In Vitro Antibacterial Activity of Propyl-Propane-Thiosulfinate and Propyl-Propane-Thiosulfonate Derived from Allium spp. against Gram-Negative and Gram-Positive Multidrug-Resistant Bacteria Isolated from Human Samples

Background The aim of this study was to compare the in vitro antibacterial activity of two compounds derived from Alliaceae, PTS (propyl-propane-thiosulfinate), and PTSO (propyl-propane-thiosulfonate), with that of other antibiotics commonly used against bacteria isolated from humans. Materials and Methods A total of 212 gram-negative bacilli and 267 gram-positive cocci isolated from human clinical samples and resistant to at least one group of antibiotics were selected. In order to determine the minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) to various antibiotics as well as PTS and PTSO, all isolates underwent broth microdilution assay. Results PTS showed moderate activity against Enterobacteriaceae with MIC50 (and MBC50) and MIC90 (and MBC90) values of 256-512 mg/L, while PTSO showed greater activity with MIC50 and MIC90 values of 64-128 mg/L and MBC50 and MBC90 values of 128-512 mg/L. These data show the bactericidal activity of both compounds and indicate that PTSO was more active than PTS against this group of bacteria. Both compounds showed lower activity against P. aeruginosa (MIC50 = 1024 mg/L, MIC90 = 2048 mg/L, MBC50 = 2048 mg/L, and MBC90 = 2048 mg/L, for PTS; MIC50 = 512 mg/L, MIC90 = 1024 mg/L, MBC50 = 512 mg/L, and MBC90 = 2048 mg/L, for PTSO) compared to those obtained in others nonfermenting gram-negative bacilli (MIC50 = 128 mg/L, MIC90 = 512 mg/L, MBC50 = 128 mg/L, and MBC90 = 512 mg/L, for PTS; MIC50 = 64 mg/L, MIC90 = 256 mg/L, MBC50 = 64 mg/L, and MBC90 = 256 mg/L, for PTSO) and also indicate the bactericidal activity of both compounds against these groups of bacteria. Finally, the activity against S. aureus, E. faecalis, and S. agalactiae was higher than that observed against enterobacteria, especially in the case of PTSO (MIC50 = 8 mg/L, MIC90 = 8 mg/L, MBC50 = 32 mg/L, and MBC90 = 64 mg/L, in S. aureus; MIC50 = 4 mg/L, MIC90 = 8 mg/L, MBC50 = 8 mg/L, and MBC90 = 16 mg/L, in E. faecalis and S. agalactiae). Conclusion PTS and PTSO have a significant broad spectrum antibacterial activity against multiresistant bacteria isolated from human clinical samples. Preliminary results in present work provide basic and useful information for development and potential use of these compounds in the treatment of human infections.


Introduction
The use of conventional antibiotics for the prevention of infectious diseases and as growth promoters in animal production has fostered the appearance of resistant bacteria and the transmission of these pathogens to humans [1]. In addition, the use and sometimes misuse of antibiotics in humans has increased the occurrence of infections (urinary tract infections, respiratory tract infections, skin and soft tissue infections, etc.) caused by multiresistant bacteria, which has reduced the therapeutic options and has made necessary the selection of new molecules with antibacterial properties [2]. Natural compounds obtained from vegetables with antibacterial properties could be considered an alternative to conventional antibiotics [3].
In recent years, the antibacterial properties of some compounds obtained from Allium plants such as garlic (Allium sativum) and onion (Allium cepa) have been described. These can inhibit the growth of a range of gram-positive and gram-negative bacteria, including both pathogenic and commensal bacteria in humans and animals [4,5]. Alliumderived products have been reported to be effective even against those strains that have become resistant to antibiotics [6].
Two of these Allium-derived compounds, propylpropane-thiosulfinate (PTS) ( Figure 1) and propyl-propanethiosulfonate (PTSO) (Figure 2), are organosulphurate products obtained by decomposition of initial compounds naturally present in garlic bulbs as alliin and allicin. In several in vitro and in vivo studies against pathogenic bacteria from animals, both compounds have showed an antibiotic activity [3,7,8]. While the precise mechanism of action is not yet known, the main antibacterial effect of thiosulfinates (as allicin) has been reported to be due to (i) its accessibility resulting from high permeability through phospholipid membranes [9]; (ii) its chemical reaction with thiol groups of various enzymes such as the bacterial acetyl-CoA-forming system, consisting of acetate kinase and phosphotransacetyl-CoA synthetase, blocking acetate incorporation into fatty acids and inhibiting the formation of lipids [10]; and (iii) the inhibition of RNA polymerase and RNA synthesis [11].
Therefore, the aim of this study was to compare the in vitro antibacterial activity of the compounds derived from garlic PTS and PTSO with that of other antibiotics commonly used against gram-negative and gram-positive multidrugresistant bacteria isolated from human clinical samples.

Antibiotics, PTS and PTSO.
All antibiotics were purchased from Sigma-Aldrich (Madrid, Spain) and each antibiotic was dissolved according to the manufacturer's recommendations.
PTS and PTSO (95% purity) were supplied by DMC Research (Alhendín, Granada, Spain) and dissolved in polysorbate-80 to a final concentration of 50%. The biosynthesis of propyl-propane-thiosulfinate (PTS) and propylpropane-thiosulfonate (PTSO) is made from propiin, an amino acid derived from L-cysteine found in Allium species. The first step of the biosynthesis is the formation of a sulfenic acid, which is highly reactive and immediately produces PTS by a condensation reaction. In the last step, oxidation of PTS induces its dismutation in PTSO and propyl disulfide that can be oxidized and transformed to PTSO and that way the oxidation of PTS to PTSO is completed.

Bacterial Isolates.
A total of 212 gram-negative bacilli and 267 gram-positive cocci isolated from clinical samples obtained from 479 different patients were selected. Identification and susceptibility studies were performed using WIDER system (Francisco Soria Melguizo, Madrid, Spain) or MicroScan system (Siemens Healthcare Diagnostics, Madrid, Spain). The susceptibility results obtained through these systems allowed the selection of isolates, based on the resistance presence to at least one group of antibiotics commonly used in the treatment of infections caused by these bacteria.
The presence of extended-spectrum beta-lactamaseproducing Enterobacteriaceae (ESBL) was confirmed by the diffusion method with disks containing cefotaxime (30 g), cefotaxime/clavulanic acid (30/10 g), ceftazidime (30 g), and ceftazidime/clavulanic acid (30/10 g). The resistance to methicillin was confirmed using the Mueller-Hinton agar diffusion procedure with 30 g cefoxitin disks. Both procedures were performed as recommended by the Clinical and Laboratory Standards Institute [12]. A

In Vitro Antibacterial Assay.
In order to determine the antibacterial susceptibilities, all 479 isolates underwent broth microdilution assay in Cation-Adjusted Mueller-Hinton Broth (CAMHB) following the guidelines of the CLSI [12]. Broth microdilution testing was performed with 96-well, round-bottom microtiter plates with a final concentration of the bacterial cell suspension equal to 1 x 10 5 colony forming units per milliliter (CFU/ml) in each well.
Each plate included negative controls (medium only) and 11 serial twofold dilutions of each antibiotic, PTS, or PTSO. The positive controls (only bacterial suspension without antibiotics) were added per well in a separate round-bottom plate.
The minimum inhibitory concentration (MIC) was defined as the lowest antibiotic concentration to completely inhibit the visible growth of a microorganism after overnight incubation and the isolates were considered to be susceptible, intermediate, or resistant, according to the recommendations of the CLSI [12]. A "susceptible" result indicates that the patient's organism should respond to therapy with that antibiotic using the dosage recommended normally for that type of infection and species. Conversely, a microorganism with a MIC interpreted as "resistant" should not be inhibited by the concentrations of the antibiotic achieved with the dosages normally used with that drug. An "intermediate" result indicates that a microorganism falls into a range of susceptibility in which the MIC approaches or exceeds the level of antibiotic that can ordinarily be achieved and for which clinical response is likely to be less than with a susceptible strain. MIC 50 and MIC 90 values were defined as the lowest concentration of the antibiotic at which 50 and 90% of the isolates were inhibited, respectively.
For minimum bactericidal concentration (MBC) testing, 100 l of broth from 1 to 5 wells containing no growth (which showed no visible turbidity) was plated onto antibioticfree Columbia agar and incubated overnight at 37 ∘ C. The highest dilution that yielded no single bacterial colony on the agar plates was taken as MBC. Allium extracts were then considered as bacteriostatic or bactericidal depending on the MBC/MIC ratio which were, respectively, greater than 2 or between 2 and 1. MBC 50 and MBC 90 values were defined as the concentration of the antibiotic which kills 50 and 90% of the isolates, respectively.
Following the CLSI guidelines, we used the following strains as quality control in the procedures: E. coli ATCC 25922, P. aeruginosa ATCC 27853, S. aureus ATCC 29213, and E. faecalis ATCC 29212.

Statistical
Analysis. Data analysis was performed using the software IBM SPSS Statistics v19. The Mann-Whitney U test was used to compare the distribution of MIC and MBC values of PTS and PTSO in the different groups of bacteria studied. A level of significance was considered with a p< 0.05. Tables 1 and 2 show the values (in mg/L) of the MIC 50 , MIC 90 , MBC 50 , and MBC 90 and percentages of resistance to the antibacterial agents tested of the 479 clinical isolates.

Results
There was 59 ESBL-producing Enterobacteriaceae (42 E. coli, 12 K. pneumoniae, and 5 K. oxytoca). The presence of this resistance phenotype in 39.1% of Enterobacteriaceae was the main determinant of the high rates of resistance to beta-lactam antibiotics, whose range oscillated from 1.   As previously mentioned, bacteria were selected for their detection of resistance to, at least, a group of antibiotics. However, a relevant characteristic of the 151 enterobacteria included in the study was the high frequency to coresistance to two or more of these groups (multidrug-resistant bacteria), as described in Table 3. Therefore, 74.0% of the isolates resistant to some beta-lactams antibiotics were also resistant to ciprofloxacin, 61.8% to trimethoprim-sulfamethoxazole, and 48.8% to some aminoglycoside. It should be noted that 22.8% of that resistant to beta-lactams was also resistant to all the other groups of antibiotics assayed.
The behaviour of PTS and PTSO against multidrugresistant enterobacteria was quite homogeneous, regardless the group analyzed (  --  Table 3, 75.0% of the isolates resistant to fluoroquinolones (ciprofloxacin) were also resistant to some beta-lactam antibiotic; 63.9% to some aminoglycoside and 55.6% showed resistance to these three groups of antibiotics.

Resistance to beta-lactams & aminoglycosides
Just as with the rest of antibiotics, when comparing the results obtained in P. aeruginosa with those obtained in others nonfermenting gram-negative bacilli, the behaviour, both of PTS and PTSO, was significantly different (Table 1 In any case, these data also indicate the bactericidal activity of both compounds, especially PTSO that showed significantly more activity than PTS (p < 0.001).
Concerning the gram-positive cocci, all the isolates were susceptible to vancomycin, teicoplanin (S. agalactiae was not tested), daptomycin, and linezolid. Besides, all the isolates of E. faecalis and S. agalactiae were also susceptible to ampicillin ( Table 2).
PTSO showed significantly more activity than PTS in the three groups of gram-positive bacteria tested (p < 0.001, in all cases) and the values for MIC 50 , MIC 90 , MBC 50 , and MBC 90 were, for both compounds, lower than those obtained against gram-negative bacteria (Table 2). However, MIC and MBC values in gram-positive bacteria differed significantly, especially for PTS (more than 2 dilutions), which indicates that these compounds could have a bacteriostatic but not a bactericidal effect against these bacteria at least at low concentrations.

Discussion
Organosulfur compounds obtained from Allium spp. such as PTS and PTSO have been proposed as an effective alternative to antibiotics to improve animal performance and prevent gastrointestinal disorders. This is due on the one hand to their greater stability in comparison to other natural compounds [13] and on the other hand to their activity against bacterial groups, such as Enterobacteriaceae, Staphylococcus spp., Enterococcus spp., Clostridium spp., Bacteroides spp., Lactobacillus spp., Bifidobacterium spp., or Campylobacter spp., among others [3,4,7]. Furthermore, it has been shown that feed supplementation with these compounds improves the digestion and absorption of nutrients in the gastrointestinal tract by modulating the intestinal microbiota and increases the villus height and mucosal thickness [7,8]. Beyond its use in animals, it is possible that these molecules may as well be useful in the human clinical practice, due to the fact that alliaceous plants have been traditionally used for their antibacterial, antioxidant, and cardiovascular properties, as has been known for centuries [6].
To our knowledge, this is the first study to evaluate the activity of PTS and PTSO against a selection of gramnegative and gram-positive multiresistant bacteria isolated from human clinical samples. Antibiotic susceptibility tests were performed in accordance with the procedure outlined by CLSI in order to determine if a bacterium is susceptible or resistant to each of the antibiotic assayed. Although the cutoff points for PTS or PTSO are unknown, perform the assay under the same conditions as the other antibiotics allow us to make comparisons with them.
Our results revealed that PTS showed moderate activity against Enterobacteriaceae with MIC 50 (and MBC 50 ) and MIC 90 (and MBC 90 ) values of 256-512 mg/L, while PTSO showed greater activity with MIC 50 and MIC 90 values of 64-128 mg/L and MBC 50 and MBC 90 values of 128-512 mg/L. These homogeneous results among the different groups of enterobacteria selected, regardless of the resistance shown to different antibiotics commonly used in clinical practice, reveal the bactericidal action of these compounds. According to these results, Ruiz et al. also proved a bactericidal effect against enterobacteria, such as E. coli and Salmonella typhimurium [3].
The activity against methicillin-resistant S. aureus, E. faecalis, and S. agalactiae was higher than that observed against enterobacteria, especially in the case of PTSO ( . It is possible that PTS and PTSO may be affected by active removal mechanisms when they come in contact with these bacteria. Further research is needed to determine with certainty the mechanisms involved in this increased resistance.
All these results are in agreement with the antibacterial effects of garlic previously described in the literature against bacterial isolates from animals and reference strains [3][4][5][6]. However, MBC determined in our experiment were much higher compared to Llana-Ruiz-Cabello et al. who demonstrated MBC lower than 5 mg/L in all cases [16]. The differences may be caused by different methodology.
In the present study, the values obtained for MIC and MBC in PTS and PTSO were very similar to those obtained in antibiotics such as nitrofurantoin, aminoglycosides, fluoroquinolones, and some beta-lactams. Based on the data obtained from MIC, the CLSI determines that a very large percentage of enterobacteria should be resistant to these antibiotics (as shown in Tables 1 and 2). It should therefore not be considered for clinical use. Likewise, we may think that the activity shown by PTS and PTSO should also not be considered for clinical use in humans considering the results obtained. However, due to the lack of susceptibility cut-off points for the compounds derived from garlic, no final conclusion can be drawn.
In correspondence with the need of discovering new potentially antibacterial natural products, the activity of these organosulfur compounds described in this study may be considered as promising. Furthermore, the use of naturally and potentially innocuous compounds that can be administered without high restrictions provided us with the possibility to discuss the viability of their application for the treatment of specific infectious pathologies, provided that adequate formulations are developed.
In our opinion, several therapeutic possibilities may be considered, i.e., superficial skin infections, such as acne, folliculitis or impetigo by topical use, the treatment of oral and gastrointestinal infections by oral administration, or even the treatment of urinary tract infections caused by multidrugresistant bacteria applied by intravesical instillation (in the same way that colistin is used). The concentration of the substance in the source of the infection should always be high enough to guarantee that it exceeds the values of MIC against the bacteria causing these processes.
It is clear that, in order to evaluate the real effectiveness of these substances, either in this or another situation, further testing would be necessary with a more diverse and larger group of bacteria. Furthermore, it would be necessary to establish suitable administration routes for the compounds and its efficacy in vivo. Finally, the concentrations that they achieve in the different tissues and fluids would also need to be known.
Lastly, PTS and PTSO are perceived as harmless since these compounds occur naturally in foods such as garlic or onion. Nevertheless, further studies on pharmacokinetic and toxicological characteristics are required before safe clinical use is considered. Some recent studies on cell lines and experimental animals reported low acute and subchronic oral toxicity in PTSO and a lack of genotoxicity, both in vitro and in vivo models [16][17][18][19].

Conclusion
Our results demonstrate that PTS, but mainly PTSO, have a significant broad spectrum antibacterial activity against a selection of gram-negative and gram-positive multiresistant bacteria isolated from human clinical samples. Further work is needed to demonstrate the effectiveness of these compounds in vivo models, although preliminary results in present work provide basic and useful information for development and its potential use in the treatment of human infections.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.