Synthesis and Antibacterial Evaluation of Novel Water-Soluble Organic Peroxides

Abstract

A set of new water-soluble organic peroxides has been synthesized and evaluated for in vitro antibacterial activity as part of an effort to develop new antibacterial agents for the treatment of acne vulgaris. The water solubility of these new dialkyl peroxides and peroxyesters was achieved by incorporating either a quaternary ammonium group or a polyethylene glycol moiety. These peroxides are effective against both gram-positive and gram-negative bacteria and have a prolonged activity compared to that of benzoyl peroxide and other peroxide-type antiseptic agents. Among them 4-[[(tert-butylperoxy)carbonyl]benzyl]triethylammonium chloride and [10-(tert-butylperoxy)decyl]trimethylammonium bromide have the broadest antimicrobial spectrums. We have shown that the oxidizing properties of the dioxy group of these compounds are responsible for their antibacterial activities.

The current treatment of acne vulgaris is accomplished primarily with antibiotics or with the concomitant use of antibiotics and benzoyl peroxide (BP). The long-term use of systemic and topical antibiotic therapy, including erythromycin, clindamycin, minocycline, and doxycycline, has been shown to be associated with an increase in the populations of highly resistant Propionibacterium acnes and coagulase-negative staphylococci, predominantly Staphylococcus epidermidis, on the skin surface (7, 18, 23, 31). The ability of coagulase-negative staphylococci to transfer resistance via plasmids to the more pathogenic Staphylococcus aureus has been demonstrated previously (27). Treatment with BP prevents the multiplication of such resistant strains (8, 31) and suppresses the growth of other susceptible acne-causing bacteria (8, 9, 16, 25, 26, 28, 29, 33).

P. acnes thrives in an oxygen-free environment and feeds on trapped oil in the sebaceous follicle. It then releases toxic, corrosive fatty acid by-products that break down the follicle walls and encourage the cellular disruptions that produce acne lesions (21, 22). BP appears to cause a significant decrease in the free fatty acids (14, 17, 18), and this reduction is used as a measure of the success of antibacterial therapy for acne vulgaris (10, 11, 30). Local application of BP also produces a keratolytic effect that causes lysis of the intercellular substance in the stratum corneum (10). In addition, BP decreases the production of sebum (10). However, successful therapy with BP is considered to be primarily a function of its oxidizing properties (8, 9, 16, 25, 26, 28, 33). BP exerts an oxidizing power and produces cell death via the interaction of oxidized intermediates with various constituents of microbial cells (6, 8).

The instability and poor solubility of BP in water limit its efficacy (9). The topical application of BP is often given in the form of an oil-in-water cream or a gel vehicle. To provide more peroxide candidates for the treatment of acnes vulgaris, a set of novel water-soluble organic peroxides has been synthesized and evaluated for in vitro antibacterial activity in comparison to that of BP and of another peroxide-type antiseptic agent, hydrogen peroxide. Four gram-positive and two gram-negative organisms were selected for the antimicrobial susceptibility tests. Among the tested organisms, the staphylococci have been found on the skin of acne patients (5, 7, 8, 18, 23, 27, 31). The selected Bacillus, Escherichia, and Pseudomonas strains which were used are commonly used for assays of antimicrobial agents (24, 32). The design and synthesis of water-soluble organic peroxides as well as the results of the antimicrobial susceptibility tests on a set of aerobic bacteria strains are presented in this paper. The antimicrobial properties of these peroxides on anaerobic P. acnes are under investigation and will be reported in due course.

MATERIALS AND METHODS

Eicosaethylene glycol isohexadecyl ether was purchased from ICI Americas, Inc. All other reagents and solvents were obtained from Aldrich. They were of the highest available purities and were used without further purification. Both ¹H nuclear magnetic resonance (NMR) and ¹³C NMR analyses were run on a Bruker AC-250 spectrophotometer. Fourier transform-infrared spectra were obtained on a Mattson Galaxy 2020 instrument.

Dialkyl peroxides and alkyl peroxyesters are known to be more stable than peroxyacids and alkyl hydroperoxides, and they are less susceptible to radical- or metal-induced decomposition (1). Therefore, the newly synthesized peroxides were designed to be dialkyl peroxides (Fig. 1, compounds 2 and 5) and peroxyesters (Fig. 1, compounds 1, 3, and 4). The water solubility of these peroxides was achieved by incorporating either a quaternary ammonium group (compounds 1, 2, and 3) or a polyethylene glycol moiety (compounds 4 and 5). The quaternary ammonium group employed in compounds 1, 2, and 3 was used because it has a widely recognized antibacterial property (13). To explore a new mode of interaction between the antibacterial agent and the cell membrane, compounds 4 and 5 were each designed to have a long alkyl chain.

FIG. 1.

Structures of water-soluble organic peroxide compounds 1, 2, 3, 4, and 5 and of their control compounds 1a, 2a, 3a, and 4a.

The syntheses of 4-[[(tert-butylperoxy)carbonyl]benzyl]triethylammonium chloride (compound 1), [10-(tert-butylperoxy)decyl]trimethylammonium bromide (compound 2), and [4-(4′-tert-butylperoxycarbonylbenzoyl)benzyl]trimethylammonium chloride (compound 3) have been reported elsewhere (19, 20). Isohexadecyl eicosaethylene glycol tert-butylperoxy phthalate (compound 4) and eicosaethylene glycol tert-butylperoxy isohexadecyl ether (compound 5) were prepared by the method outlined in Fig. 2.

FIG. 2.

Syntheses of isohexadecyl eicosaethylene glycol tert-butylperoxy phthalate and eicosaethylene glycol tert-butylperoxy isohexadecyl ether.

Both cationic (compounds 1, 2, and 3) and nonionic (compounds 4 and 5) peroxides are more water soluble than BP (0.005% [wt/vol]) (6). The solubility of compound 1 was more than 100% (wt/vol), while the solubilities of compounds 2 and 3 were more than 5% (wt/vol). The solubilities of compounds 4 and 5 were more than 20% (wt/vol). To compare the in vitro activities of these compounds with those of BP and H₂O₂, we used both dimethyl sulfoxide (DMSO) and water as carriers in the experiments described below. For the comparisons between the peroxides (compounds 1, 2, 3, 4, and 5) and their controls (compounds 1a, 2a, 3a, and 4a [compound 4a was used as a control for compounds 4 and 5]), solutions were prepared by dissolving these compounds into sterile water to the required concentrations before use.

Synthesis of isohexadecyl eicosaethylene glycol tert-butylperoxy phthalate.

Sublimated phthalic anhydride (0.85 g, 5.7 mmol) was dissolved in 25 ml of acetonitrile. Sodium t-butyl peroxide (0.65 g, 5.8 mmol) was gradually added to the solution. The reaction mixture was stirred at room temperature for 2 h. Ethyl chloroformate (0.71 g, 6.5 mmol) was added dropwise to the mixture, and 1 small drop of pyridine was added afterwards. The mixture was stirred for another 4 h at room temperature. Meanwhile, eicosaethylene glycol isohexadecyl ether (5.9 g, 5.2 mmol) was mixed with sodium hydride (0.22 g, 5.5 mmol) in 30 ml of acetonitrile, and the mixture was stirred at room temperature for 40 min. After evaporation of the solvent from the reaction mixture, freshly prepared sodium alkoxide solution was gradually added to the mixture in an ice bath. The reaction mixture was stirred at room temperature for 3 h and concentrated. The residue was chromatographed on a silica gel column with acetone as eluent to obtain 3.5 g (59%) of compound 4. Found: infrared 1772, 1729 cm⁻¹; ¹H NMR (CDCl₃) δ 7.85 (1H, br), δ 7.60 (2H, br), δ 7.35 (1H, br), δ 4.47 (2H, t), δ 3.56 to 3.90 (78H, m), δ 3.31 (2H, t), δ 1.56 (2H, br), δ 1.39 (9H, s), δ 1.26 (23H, br), δ 0.88 (6H, t); ¹³C NMR (CDCl₃) δ 165.73 (s), δ 165.55 (s), δ 128.48 to 131.1 (m), δ 83.84 (s), δ 61.33 to 74.43 (m), δ 22.4 to 37.83 (m), 13.92 (s).

Synthesis of isohexadecyl eicosaethylene glycol p-toluenesulfonate.

To 200 ml of benzene solution containing eicosaethylene glycol isohexadecyl ether (22.4 g, 0.02 mol), 60% sodium hydride (0.8 g, 0.02 mol) was added at room temperature. After stirring for 40 min, p-toluenesulfonal chloride (4.6 g, 0.02 mol) was gradually added to the mixture. The reaction mixture was stirred for 24 h at room temperature. After evaporation, the crude product was chromatographed on a silica gel column with acetone as an eluent to yield 13.5 g (53%) of p-toluenesulfonate ester (compound 6). Found: ¹H NMR (CDCl₃) δ 7.7 (2H, d), δ 7.35 (2H, d), δ 4.13 (2H, t), δ 3.30 to 3.90 (78H, m), δ 3.10 (2H, br), δ 2.44 (3H, s), δ 1.50 (2H, br), δ 1.26 (23H, br), δ 0.85 (6H, t); ¹³C NMR (CDCl₃) δ 144.06 (s), δ 132.58 (s), δ 129.24 (s), δ127.31 (s), δ 73.94 (s), δ 67.99 to 72.07 (m), δ 60.85 (s), δ 37.49 (s), δ 22.04 to 31.27 (m), δ 13.56 (s).

Synthesis of eicosaethylene glycol tert-butylperoxy isohexadecyl ether.

To 5 ml of a dimethylene chloride solution of p-toluenesulfonate ester (compound 6), t-BuOO⁻Na⁺ (1.2 g, 0.02 mol) and 3 ml of decane solution containing t-BuOOH (5 to 6 M) were added. The reaction mixture was stirred at room temperature for 36 h. After evaporation, the residue was chromatographed with acetone as eluent to obtain 3.8 g (81%) of the peroxide product compound 5. Found: ¹H NMR (CDCl₃) δ 4.07 (2H, t), δ 3.55 to 3.70 (78H, m), δ 3.3 (2H, t), δ 1.50 (2H, t), δ 1.26 (23H, br), δ 1.24 (9H, s), δ 0.88 (6H, t); ¹³C NMR (CDCl₃) δ 79.67 (s), δ 68.07 to 74.20 (m), δ 37.68 (s), δ 22.27 to 31.48 (m), δ 13.75 (s).

Bacteria.

Bacillus cereus ATCC 11778, Bacillus subtilis ATCC 6633, Staphylococcus epidermidis ATCC 12228, Staphylococcus aureus ATCC 25923, and Pseudomonas aeruginosa ATCC 27853 were purchased from the American Type Culture Collection. Escherichia coli DH5α was obtained from Life Technologies (Gaithersburg, Md.). The antibacterial activities of both cationic (compounds 1, 2, and 3) and nonionic (compounds 4 and 5) peroxides were assessed by a standard disk diffusion procedure (2, 3). Bacteria were grown on Mueller-Hinton agar and Luria-Bertani (LB) agar. The sizes of the inhibition zones on both media were found to be similar. The B. cereus, B. subtilis and E. coli inocula for the disk diffusion tests were derived from cultures that had been aerobically grown for 6 h in LB medium (Difco) at 37°C. S. aureus and P. aeruginosa inocula were obtained from cultures aerobically incubated for 11 h at 37°C in Nutrient broth (Difco), and S. epidermidis inocula were obtained from cultures aerobically incubated for 11 h at 37°C in Trypticase Soy broth (BBL).

Width of the inhibition zone around the disk.

The agar disk diffusion tests were performed on LB agar plates. A total of 50 ml of agar medium was poured into a petri dish (150 by 15 mm) to a depth of 4 mm. Plates were dried at 35 to 36°C for about 30 min in an incubator before inoculation. One freshly grown colony of the bacterial strain was inoculated into 2 ml of LB medium. The culture was then incubated as described above. The density of the inoculum was determined by standard bacterial plate counting. The inoculum (200 μl) was applied to each plate and uniformly spread with a glass spreader over the surface. After the inoculum dried, autoclaved filter paper disks (250 mm; Fisher) were placed on the agar, and 60 μl of a known concentration of each compound was loaded onto the disks. To compare the inhibition zone diameters for compounds 1, 2, and 4 with that for BP, all chemicals were dissolved in DMSO. For the comparisons between compounds 1, 2, 3, and 5 and H₂O₂, the chemicals were dissolved in sterile water. The plates were inverted and allowed to incubate at 35 to 36°C. The inhibition zone around the disk (χ) was calculated by subtracting the disk size (in millimeters) from the zone diameter and then dividing by two.

Critical concentration.

The critical concentration (m′) was determined experimentally by testing a bacterial isolate against six concentrations of a compound and plotting the natural logarithm of the compound concentration against the square of χ as described by Barry (2). Four well-isolated colonies of the same morphological type were selected and transferred to a tube containing 4 ml of a suitable broth medium. The broth culture of individual bacteria was incubated under the conditions mentioned above. The turbidity of the actively growing broth culture was adjusted with sterile saline to a density equivalent to that of a 0.5 McFarland standard by measuring the absorbance at 650 nm. Within 15 min of adjusting the turbidity of each inoculum suspension, the inocula were applied to agar plates. Sterile blank test disks (BBL) were placed on the inoculated agar surfaces. A solution (20 μl) of each experimental compound was loaded onto the disks. Compounds 1, 2, 3, 4, and 5 were dissolved in sterile water to the required concentrations. BP was dissolved in DMSO. The plates were inverted and incubated aerobically for 18 h at 36 to 37°C. The sizes of the inhibition zones were measured and used to calculate the m′s.

RESULTS AND DISCUSSION

The m′s of each compound for the tested bacteria, which concentration are the amounts that are just capable of inhibiting microbial growth under the test conditions (2), are presented in Table 1. All peroxides were effective against B. cereus and B. subtilis. Among the peroxides studied, compound 3 showed the highest activity against both Bacillus strains. Compound 3 had a structure which was similar to that of compound 1. Incorporation of an additional benzoyl group to compound 1 enhanced the level of activity. Compounds 1 and 2 were effective against E. coli. The peroxide compounds 1, 2, and 3 exhibited activities against the Staphylococcus strains. Again, peroxide compound 3 showed the highest antibacterial activity, with an m′s on S. aureus which was superior to those of compounds 1 and 2. Compounds 2 and 4 had activities against P. aeruginosa as well. Since compounds 1 and 2 were found to be active against both the gram-positive and the gram-negative strains tested, they were considered to have a broader spectrum of activity than the other peroxides.

TABLE 1.

Antimicrobial susceptibilities of aerobic strains^a

Bacterial strains	m′ (μg/ml) of peroxide compound:
	1	2	3	4	5
B. cereus ATCC 11778	115	280	70	90	75
B. subtilis ATCC 6633	139	168	77	188	1,300
E. coli DH5α	212	310	>970	>2,600	>3,600
S. epidermidis ATCC12228	237	107	198	>2,600	>3,600
S. aureus ATCC 25923	141	181	10	>2,600	>3,600
P. aeruginosa ATCC 27853	>600	90	>970	68	NTb

^aThe tests to determine the m′ of BP were performed on the bacteria listed. However, there was no linear relationship between the natural logarithm of the concentration and the square of the zone diameter observed. 

^bNT, no test was performed. 

The m′s indicate that all peroxides bearing the positively charged quaternary ammonium group (compounds 1, 2, and 3) were effective against the gram-positive strains tested, such as the bacilli and staphylococci. Considering the positive charges of compounds 1, 2, and 3, these compounds were not expected to diffuse through the cell membrane to a significant extent. Therefore, they may have exerted their antibacterial activities by oxidizing the components of the cell envelope. Further investigation is needed to elucidate the mode of action. We attached a polyethylene glycol isohexadecyl ether moiety to compounds 4 and 5 to enhance this solubilities in water and to explore a new interaction mode between antimicrobial agents and bacteria cell membranes. The long hydrophobic hexadecyl chain was expected to insert into the bilipid layer, thus anchoring the oxidizing dioxy group to cell membranes.

Comparison of the biological activities of the compounds having a tert-butylperoxy group (compounds 1, 2, 3, 4, and 5) with that of the control compounds without the dioxy group (compounds 1a, 2a, 3a, and 4a) (Fig. 1) demonstrated that the powerful oxidizing properties of the peroxides contributed significantly to their antibacterial activities. The control compounds 1a, 2a, and 3a lacked activity due to the absence of the dioxy functional group. Although the antibacterial properties of quaternary ammonium compounds are widely recognized, their activities depend on the length of the aliphatic chain attached to the cationic head. Optimal activity is usually reached with an alkyl chain length of ca. 12 carbons (34). Therefore, it is not surprising to observe that the quaternary ammonium compounds 1a, 2a, and 3a did not show any activity under the test conditions because of the shorter length of the alkyl chain attached to the ammonium group in each of these compounds. Peroxides containing a quaternary ammonium group with a longer alkyl chain could have higher activities than those of the peroxides used in this study.

The control compound 4a, polyethylene glycol isohexadecyl ether, generated an inhibition zone. It had rather short-lived activity, and the inhibition zone disappeared completely after 20 h of incubation. Therefore, the antimicrobial activities of the nonionic peroxides, compounds 4 and 5, are attributed to both the dioxy group and the polyethylene glycol isohexadecyl ether moiety. The polyethylene glycol isohexadecyl ether moiety of compounds 4 and 5 possibly played a causative role in regard to the activities in the early time of incubation. The remaining activities of compounds 4 and 5 after 20 h are only due to the contribution from the dioxy functional group.

The durations of the antibacterial effects of these peroxides were studied and are shown in Fig. 3 and 4 in comparison with those of BP and the common peroxide antiseptic agent H₂O₂. Figure 3 displays the widths of the inhibition zones around the disks (χ) for peroxide compounds 1, 2, and 4 and for BP against B. cereus over the incubation time. Figure 4 shows the χ values of peroxide compounds 1, 2, 3, and 5 and H₂O₂ against B. cereus over the incubation time. Each compound diffuses through the agar medium at a different rate when it is applied to the disk in a different solvent medium. As a result, the observed zone sizes around disks for the same compound differ in Fig. 3 and 4 due to the use of different solvents as carriers. The antibacterial effects of the compounds containing the quaternary ammonium group (compounds 1, 2, and 3) persisted for the entire period of incubation as shown in Fig. 3 and 4. The sizes of the inhibition zones around the disks for compounds 1 and 3 appeared to increase during the incubation because of actual lysis of the initial growth within the inner ring of the zone (4). The activities of the nonionic peroxide compounds 4 and 5 lasted for the entire period of incubation as well. However, their χ values gradually deceased over time, but some residual activity remained until the end of incubation. The decrease of the χ value during the incubation might be due to the dissipation of the antimicrobial activity from the polyethylene glycol isohexadecyl ether moiety. The inhibition zone around the disk for BP became apparent at about 8 h, but its size decreased dramatically over the next 2 h. It dissipated completely by 40 h (Fig. 3). The effect of hydrogen peroxide lasted for only 10 h (Fig. 4).

FIG. 3.

Comparison of the duration of the antibacterial activities of peroxide compounds 1, 2, and 4 with that of BP. All compounds were dissolved in DMSO, and 3.6 μmol of each compound was loaded onto a 2.5-cm-diameter disk. The tested B. cereus inoculum contained 1.1 × 10⁸ CFU/ml. Measurements were first made at 8 h and were made periodically thereafter. Values are means for four independent experiments. Error bars indicate standard deviations. No inhibition zone was observed with DMSO on the disk.

FIG. 4.

Duration of the antibacterial activities of peroxide compounds 1, 2, 3, and 5 and H₂O₂. All compounds were dissolved in sterile water, and 3.6 μmol of each compound was loaded onto a 2.5-cm-diameter disk. The tested B. cereus inoculum contained 1.1 × 10⁸ CFU/ml. Measurements were first made at 8 h and were periodically made thereafter. Values are means for four independent experiments. Error bars indicate standard deviations. No inhibition zone was observed with sterile water on the disk.

These novel water-soluble organic peroxides have long-lasting antimicrobial activities which are superior to those of BP and H₂O₂. These compounds exerted immediate antibacterial effects which were capable of strongly suppressing the growth of a dense population of bacteria. The inhibition effects lasted over the entire period of incubation. Such long-lasting antimicrobial activity was probably due to the thermal stability and hydrophilic property of these peroxide. Peroxide compounds 2 and 5 are dialkyl peroxides, and peroxide compounds 1, 3, and 4 are alkyl peroxyesters. The thermal stabilities of dialkyl peroxides and alkyl peroxyesters are higher than those of diacyl peroxides such as benzoyl peroxide (15). The higher stability of these peroxides prevented the undesirable degradation of the compounds and led to longer-lasting antimicrobial effects. The hydrophilic nature of these compounds permitted them to stay in the aqueous media where they could oxidize various constituents of cell membranes instead of being dissolved in the lipid bilayer like BP (6). The hydrophilicity of the quaternary ammonium group of compounds 1, 2, and 3 is not necessarily an advantage for in vivo anti-acne application due to its limited solubility in the sebaceous follicles. However, the high antimicrobial activities and stability of these compounds could compensate for such a disadvantage. The nonionic peroxide compounds 4 and 5 provide the necessary solubility in the sebaceous follicles due to the possession of a long hexadecyl chain.

The experimental results presented here clearly demonstrate that the newly synthesized organic peroxides do have antibacterial activities and can potentially be used in treatment against P. acnes. Unlike BP, which is lipophilic, the new organic peroxides have a lipophilic dioxy moiety and a hydrophilic functional group. The hydrophilic functional group in these molecules provide them with a possible new mode of action against the targeted bacteria, i.e., oxidizing the cell surface instead of the lipid bilayer. This new mode of action may be responsible for the higher and longer-lasting antibacterial activities. Considering the lipophilic environment of P. acnes in the follicle, oil-compatible and surface-active peroxides such as compounds 4 and 5 may present measurable advantages in acne treatment.