Effect of protease inhibitors on the quantitative and qualitative assessment of oral microbes

Abstract

Protease inhibitor cocktails are routinely added to clinical samples used for proteomic studies to inactivate proteases. Since these same samples are often used for microbial studies, we determined whether addition of protease inhibitors could affect the quantitative or qualitative assessment of microbial profiles. Twenty-two saliva samples were collected and processed immediately with or without the addition of a protease inhibitor cocktail. Conventional cultivation methods were used to evaluate total bacterial growth. Total genomic DNA was isolated and a specific 16S rRNA gene-targeted region was PCR-amplified and separated by denaturing gradient gel electrophoresis. A combination of 1D SDS-PAGE and LC-MS/MS methods was used to determine the effect of the protease inhibitors on the integrity of salivary proteins and peptides. Interestingly, no significant differences were observed in either the bacterial growth and composition or the integrity of salivary proteins between the two groups. Correlation coefficients between the paired samples for total cultivable microbiota (r²=0.847), total mutans streptococci (r²=0.898), total oral lactobacilli (r²=0.933), and total Streptococcus mutans (r²=0.870) also exceeded expected values. The results suggest that the addition of a protease inhibitor cocktail in saliva samples does not impact the growth of oral microbiota or compromise the ability to characterize its composition.

Introduction

Proteases are widely distributed in most animals, plants, and microorganisms. They constitute one of the largest functional groups of proteins (Rawlings, et al., 2010). Proteases play critical roles in regulating the activity of proteins and enzymes during the life cycle of cells. Proteases inhibitors (PI) inhibit the proteolytic cleavage of proteins, and they are generally grouped into five major types: cysteine, serine, threonine, aspartate, and metalloproteinase, according to the amino acid active site responsible for proteolytic cleavage (Supuran, et al., 2002). Since proteases are involved in intra- and extracellular biological processes, protease inhibitors also play an important role in modulating multiple molecular events in all forms of organisms. It has been reported that some protease inhibitors affected bacterial morphological differentiation, evasion ability, biofilm formation, and the acquisition of nutrients (Curtis, et al., 2002, Armstrong, 2006, Tsang, et al., 2008). Studies have demonstrated that the presence of endogenous protease inhibitors, such as cystatins in saliva, inhibited the growth of some oral bacteria, but caused others to proliferate (Blankenvoorde, et al., 1998, van Nieuw Amerongen, et al., 2004). Other studies reported that exogenous PI, such as aprotinin, bestatin, and leupeptin, could also interfere with the growth and proliferation of oral bacteria (Grenier, et al., 2001, Labbe, et al., 2001, Ibrahim, et al., 2002).

In recent years, adding a PI to clinical samples has been recommended as a means of controlling enzymatic protein degradation caused by liberated or activated endogenous protease during cell membrane disruption and protein preparation. However, it remains unknown whether this routine protocol can interfere with either a count of total cultivable bacteria or an analysis of changes in oral bacterial composition.

Over 500 bacterial species have been identified in human oral cavity (Aas, et al., 2005). Quantifying total cultivable bacteria or a specific bacterial species has typically relied on in vitro cultivation methods. Recently, our group and others have demonstrated the use of denaturing gel gradient electrophoresis (DGGE) to evaluate the composition of cultivable and uncultivable oral microbial communities (Li, et al., 2005, Li, et al., 2006, Li, et al., 2007). The DGGE approach extracts genomic DNA and specifically targets regions of 16S rRNA gene which are amplified by PCR. Subsequently, the PCR amplicons are analyzed on a denaturing gel that separates DNA fragments according to their nucleotide composition. The present study used both in vitro cultivation and PCR-DGGE methods to evaluate the effect of a PI cocktail on total cultivable bacterial growth and composition in saliva as well as the effect of PI on salivary proteins.

Materials and methods

Saliva sample collection

This study was approved by the Institutional Review Board of New York University School of Medicine for Activities Involving Human Subjects. Twenty-two stimulated whole salivary samples were obtained from 10 adult subjects. The subjects were first asked to rinse their mouth with water and then chew a piece of neutral gum base to stimulate saliva flow. On average, 4~5 ml of saliva were collected from each subject into a 50 ml sterile plastic conical tube held on ice. A 2 ml aliquot was mixed with 20 μl protease inhibitor cocktail (Halt™, Thermo Scientific, USA; Stock inhibitor concentrations are as follows: AEBSF, 1mM; Aprotinin, 800 nM; Bestatin, 50 μM; E64, 15 μM; Leupeptin, 20 M; and Pepstatin A, 10 μM). A second 2 ml aliquot was preserved without inhibitors. The samples were maintained on ice and processed within 1 h after collection.

Bacterial cultivation

After each saliva sample was vortexed briefly for 10s, 200 μl were mixed with 1.8 ml reduced transport fluid (RTF) buffer (Syed & Loesche, 1972). Finally, 50 μl of serially diluted (1/10, 1/100 and 1/1000 with 1× phosphate buffered saline) samples were plated, using an Autoplate™ 4000 (Spiral Biotech, Bethesda, MD), onto an enriched tryptic soy agar (ETSA) and three selective media: mitis-salivarius (MSA), mitis-salivarius-bacitricin (MSB) and Rogosa, respectively. After 72 h of anaerobic incubation (85% N₂, 10% CO₂, and 5% H₂) at 37°C, the colony-forming units (CFUs) were counted to provide an estimate of the level of total cultivable bacteria on the different media (ETSA for total cultivable, MSA for total mutans streptococci, MSB for S. mutans (Gold, et al., 1973), and Rogosa for total oral lactobacilli).

Genomic DNA isolations

Total genomic DNA of the saliva samples was extracted from two sets of bacterial samples: whole saliva and total cultivable bacterial colonies grown on ETSA plates. More specifically, the whole saliva sample was centrifugated for 3 min at 18000 × g. The supernatant was discarded, and total bacterial genomic DNA was extracted from the pellet. The total cultivable bacterial colonies grown on ETSA media were collected with a cotton swab and washed in 1.5 ml TE buffer for DNA isolation. DNA purification kit (MasterPure, Epicentre, Madison, WI) combined with a solution of phenol/chloroform/isoamyl alcohol (25:24:1) at pH 8.0 was used for all isolation procedures, as previously described by our group (Li, et al., 2007). The quality and quantity of the DNA were measured using a UV-spectrophotometer at 260 nm and 280 nm (Nanodrop 1000, Thermo Scientific, USA). The final concentration of each DNA sample was adjusted to10 ng/μl for all PCR applications.

PCR Assay

PCR amplification of bacterial 16S rRNA gene fragments used the GeneAmp® PCR System 9700 (PE Applied Biosystems). Initially, the complete 16S rDNA gene locus (~1500-bp) was pre-amplified for DNA extracts with a set of universal 16S rDNA gene primers (Lane, 1991). A second PCR reaction was performed after using a different set of universal bacterial 16S rDNA gene primers (prbac1 and prbac2) (Rupf, et al., 1999) with a 40-nucleotide GC-clamp as previously described (Sheffield, et al., 1989). Each PCR reaction mixture and PCR condition have been previously published with details (Li, et al., 2007).

Assay for microbial profile analysis

The characterization of the total bacterial composition in saliva for both cultivable and non-cultivable microorganisms was based on 16S rRNA gene profiles obtained from gradient gels as previously described (Li, et al., 2005, Li, et al., 2006, Li, et al., 2007), using DGGE (Bio-Rad Dcode System; Hercules, CA, USA). A 40% to 60% linear DNA denaturing gradient, where 100% denaturant is equivalent to 7 mol/l urea and 40% deionized formamide formed in 8% (w/v) polyacrylamide gels, was used to separate amplicons, and electrophoresis was performed at constant 60 V, 58°C for 16 h in Tris-acetate-EDTA (TAE) buffer, pH 8.5. After electrophoresis, the gels were rinsed in H₂O and stained in ethidium bromide (0.5 μg/ml), followed by 10 min of de-staining in water. The DGGE profile images were digitally captured (AlphaImager™ 3300 System, Alpha Innotech Corporation, San Leandro, CA, USA), and analyzed by means of Fingerprinting II Informatix Software (Bio-Rad). The similarity coefficient (C_s) between fingerprinting profiles of paired samples was calculated according to the Dice coefficient of pairwise comparisons (Fromin, et al., 2002).

Assay for proteomic analysis

Saliva samples treated with or without protease inhibitor cocktail were analyzed by a combination of 1D SDS-PAGE and LC-MS/MS analysis. Two commercially available inhibitor cocktails (Halt™, Thermo Scientific, and SIGMAFAST™ Protease Inhibitor, Sigma-Aldrich) were used. Briefly, 1 ml of saliva sample with or without PI was concentrated with a 10 kDa membrane cut-off filter (Millipore). The fractions with high-molecular–weight proteins were separated on 1D SDS-PAGE (Novex Bis-Tris 4–12% gel, Invitrogen) and stained with Coomassie Blue. The protein gel bands were excised from the 1D SDS-PAGE and subjected to in-gel reduction, alkylation and trypsin digestion. The digestion was performed for 16 h at 37 °C. The generated peptides were extracted with 50% acetonitrile, washed with a solution containing 0.1% TFA twice and dried with a Speed-Vac. The dried peptide mixture was subjected to LC-MS/MS analysis. For LC-MS/MS analysis, the peptide mixture was separated by a 60 min gradient elution with the Dionex U3000 capillary/nano-HPLC system (Dionex, Sunnyvale, California) at a flow rate of 0.25 μL/min directly interfaced with a Thermo-Fisher LTQ-Orbitrap mass spectrometer (Thermo- Fisher, San Jose, California) operated in data-dependent scan mode. The analytical column was a homemade fused silica capillary column (75 μm ID, 100 mm length; Upchurch, Oak Harbor, Washington) packed with C-18 resin (300 A, 5 μm, Varian, Palo Alto, California). Mobile phase A consisted of 0.1% formic acid, and mobile phase B consisted of 100% acetonitrile and 0.1% formic acid. The 60 min gradients at 0.250 μL/min flow rate for solvent B went from 0 to 55% in 30 min and then to 80% in 10 min. The experiment consisted of a single full-scan mass spectrum in the Orbitrap (400–1600 m/z, 30,000 resolutions), which was followed by 6 data-dependent MS/MS scans in the ion trap at 35% normalized collision energy. Data were analyzed by MASCOT software and manual inspection. The fraction with low-molecular–weight species was directly analyzed by LC-MS/MS using the method described above with a LTQ-Orbitrap mass spectrometer (Thermo-Fisher).

Statistical analyses

Data management and analyses were performed using SPSS 17.0 software (SPSS Inc., Chicago, IL). All cultivable bacterial data were compiled and logarithmically transformed to normalize the variance distribution. Correlation analyses were performed to determine the correlation coefficients of the mean bacterial levels in the samples with and without PI addition. For DGGE profile analysis, levels of similarity between fingerprints were calculated according to the Dice coefficient. Dendrograms were constructed from the average matrix using the unweighted pair group method by means of arithmetic averages (UPGMA). Differences in mean bacterial counts (log₁₀ value), the number of detected DGGE bands, and the degree of similarity were evaluated by using the paired t-test. All p values <0.05 were two-tailed and considered significant.

Results

Based on conventional culturing techniques, the log₁₀ values of the total cultivable bacteria in saliva with PI were similar to those of saliva without PI. Similar results were found for mutans streptococci, oral lactobacilli, and S. mutans (Table 1). Spearman's correlation coefficients (r²) obtained from the paired samples with or without PI demonstrated a high degree of correlation in the mean CFU counts (Fig. 1 A–D).

Table 1.

Comparison of the mean level of bacterial colonization (log₁₀ value of CFU/ml) in saliva samples with and without addition of protease inhibitors (N=22 pairs)

Microorganism	PIs group (mean ± SD)	No-PIs group (mean ± SD)	P value*	r2	P value†
Total cultivable	7.88 ± 0.48	7.89 ± 0.45	0.869	0.847	< 0.001
Mutans streptococci	7.52 ± 0.63	7.47 ± 0.53	0.656	0.898	< 0.001
Oral lactobacilli	5.06 ± 1.92	5.20 ± 1.46	0.944	0.933	< 0.001
S. mutans	5.49 ± 1.24	5.42 ± 1.12	1.000	0.870	< 0.001

^*A two-tailed significance level of paired t-test for two independent means.

^†A two-tailed significance level of nonparametric Spearman's rho correlation coefficient.

Fig 1.

Comparison of cultivable bacterial counts with or without addition of protease inhibitors. The correlation of bacterial growth (measured by CFU counts) on various media in the presence or absence of protease inhibitor cocktail was significant between the paired samples. (A) The Spearman's rho correlation coefficient was 0.847 for total cultivable microbiota (p <0.001); (B) 0.898 for total mutans streptococci (p <0.001); (C) 0.933 for total oral lactobacilli (p <0.001), and (D) 0.870 for total S. mutans (p <0.001). In each graph, the red diagonal line () is the linear regression line from the equation; the solid diagonal line () represents the line of identity where all data points would fall assuming no effect of the protease.

PCR amplification of 16S rRNA gene fragments of 300 bp from 22-paired saliva and 22-paired ETSA plates were profiled by DGGE. The banding patterns were first normalized and then compared between the two groups (with or without PI), based on the position and intensity of each detected band. No difference between the two groups was observed in the numbers of detected DGGE bands (Table 2), or in the total DGGE profiles, for either the saliva samples (Fig. 2A) or the total cultivable samples from ETSA plates (Fig. 2B). The dendrograms clearly demonstrated that all 22 pairs were placed in the same branch. The mean similarity coefficient (C_s) between the paired samples was 97.4% (range from 92.7% to 100%) for the saliva samples and 95.8% (range from 85.7% to 100%) for the total cultivable samples.

Table 2.

Comparison of total bacterial composition profiles in saliva with and without addition of protease inhibitors

N = 22 pairs	Number of bands detected*	Similarity coefficient (Cs)
	(mean ± SD)	(mean ± SD)	Range
Whole saliva sample (cultivable + non-cultivable)
Plus PI	34.0 ± 4.8	97.4% ± 2.2%	92.7% ~ 100%
Minus PI	33.0 ± 4.7
ETSA sample (total cultivable)
Plus PI	23.6 ± 5.8	95.8% ± 4.2%	85.7% ~ 100%
Minus PI	24.0 ± 5.6

^*Paired t-test, both p values greater than 0.5.

Fig 2.

Fingerprint of bacterial composition. Comparison of the DGGE banding patterns of (A) the saliva samples and (B) the total cultivable samples from ETSA plates, with or without protease inhibitors. The dendrograms clearly showed a high degree of similarity between the 22 paired samples.

To determine the effects of PI on the integrity of saliva proteins, the saliva samples treated with and without PI were analyzed by 1D SDS-PAGE and LC-MS/MS (Figure 3). No significant differences were observed among the protein bands between the treated and untreated samples. Using a combination of in-gel digestion and LC-MS/MS analysis, we identified approximately 600 proteins with high confidence for each gel lane. The spectra counts of the major saliva proteins do not show any changes larger than 2-fold, indicating that the inclusion of PI did not have a significant impact on the integrity or stability of salivary proteins. To investigate any effects of the inhibitors on peptidase activity, we analyzed the low-molecular-weight species in the saliva. The molecular ions of the low-molecular-weight species were detected (Figure 4). We found the major ions to be identical for both treated and untreated saliva samples. By a database search, it was observed that of the ions detected in the LC-MS/MS analysis are fragments of proline-rich proteins.

Fig 3.

1D SDS-PAGE image of saliva samples treated with or without protease inhibitor. No significant differences were observed in the high-molecular-weight protein patterns in saliva between the untreated samples (lane 1) and those treated with Halt™ (lane 2) and SIGMAFAST™ (lane 3) protease inhibitors cocktail. “M” stands for the molecular weight markers.

Fig 4.

The mass spectra of low-molecular-weight ions detected by LC-MS/MS analysis. (a) Untreated saliva and (b) SIGMAFAST™ Protease Cocktail-treated saliva sample. The major ions are fragments of proline-rich proteins as identified by database search.

Discussion

Proteases play important roles in a multitude of physiological reactions and biological functions of most microorganisms. Intracellularly, they maintain whole protein homeostasis by 1) controlling the degradation of proteins, which are involved in cell cycle and bacterial development, and 2) responding properly to such environmental changes as stress (Gottesman, 1996, Prepiak & Dubnau, 2007). Extracellularly, a direct relationship with the inactivation of foreign proteins and the destruction of connective tissue components has been reported (Supuran, et al., 2002). Protease inhibitors can alter cell regulation, differentiation, and physiologic functions of microorganisms (Travis & Potempa, 2000), and they have been used as antibacterial agents. For example, anti-leukoprotease (ALP), an endogenous inhibitor of serine proteases produced by serous cells in the submucosal glands, was shown to display antibacterial activity against Escherichia coli and Staphylococcus aureus in vitro (Hiemstra, et al., 1996). Also, cystatins, a superfamily of cysteine PI in human saliva, is known to interfere with the growth of oral bacteria such as Porphyromonas gingivalis (Blankenvoorde, et al., 1998). Synthetic PIs have been developed against a number of proteolytic enzymes as potential antibiotics to retard the growth and proliferation of bacterial pathogens and viruses. Based on human cysteine protease inhibitors, Lars Björek and others synthesized a peptide derivative known as Z-LVG-CtlN2 and showed its specific inhibitory effect on the growth of group A Streptococci strains, both in vivo and in vitro (Bjorck, et al., 1989). Aprotinin was found to have antibacterial activity by its ability to permeate the cell walls of Gram-positive and Gram-negative bacteria and disintegrate the cytoplasm (Pellegrini, et al., 1992). Andréa Loped et al. pointed out the direct correlation between the action of aprotinin and inhibiting growth of the Gram-positive bacterium Streptomyces alboniger (Lopes, et al., 1999). The growth of P. gingivalis and Fusobacterium nucleatum was specifically inhibited by amino-protease inhibitors, such as bestatin (Rogers, et al., 1998, Grenier, et al., 2001).

Because protease inhibitors are widely added during standard purification procedures for proteomic studies, we examined whether such protocol might ultimately affect our ability to qualitatively and quantitatively evaluate the growth and diversity of oral bacteria. To address this question, we used a commercially available PI cocktail which consists of the serine protease inhibitors AEBSF and aprotinin, the cysteine protease inhibitor E-64, the serine and cysteine protease inhibitor leupeptin, the amino-protease inhibitor bestatin, and the aspartic acid protease inhibitor pepstatin A. Based on previously published studies, we hypothesized that the cocktail would affect total cultivable bacterial growth and, therefore, would interfere with the evaluation of bacterial composition in whole oral saliva samples. Unexpectedly, however, the results of our study showed that neither oral bacterial counts nor DGGE profiles of the bacterial composition with protease inhibitors displayed significant statistical differences when compared to the non-protease inhibitor group, suggesting that the addition of PI in saliva samples has no effect on either the growth or composition of oral microbiota. Pellegrini et al. tested the antibacterial properties of a wide variety of protease inhibitors and found that most did not inhibit the growth of bacteria (Pellegrini, et al., 1990). Their observation suggested that the concentration of aprotinin and bestatin might dictate selective antibacterial activities. For instance, bestatin was only found to completely affect the proliferation of Porphyromonas gingivalis in the oral cavity at a concentration of 2.5 μg/ml (Grenier & Michaud, 1994). Such selective PI bacterial killing property would therefore not cause a significant change in overall bacterial diversity profiles. It is also plausible that the concentration of protease inhibitors in a given cocktail may be sufficient to prevent protein degradation, but insufficient to inhibit or kill bacteria in the samples. Accordingly, several investigators have reported that a concentration-dependent relationship exits between protease inhibitor and bacterial growth (Labbe, et al., 2001). Grenier's group showed that there was no inhibition of bacteria with bestatin at 0.02 μg/ml, but when the concentration of bestatin approached 10 μg/ml, the bactericidal function reached a maximum (Grenier, et al., 2001) In the presence of aprotinin, the growth of S. alboniger was also partially or completely inhibited, depending on the concentration of the protease inhibitor (Lopes, et al., 1999). Clearly, then, a higher dose of protease inhibitor has the potential to interfere with the proliferation of bacteria, resulting in an alteration of bacterial composition.

Because massive degradation of proteins caused by proteases has been observed in proteomic studies, it was suggested that PI should be added in the preparation of samples. Here, we have presented results indicating that saliva samples with and without PI showed similar protein diversity in fractions both with high-molecular-weight proteins and low-molecular-weight species as judged by both 1D SDS PAGE and LC-MS/MS analysis. Addition of protease inhibitors seemed to have no significant effect on the integrity of salivary samples. Alternatively keeping the samples on ice and processing them in less than I hr may have been sufficient to preserve protein integrity.

In summary, our study lends considerable evidence that a protease cocktail containing AEBSF, aprotinin, bestatin, E64, leupeptin, and pepstatin A has no effect on oral bacterial growth or total bacterial composition. These findings suggest that the addition of protease inhibitors in the preparation of saliva samples for protein research will not interfere with microbial DNA analysis.

Abbreviations

PI

Protease Inhibitors