with a 5.0-μm pore size resulted in about a 50% loss of intrinsic fluorescence but remained unchanged after filtration with smaller pore sizes (Fig. (Fig.3D).3D). This observed decrease most likely results from the loss of total protein concentration rather than from degradation of intrinsic fluorophores because it parallels the drop in protein concentration seen in Fig. Fig.3B.3B. The enzymatic activity of amylase was decreased by irradiation to less than 50% of its original level at 10 kGy (Fig. (Fig.3E),3E), whereas filtration had no apparent effect on amylase activity (Fig. (Fig.3F).3F). Interestingly, the opposite was observed for lysozyme, in that irradiation had only marginal effects on enzyme activity whereas filtration reduced lysozyme activity to 10% or less of its original level already after the initial filtration step through the 5.0-μm pore size.
For investigating functional effects of saliva, filtration has been widely, and without much critical evaluation, adopted as a standard method whenever sterile saliva is needed. With the current advances in protein sterilization by gamma irradiation, developed largely in the food and medical industries, this technique deserves to be reevaluated for the sterilization of whole human saliva. The results of the present study show that radiation doses can be delimited that will sufficiently inactivate bacterial viability in saliva but will not considerably influence salivary protein composition, integrity, or function. In comparison, filtration caused not only a substantial decrease in overall protein concentration but also a selective loss of certain proteins. Thus, depending on the intended experimental purpose, gamma irradiation sterilization of saliva could in certain instances be preferable to filter sterilization.
Irradiation doses of 3.5 kGy, which were found to be sufficient for suppressing bacterial growth in salivary samples, caused virtually no protein denaturation, as seen by the electrophoretic banding pattern, and only a modest decrease in enzymatic activities. However, not all proteins appeared to be equally sensitive to the effects of radiation, as evidenced by the much higher radio-tolerance of lysozyme compared to amylase. On one hand, this may be due to intrinsic structural differences that endow a given protein with less sensitivity to radiation. On the other hand, it could be that more-dilute salivary proteins, exemplified by lysozyme (80 μg/ml in whole saliva ), tend to be less affected by radiation damage than those which occur at a relatively higher concentration, i.e., salivary amylase (650 to 800 μg/ml in parotid saliva ). The window of radiation dosage sufficient for achieving sterility of saliva without causing significant damage to salivary proteins can be estimated to range from 3 to <5 kGy and is similar to what was found in a previous study that used electron beam radiation derived from a linear accelerator. The disadvantage of the latter method is the long time required to accumulate the desired doses necessary for sterilization (10 h for 2.5 kGy) (40), whereas less than 20 min of exposure to the cobalt-60 source used in the present study was required for sterilizing the salivary samples. Such short irradiation periods also allow for better control of collateral heat generation throughout the irradiation process, because the samples are kept in an ice bath or exposed in a frozen state to the radiation source (47).
Filtration reduced the total protein concentration by about one-half. This confirmed the findings in a previous report in which a similar decrease in protein concentration was found (40). In that previous study, only a pore size of 0.45 μm was used, whereas in the present investigation, the most significant decrease in protein concentration was observed already by use of a much greater filter pore size of 5.0 μm. It is unlikely that the effect is solely due to protein being trapped by bacteria, as had been suggested earlier (38), because for microbial reduction the most significant drop was observed only after filtration with a 10-fold-smaller pore size of 0.45 μm (Fig. (Fig.1C).1C). Since a pore diameter of 5.0 μm is above the average size of a single salivary protein by several orders of magnitude, this suggests that a large part of protein in whole saliva is organized in very large multicomponent protein aggregates. Indeed, formation of supramolecular complexes involving salivary mucins has been reported (29, 37, 44). How those supramolecular structures are related to salivary micelle-like globules (48) or the recently discovered exosomes in saliva (10, 26), and also what role bacterial aggregates play, deserves further investigation. The selective loss of lysozyme after filtration with the 5.0-μm pore size is perplexing as well and raises the question as to how such a small molecule could be filtered out by a comparatively gigantic pore size. Selective binding of lysozyme, which has a positive net charge, to the negatively charged cellulose acetate membrane of the filter may occur but is unlikely to account for the drastic drop to only 10% of its original activity, since the membrane's binding capacity will be saturated while saliva still continues to percolate. It is also possible that lysozyme is bound to larger bacterial aggregates retained in the 5-μm filter, but an alternative explanation could be that it is bound to protein aggregates too large to pass the 5-μm pore size of the filter. In this regard, it was suggested previously that the supramolecular salivary mucin matrix is decorated with protective factors which also include lysozyme (15, 43, 44). Indeed, in human bronchial secretions, lysozyme shows a strong ionic interaction with mucins (4, 42).
If a sterile saliva is wanted that most closely resembles the true in vivo oral fluid, sterilization of saliva by gamma irradiation is preferable over filtration because it alters the original composition and biological activity the least. This type of fluid will still contain particulate matter and bacterial remnants as they were present under in vivo conditions. The question ultimately will be whether this is appropriate for the intended experimental purpose. It will also be necessary to decide for each experimental setting, depending on the bioburden of the saliva and duration of the experiment planned, how high a sterility assurance level will be needed versus how much of the original biological or enzymatic activity needs to remain. This will eventually determine the dose of radiation to which a particular sample has to be exposed. If filtration is chosen to sterilize saliva, one has to be aware that the resulting fluid will be significantly altered in its composition, particularly if salivary proteins are of importance. Fortunately, difficulties encountered in the past when sterilizing saliva by gamma irradiation, including the scarcity of conveniently available and powerful Co60 sources, can now be overcome by using specially designed industrial facilities for irradiation that achieve safety standards and system requirements set out by the Nuclear Regulatory Commission.
We are grateful to Verena Wittmann and Barbara Kellerer for excellent technical help and to the employees of Gammamaster Deutschland GmbH in Allershausen, Germany, particularly Reiner Eidenberger, Albert Löw, and Michael Spies for their friendly assistance in performing the irradiation experiments. We further thank Frank Scannapieco and Molakala Reddy for their critical reviews of the manuscript.
This work was supported by Deutsche Forschungsgemeinschaft grants Ru409/4-1 and SFB 585, B5 (S.R.).
Published ahead of print on 10 December 2010.