Evaluation of diffusion and dilution methods to determine the antimicrobial activity of water-soluble chitosan derivatives
Beilei Ge, Division of Animal and Food Microbiology, Office of Research, Center for Veterinary Medicine, US Food and Drug Administration, 8401 Muirkirk Road, Laurel, MD 20708, USA. E-mail: email@example.com
Chitosan has gained wide applications in the food industry and biomedical field owing to its biodegradability, biocompatibility, nontoxicity and its antimicrobial activity against a wide spectrum of micro-organisms. However, the methods used to investigate antimicrobial effects of chitosan vary considerably among studies, making comparisons difficult.
Methods and Results
One diffusion (disc diffusion) and two dilution (agar dilution and broth microdilution) methods commonly used in clinical laboratories to assess microbial susceptibility/resistance to antimicrobial agents were comparatively used to determine the antimicrobial activity of two water-soluble chitosan derivatives (molecular weights of 43 and 67 kDa) against 31 representative foodborne pathogens. When tested at 1·6% for the 43-kDa chitosan and 3·2% for the 67-kDa chitosan, by disc diffusion, approximately 10- to 11-mm-diameter inhibition zones were observed for all of the bacterial groups, except for Salmonella tested for the 67-kDa chitosan where no inhibition zone was observed. By agar dilution and broth microdilution, the minimal inhibitory concentration (MIC) values varied largely dependent upon the molecular weight of chitosan, bacterial genus/species and the testing method. The agreement between MIC values obtained by the two methods was poor, with broth microdilution generally having lower MIC values than agar dilution. Regardless of the testing method, Salmonella strains were the least susceptible among Gram-negative strains for both chitosans, followed by Escherichia coli and Vibrio.
Besides chitosan's molecular weight and bacterial genus/species, the antimicrobial activity of chitosan was also influenced largely by the susceptibility testing method used.
Significance and Impact of the Study
This is the first study that comparatively evaluated these diffusion and dilution methods, particularly two quantitative methods (agar dilution and broth microdilution), to assess the antimicrobial activity of two water-soluble chitosans against a large number of foodborne pathogens. The study highlights the need for standardized methods to be used in evaluating chitosan's antimicrobial properties in future studies.
Chitosan is a collective name for a group of polysaccharide biopolymers obtained by deacetylation of chitin to various degrees (No and Meyers 1995). Major sources of chitin include exoskeletons of crustaceans (crab, shrimp, lobster, crawfish, etc.) and cell walls of fungi and insects (Tharanathan and Kittur 2003). As a natural compound with desirable functional properties (e.g. biodegradability, biocompatibility and nontoxicity) (Aranaz et al. 2009), chitosan has gained wide applications over the past two decades in food, biomedical, textile and other industries (Tharanathan and Kittur 2003; No et al. 2007; Kong et al. 2010). Examples of chitosan's applications in food include water and fat uptake, emulsification, dye binding, gelation among others (Shahidi et al. 1999; No et al. 2007).
Recently, chitosan and its derivatives have attracted great attention as antimicrobial agents against a wide spectrum of micro-organisms, including bacteria, fungi and viruses (No et al. 2002; Seo et al. 2008; Kong et al. 2010). Although the exact mode of action remains to be fully elucidated, the polycationic nature of chitosan, its chelating capacity and the binding/interaction with cell membrane structures are regarded as important factors resulting in a sequence of events that eventually lead to bacterial death (Raafat et al. 2008; Kong et al. 2010). Simultaneously, commercial applications exploring the antimicrobial activity of chitosan have been developed such as food preservatives, wound dressings and antimicrobial textiles (Tharanathan and Kittur 2003; No et al. 2007; Kong et al. 2010).
Many factors may affect the antimicrobial activity of chitosan, such as intrinsic factors of chitosan (molecular weight, degree of deacetylation, solubility, etc.), microbial factors (bacterial genus, species, age, etc.) and extrinsic factors (pH, temperature, etc.) (Kong et al. 2010). Nonetheless, the methods selected to test for the antimicrobial activity of chitosan may also influence the results obtained. Currently, agar-based diffusion assays (e.g. disc diffusion) are commonly used to assess antimicrobial activity of chitosan and other natural compounds (No et al. 2002; Seo et al. 2008; Klancnik et al. 2010). However, this method is not quantitative, that is, no minimal inhibitory concentration (MIC, the lowest concentration of the compound that completely inhibits bacterial growth) data are generated. Besides, there are a large number of testing variables and different interpretive criteria used in those studies, making comparisons difficult.
In clinical laboratories, the MIC determination of antimicrobial agents is routinely conducted by in vitro susceptibility testing using dilution methods (agar dilution or broth microdilution) standardized by the Clinical and Laboratory Standards Institute (CLSI) (Clinical and Laboratory Standards Institute 2008a, 2006,b). Unlike disc diffusion, these standardized methods have not been widely adopted to test for the antimicrobial activity of chitosan or other natural compounds, although various formats of broth-based, to a lesser extent agar-based, co-incubation methods have been used (No et al. 2002; Burt 2004; Klancnik et al. 2010; Oshima et al. 2012). Also, to our knowledge, performances of these methods have not been compared side by side to determine the antimicrobial effect of chitosan using a large number of bacterial strains of food safety concerns.
The objectives of this study were twofold: (i) to comparatively evaluate the performance of these diffusion and dilution methods (disc diffusion, agar dilution and broth microdilution) as standardized by CLSI for determining the antimicrobial activity of two water-soluble chitosan derivatives (molecular weights of 43 and 67 kDa) and (ii) to obtain the susceptibility profiles of 31 representative foodborne pathogens against the two compounds.
Materials and methods
Bacterial strains and culture conditions
A total of 31 bacterial strains (26 Gram-negative and 5 Gram-positive; Table 1) were used in this study. The Gram-negative strains tested included Escherichia coli, Salmonella and Vibrio, whereas the Gram-positive strains were Enterococcus faecalis, Listeria monocytogenes and Staphylococcus aureus. The strains were cultured at 35°C overnight on trypticase soy agar (TSA; BD Diagnostic Systems, Sparks, MD, USA) except for Vibrio strains, for which TSA supplemented with 2% NaCl was used.
Table 1. Bacterial strains used in this study and their susceptibility profiles for two water-soluble chitosan derivatives as determined by disc diffusion, agar dilution and broth microdilution
|Escherichia coli (n = 9)||P132|| ||10·84||12·12||0·06||0·06||0·003||0·006|
|ATCC 25922|| ||10·61||10·44||0·06||0·125||0·0125||0·1|
|ATCC 35218|| ||10·73||10·73||0·125||0·125||0·025||0·2|
|Salmonella (n = 5)||H9812||Braenderup||10·2||0||0·25||>1||0·2||0·8|
| Vibrio cholerae ||ATCC 14035||O1||10·86||11·21||0·125||0·125||0·025||0·0125|
| Vibrio fluvialis ||ATCC 33809|| ||10·50||11·31||0·06||0·06||0·003||0·006|
|Vibrio harveyi (n = 2)||BB120|| ||10·13||10·13||0·06||0·125||0·003||0·0125|
| Vibrio mimicus ||ATCC 33653|| ||10·73||9·78||0·125||0·125||0·025||0·0125|
|Vibrio parahaemolyticus (n = 3)||ATCC 33847|| ||11·64||11·48||0·06||0·125||0·006||0·006|
|Vibrio vulnificus (n = 4)||ATCC 27562|| ||11·43||10·92||0·06||0·125||0·006||0·006|
|ATCC 33815|| ||12·01||14·24||0·06||0·06||0·003||0·006|
|Enterococcus faecalis (n = 2)||ATCC 29212|| ||10·68||11·03||0·25||0·25||0·2||0·025|
|ATCC 19433|| ||11·38||11·78||0·25||0·125||0·025||0·006|
|Listeria monocytogenes (n = 2)||ATCC 19112||2||11·03||11·92||0·03||0·06||0·2||0·025|
| Staphylococcus aureus ||ATCC 29213|| ||11·20||11·48||0·125||0·125||0·025||0·006|
Two water-soluble chitosan derivatives (referred to as chitosans thereafter; Mw = 43 and 67 kDa with degree of deacetylation values of 76·59 ± 0·42 and 61·43 ± 0·45, respectively; Keumho Chemical, Seoul, Republic of Korea) produced from crab shells were used in this study. The chitosans were placed in separate vials and dried in an oven under 60°C for 1 h before use. Stock solutions (10%, w/v) were made by dissolving dried chitosans in water and filter-sterilized through a 0·2-μm filter (BD Diagnostic Systems). The pH value of the stock solutions was adjusted to 5·9 with 1 mol l−1 HCl (No et al. 2002). The stock solutions were kept at 4°C until use.
Assays to determine the antimicrobial activity of chitosans
Disc diffusion, agar dilution and broth microdilution were performed following CLSI guidelines for antimicrobial preparation, inoculum preparation, inoculation, incubation and data interpretation (Clinical and Laboratory Standards Institute 2008a, 2006,b). All experiments were conducted in triplicate. Briefly, bacterial suspensions were prepared by suspending 3–5 well-separated overnight colonies from TSA plates into 3 ml of cation-adjusted Mueller–Hinton broth (CAMHB, pH adjusted to 5·9; BD Diagnostic Systems), and the turbidity was adjusted to equivalent to a 0·5 McFarland standard. The suspensions were used directly as the final inocula for the disc diffusion and agar dilution methods. For broth microdilution, the suspensions were further diluted 1 : 100 in CAMHB (pH 5·9) before inoculation. Based on preliminary testing (data not shown), two concentrations of chitosans (1·6 and 3·2% for the 43-kDa and 67-kDa chitosans, respectively) were chosen for disc diffusion. The chitosan's test ranges for agar dilution and broth microdilution were 0·03–1% and 0·003–1·6%, respectively.
For disc diffusion, Mueller–Hinton agar plates (MHA, pH adjusted to 5·9; BD Diagnostic Systems) were inoculated with the bacterial suspensions described above by swabbing evenly in three directions to form a lawn. Two sterile 6-mm paper discs (BD Diagnostic Systems) each impregnated with 20 μl of respective chitosan solutions (1·6 and 3·2% for the 43-kDa and 67-kDa chitosans, respectively) were placed on the surface of each inoculated MHA plate using a pair of sterile forceps. A control disc without chitosans was placed similarly on the plate. The plates were incubated at 37°C for 24 h, and the diameters of inhibition zones were measured using a calliper.
For agar dilution, MHA agar plates incorporating twofold serial dilutions of chitosan solutions (0·03–1%) were prepared and inoculated with approximately 0·15 μl of bacterial suspensions described above using a Cathra replicator with 1-mm pins (Oxoid, Lenexa, KS, USA) following CLSI recommendations (Clinical and Laboratory Standards Institute 2006). The plates were incubated under the same condition as for disc diffusion. The MICs of chitosans were recorded as the lowest concentration of chitosans that completely inhibited bacterial growth on the agar plates.
For broth microdilution, 96-well microtitre plates containing CAMHB (pH 5·9) with twofold dilutions of chitosan solutions (0·003–1·6%) were prepared and inoculated with 50 μl of the final inocula described above following CLSI recommendations (Clinical and Laboratory Standards Institute 2006) with an eight-channel pipette. The plates were sealed using a perforated plate seal (TREK Diagnostic Systems Inc., Cleveland, OH, USA) and incubated similarly as described above. The MICs of chitosans were recorded as the lowest concentration of chitosans where no visible growth was observed in the wells of the microtiter plates.
Means and standard deviations of inhibition zone diameters obtained by disc diffusion based on triplicate experiments were calculated by Microsoft Excel (Microsoft, Seattle, WA, USA). The inhibition zone diameters sorted by bacterial genus and chitosan tested were compared using the analysis of variance followed by post hoc comparisons using the least significant difference (LSD) test (SAS for Windows, ver. 9·2; SAS Institute Inc., Cary, NC, USA). Differences between the mean values were considered significant when the P value was <0·05.
For agar dilution and broth microdilution, the MIC50 values were calculated to be the MIC values where 50% of the bacterial population was inhibited, that is, median MIC (Schwarz et al. 2010a,b). MIC agreement between agar dilution and broth microdilution was defined as the differences between MIC values to be within two log2 dilutions (Ge et al. 2002; Oncul et al. 2003; Halbert et al. 2005; Valdivieso-Garcia et al. 2009). Off-scale MIC results, that is, out of the chitosan's test ranges, obtained from any of the two methods were excluded from the agreement calculation.
Antimicrobial activity of chitosan determined by disc diffusion
Table 1 presents the inhibition zone diameters determined by disc diffusion for each bacterial strain, whereas Table 2 summarizes the data by bacterial group. On average, for the 43-kDa and 67-kDa chitosans tested at a concentration of 1·6 and 3·2%, respectively, approximately 10- to 11-mm-diameter inhibition zones were observed for all of the bacterial genera tested, except for Salmonella tested for the 67-kDa chitosan where no inhibition zone was observed (Table 2). In addition, the average inhibition zone size of Salmonella tested for the 43-kDa chitosan was significantly smaller (P < 0·05) than that of Vibrio spp. (Table 2).
Table 2. Antimicrobial activity parameters of two water-soluble chitosan derivatives against each bacterial group as determined by disc diffusion, agar dilution and broth microdilution
| E. coli ||9||10·74 ± 0·29AB||10·74 ± 0·68A||0·125||0·125||0·0125||0·2|
| Salmonella ||5||9·82 ± 1·19B||0||0·25||>1||0·2||1·6|
| Vibrio ||12||11·09 ± 0·59A||11·34 ± 1·09A||0·06||0·125||0·003||0·006|
|G+ bacteria||5||10·81 ± 0·64AB||11·45 ± 0·41A||0·125||0·125||0·2||0·025|
MIC values determined by agar dilution and broth microdilution
The MIC values obtained by agar dilution and broth microdilution for each bacterial strain are shown in Table 1, whereas Table 2 summarizes the MIC50 data by bacterial group. Similar to results obtained using disc diffusion, the MIC values of chitosans also differed with the bacterial group and chitosan tested, and the testing method used. Among Gram-negative genera tested, Salmonella remained the most resistant, followed by E. coli and Vibrio spp., which was most susceptible (Table 2). For Gram-positive bacteria, the MIC50 values were the same as those for E. coli when tested for both chitosans by agar dilution. However, by broth microdilution, the 43-kDa chitosan appeared to be much less effective against Gram-positive bacteria than the 67-kDa chitosan, with MIC50 8-fold higher (Table 2).
When examining the MIC values at the individual strain level, for E. coli, most of the differences in MIC values between the two chitosans were observed only using the broth microdilution method, with the 43-kDa chitosan being more effective by up to 8-fold lower MICs for five generic E. coli strains and 64- to 256-fold lower MICs for four E. coli O157:H7 strains (Table 1). For Salmonella, clearly the 43-kDa chitosan was more effective against all of the five strains tested, regardless of the testing method (Table 1). In contrast, for Vibrio strains, the differences in MIC values between the two chitosans were minimal using either testing method (Table 1). Interestingly, for the five Gram-positive strains tested, similar to E. coli, the majority of MIC variations between the two chitosans were observed only by the broth microdilution method. However, contrary to E. coli results, the chitosan with a stronger antimicrobial effect against the Gram-positive strains was the 67-kDa one with MIC values 4- to 8-fold lower than those of the 43-kDa chitosan (Table 1). This finding does agree with results obtained by the disc diffusion method, where a larger inhibition zone for Gram-positive bacteria was observed for the 67-kDa chitosan (Table 2).
Agreement between MIC values obtained by agar dilution and broth microdilution
Regardless of chitosans tested, MIC values generated by broth microdilution were generally smaller (by up to 20-fold) than those by the agar dilution method (Tables 1 and 2). To quantitatively compare the MIC agreement between the two methods, log2-transformed MIC data were compared (Table 3). The overall MIC agreements between the two methods were 22·6 and 15·4% for the 43-kDa and 67-kDa chitosans, respectively, suggesting poor agreement. When examined by bacterial group, complete mismatches were observed for E. coli tested for the 43-kDa chitosan and Vibrio for both chitosans, although 100% match was observed for Salmonella tested for the 43-kDa chitosan (Table 3).
Table 3. Agreements of MIC values for two water-soluble chitosans obtained by agar dilution and broth microdilution for different bacterial groups
| E. coli ||9||0||20|
| Salmonella ||5||100||Excluded for comparison|
| Vibrio ||12||0||0|
In this study, one diffusion (disc diffusion) and two dilution (agar dilution and broth microdilution) methods commonly used in clinical laboratories to assess microbial susceptibility/resistance to antimicrobial agents were performed side by side to evaluate the antimicrobial activity of two water-soluble chitosan derivatives (molecular weights of 43 and 67 kDa) against 31 foodborne pathogens. The findings indicate that the antimicrobial activity of chitosan is dependent upon chitosan's molecular weight, bacterial genus/species tested and the testing method. By the disc diffusion method, the 43-kDa chitosan resulted in bigger inhibition zones for Salmonella than the 67-kDa chitosan where no inhibition zone was observed. This lack of antimicrobial activity of the 67-kDa chitosan against Salmonella is surprising as previously chitosan with an even higher molecular weight (150 kDa) was shown to be bactericidal when co-incubated with two Salmonella typhimurium strains in chitosan oil-in-water emulsions (Zivanovic et al. 2004). Other factors, such as degree of deacetylation, chitosan concentration, solvents used to dissolve chitosan, the bacterial strains and the level of inoculum tested may have played a role. By broth microdilution, the 43-kDa chitosan showed stronger activities than the 67-kDa chitosan against Gram-negative bacteria but not Gram-positive ones. Regardless of the testing method, Salmonella was the least susceptible Gram-negative genus for both chitosans, followed by E. coli and Vibrio. Additionally, broth microdilution generally had lower MIC values than agar dilution, and the two methods agreed poorly on the MIC values obtained. To our knowledge, this is the first study that comparatively evaluated these three methods, particularly the inclusion of two quantitative methods (agar dilution and broth microdilution), to assess the antimicrobial activity of two water-soluble chitosans against a large selection of bacteria of food safety concerns.
Previously, numerous studies (Papineau et al. 1991; Sudharshan et al. 1992; Wang 1992; Darmadji and Izumimoto 1994; Simpson et al. 1997; Chen et al. 1998; Tsai and Su 1999; Jeon et al. 2001; Liu et al. 2001, 2006; No et al. 2002; Devlieghere et al. 2004; Tsai et al. 2004; Wang et al. 2004; Zivanovic et al. 2004; Chhabra et al. 2006) have demonstrated the antimicrobial activity of chitosan and its derivatives against a wide range of foodborne pathogens using either broth- or agar-based methods. Jeon et al. (2001) and No et al. (2002) reported that chitosan at a concentration of 0·1% (w/v) had stronger antimicrobial effects against Gram-positive bacteria than Gram-negative, although conflicting results were reported in other studies (Devlieghere et al. 2004; Ganan et al. 2009). In the present study, no clear trend between Gram-positive and Gram-negative bacteria was found. Such discrepancy may be partly explained by the different bacterial genera and numbers of strains tested from those genera as well as the different chitosans tested (Devlieghere et al. 2004). Salmonella was found to be the least susceptible Gram-negative genus for both chitosans, followed by E. coli and Vibrio, corroborating findings reported previously using broth-based methods (Chen et al. 1998; Chhabra et al. 2006). However, in the present study, the MIC values of Salmonella for the two chitosans (0·2 and 1·6% by broth microdilution) were higher than those (between 0·01 and 0·1%) reported previously by either broth- or agar-based methods (No et al. 2002; Tsai et al. 2004; Wang et al. 2004; Zivanovic et al. 2004). For E. coli, some studies (Wang 1992; Darmadji and Izumimoto 1994; Simpson et al. 1997; No et al. 2002) reported that chitosans at concentrations ranging from 0·0075 to 1% completely inhibited E. coli growth, while others (Papineau et al. 1991; Sudharshan et al. 1992; Chen et al. 1998) found complete inhibitions with 200 ppm (i.e. 0·02%) of chitosan in broth. Here, the MIC ranges for E. coli were 0·06–0·125% and 0·003–1·6% by agar dilution and broth microdilution, respectively, overlapping with the range reported previously. Similarly, earlier studies examining the effect of chitosan and its derivatives on Vibrio spp. reported an MIC value of 0·01–0·1% using broth- or agar-based tests (Chen et al. 1998; No et al. 2002; Chhabra et al. 2006), overlapping with the ranges (0·06–0·125% for agar dilution and 0·003–0·025% for broth microdilution) observed in the present study.
Among intrinsic factors of chitosan, molecular weight plays a critical role in determining its antimicrobial activity (Jeon et al. 2001; Liu et al. 2001; No et al. 2002; Zivanovic et al. 2004; Seo et al. 2008). Stronger activities were observed among chitosans (Mw of 28–1671 kDa) than chitosan oligomers (Mw of 1–22 kDa) (No et al. 2002), and a minimum of 10 kDa was required to have antimicrobial activities (Jeon et al. 2001). Liu et al. (2001) reported that the antimicrobial activity of chitosan increased with Mw increase from 5 kDa to 91·6 kDa but then decreased when the Mw further increased to 1080 kDa. Additionally, Seo et al. (2008) reported that the most effective molecular weight of chitosan varied with micro-organisms tested. The two chitosans tested in the present study have Mw of 43 and 67 kDa. By disc diffusion, the 43-kDa chitosan appeared more effective against all bacterial genera than the 67-kDa chitosan except for E. coli. While by broth microdilution, it showed stronger activities against Gram-negative bacteria but not Gram-positive ones. By agar dilution, no discernible differences were observed between the two chitosans for all genera tested except for Salmonella (Table 1).
Although great attention has been paid recently to utilize the antimicrobial effects of natural compounds such as chitosans in food, biomedical and textile industry, there are currently no standardized methods to assess their antimicrobial activities (Liu et al. 2007; King et al. 2008; Klancnik et al. 2010). Often times, a large number of testing variables were introduced such as inoculum size and growth media, rendering the comparison between studies difficult. Disc diffusion method is attractive because of its ease of performance and low cost; however, the inhibition zone size is largely determined by how well the compounds uniformly diffuse into the agar media, which may be problematic for many natural compounds that are large and insoluble in water (Klancnik et al. 2010). Also, a linear or logarithmic relationship between the inhibition zone size and the concentration of compounds impregnated on the disc usually does not exist, making it only a qualitative method (King et al. 2008). On the other hand, MIC determination by dilution methods (agar dilution and broth microdilution) is quantitative, sensitive, reproducible and can be easily standardized. Agreeable with a previous report (Klancnik et al. 2010), we also found broth microdilution generated much lower MICs than agar dilution. Additionally, agar dilution method failed to detect any differences between the two chitosans against most bacterial genera tested. This is probably due to that chitosans are more mobile in broth solution than in agar mixture. Therefore, we support that broth microdilution could be used as a screening method for preliminary MIC determination, which is to be followed by macrodilution method at selected MIC concentrations to confirm bacterial inactivation (Klancnik et al. 2010). Nonetheless, standardized methods are desired to ensure a more comparable evaluation of chitosan's antimicrobial properties in future studies.
Regardless of the testing method used, both chitosan derivatives appeared to be more effective against Vibrio species than other bacterial genera tested. As Vibrio parahaemolyticus and Vibrio vulnificus continue to cause seafood safety concerns particularly in oysters that are often consumed raw (Han et al. 2007), and given chitosan's other desirable functional properties (e.g. biodegradability, biocompatibility and nontoxicity), the antimicrobial property of chitosan may be applied in oyster products to ensure safer products for consumers.
This study was supported in part by an Aquaculture Special Grant from the US Department of Agriculture.