®
Akademik Gıda / Academic Food Journal ISSN Print: 1304-7582 http://www.academicfoodjournal.com
Akademik Gıda 9(5) (2011) 11-20 Research Paper / Araştırma Makalesi
Numerical Taxonomy of Lactic Acid Bacteria Isolated from Fermented Foods Özlem Ertekin1, Ahmet Hilmi Çon2 1 Tunceli University, Engineering Faculty, Food Engineering Department, Tunceli, Turkey Ondokuz Mayıs University, Engineering Faculty, Food Engineering Department, Samsun, Turkey
2
Received (Geliş Tarihi): 03.08.2011, Accepted (Kabul Tarihi): 25.10.2011 Corresponding author (Yazışmalardan Sorumlu Yazar):
[email protected] (A.H. Çon) + 90 362 312 19 19 +90 362 457 60 35
ABSTRACT Lactobacillaceae family members have identification problems depending upon isolated material and traditional carbohydrate fermentation tests used. In this study, it is aimed to overcome these problems by using numerical taxonomy method. Morphological, metabolical and physiological characteristics of bacterial isolates were determined with standard type strains, and they were defined and classified into homogenous groups by using fundamental principles of numerical taxonomy. In this research 32 test isolates and 10 type isolates were used and defined at the 89.4% similarity level according to SSM-UPGMA analysis and classified into twenty clusters being the first cluster with 7 members, the second cluster with 15 members, the third cluster with 3 members and 17 clusters that have only one member. Based on these results, 25 of total 42 isolates (59.52%) had multiple members, 17 of them (40.48%) were grouped into single member clusters. In multi-member clusters, similarity ratio of the first cluster was 89.4%, the second cluster’s was 89.4% and the third cluster’s was 90.7%. Three member clusters’ similarity ratio was calculated 80.0%. After conducting numerical taxonomy study, test error was calculated as 1.11%. Results showed that reliability of this research is considerably higher than previous assumption. Key Words: Lactic acid bacteria, Numerical taxonomy, Dendrogram, Fermented food
Fermente Gıdalardan İzole Edilen Laktik Asit Bakterilerinin Numerik Taksonomisi ÖZET Lactobacillaceae familyası üyelerinin tanımlanması uygulanan karbonhidrat fermantasyon testleri ve izole edildikleri materyallere göre önemli problemlere neden olmaktadır. Bu çalışmada numerik taksonomi uygulaması ile bu problemlerin aşılması hedeflenmiştir. Bakteriyal izolatların morfolojik, metabolik ve fizyolojik karakterleri standart tip suşlar ile birlikte belirlenmiş, nümerik taksonomi esasları çerçevesinde homojen gruplar halinde sınıflandırılarak tanımlanmıştır. Çalışmada 32 test izolatı ve 10 tip izolatı SSM-UPGMA analizine göre %89.4 benzerlik seviyesinde tanımlanmış ve 20 kümeye sınıflandırılmıştır. 7 üyeli 1. küme, 15 üyeli 2. küme, 3 üyeli 3. küme ve 17 adet de tek üyeli kümeye sınıflandırılmıştır. Bu sonuca göre toplam 42 izolatın 25 adedi (% 59.52) çok üyeli, 17 adedi de (% 40.48) tek üyeli kümelere gruplandırılmıştır. Çok üyeli kümelerden 1. kümenin benzerlik oranı %89.4, 2. kümenin %89.4, 3. kümenin %90.7 olarak belirlenmiştir. SSM-UPGMA analiz sonucuna göre tanımlanan 3 adet çoklu kümenin birbiri arasında benzerlik oranı %80 olarak bulunmuştur. Yapılan nümerik taksonomi çalışması sonucunda test hatasının tespit edilmesi için de test varyansı (Si2) hesaplanmış ve test hatası %1.11 bulunmuştur. Bu sonuç çalışmanın güvenirliliğinin önceki kabullere göre oldukça yüksek olduğunu göstermiştir. Anahtar Kelimeler: Laktik asit bakterisi, Nümerik taksonomi, Dendogram, Fermente gıda
11
Ö. Ertekin, A.H. Çon Akademik Gıda 9(5) (2011) 11-20
INTRODUCTION
In a study, 120 Lactobacillus isolates from ogi and 3 traditional starch fermented alcoholic beverages produced in Nigeria, 18 reference strains from sour Sweden dough and 50 types of Lactobacilli strains were classified by using phenotype taxonomy depending upon their fermentation ability for 49 different types of carbohydrate. Isolates were divided into 7 main groups that had 82% similarity ratio. Three of them were similar to L.plantarum and L.plantarum quasi. Others were L.confosus, L.murinis, L.agilis and Leuconostoc mesenteroides subsp. mesenteroides. It was indicated that L.plantarum’s phenotype features changed in wide range of limits [9].
The definition of numerical taxonomy is expressed to apply various mathematical methods for data which are coded numerically and expressed as individual characters and assigning organisms into cluster based on their comprehensive similarities, afterwards presenting relationships in the form of dendrograms. This classification is used by many researchers for bacteria systematic [1-4]. This classification converts information about properties of microorganisms into numerical data, which are more convenient for using numerical analysis methods. Then this data can be compared with the help of a computer. Resulting classification is based on evaluation of general similarities and comparison of many properties that are weighted equally. To get correct and reliable classification, at least 50, ideally 100-200 characters including morphological, biochemical and physiological data have to be compared [4, 5].
Two molecular methods (RAPD and Southern hybridization) were used to study differences of 140 strains that have relative relation with L.plantarum. More than 93% (56/60) of studied strains gave close results after classification depending on results of RAPD and hybridization. When compared to fermentation tests, in group, which was classified as L. plantarum, some differences were found in usage of melesitose, αmethyl-D-mannoside and dulcitol [10].
To get reliable and successful results by using numerical taxonomy, it is very important to choose characters that are not affected from environmental conditions, and they have to represent single gene or operon’s expression. These kinds of characters are stable, so it is easy to get reliable natural classification. In practice, it is necessary to perform a series of tests that represent biological properties of organism. An ideal test list includes colony and micro-morphological data, growth characteristics, biochemical tests, effects of inhibitor agents, single carbon source compounds for energy and growth, serologic, chemotaxonomic and molecular genetic information. The objective is gathering enough information to determine relationships or differences between taxonomic ranks. Due to this reason, it is believed that when test results are totally positive or totally negative, these tests don’t have any value for taxonomic studies as they don’t have distinctive intrataxonomic properties. That’s why last data should not be included into evaluation matrix [6].
Identification of Lactobacillus, which was isolated form healthy horse bowels’ mucous membrane, was made by using API 50CH kit. Total of 51 Lactobacillus sp. were isolated in selective culture environment, and their capability of fermentation with different carbohydrates was characterized. The results were compared with previous results that had been gathered from 200 strains of Lactobacillus fermentation test results, and numerical taxonomy was applied. Isolates were classified into 10 different groups. 4% of isolates could not be identified [11]. This is much higher success ratio than that of obtained from traditional definition. In a study, 400 probable Lactobacillus isolates were obtained from 12 Italian ewe cheeses (6 different kinds) produced by different producers, which were not starter. Phenotype, genetic and cell wall protein characterization analyses were applied to 123 isolates and 10 type strains. Phenotype features of cheese isolates were determined as 32% of L.plantarum, 15% L.brevis, 12% L.paracasei subsp. paracasei, 9% L.curvatus, 6% L.fermentum, 6% L.casei subsp. casei, 5% L.pentosus, 3% L.casei subsp. pseudoplantarum and 1% L.rhamnosus. Phenotype of 11% isolates could not be identified [12].
Applied test method and optimization of this method has importance at numerical taxonomy. Medium formulation, methods and amounts of inoculation, duration and temperature of incubation and growth phase of microorganism (has to be logarithmic phase) have to be standard [6, 7]. Nowadays, it is assumed that numerical taxonomy is one of the most reliable approaches to explain relationship among species, so it can be used for the classification of lactic acid bacteria. Because, it has been characterized by popularity of usage at food fermentation process of lactic acid bacteria, metabolic properties, growing performance, resistance to industrial processing, ability to stay alive in end product, shelf life etc. In addition, technological assumptions, security and quality control improvements also became important criteria [8]. At this point, numerical taxonomy’s importance increased for the classification of lactic acid bacteria. Many taxonomic studies have been conducted by using different features of lactic acid bacteria.
In another study, 4 lactic acid bacteria isolates from industrial fermented milk and 10 dominate from conventionally fermented milk were tested in 32 different characteristic dimensions with 14 references Lactobacillus strains and 3 Lactococcus strains together. All of the isolates obtained from conventionally fermented milks turned out to belong to Lactobacillus species. Among them, other bacteria types such as L.helveticus, L.plantarum, L.delbrueckii subsp. lactis, L.casei subsp. casei, and L.casei subsp. pseudoplantarum and also 3 isolates were identified as beta bacteria or streptobacteria, and 4 isolates were identified as L.lactis [13].
12
Ö. Ertekin, A.H. Çon Akademik Gıda 9(5) (2011) 11-20
obtained distinctive features test results were evaluated [18]. As summarized above, it is possible to find some research related to this subject around globe, and still it is considered as a hot research topic.
In a study, 94 Gram-positive, catalase negative bacteria from pork and chicken meats packaged under modified atmospheric conditions, were isolated. Afterwards, an irradiation dose of 1.75 and 2.5kGy were applied to the samples. As a result of numerical taxonomy group 1 contained 3 L.sake strains with 78 test strains. Group 2 included 3 test strains and type strains of C.piscicola and C.diverge; group 3 contained two chicken strains and L.curvatus. Group 4 contained pork strains and Leuconostoc dextranicum, group 5 and 6 contained 4 pork and 2 chicken strains respectively. On the other hand, 4 test strains were not included in any of the groups indicated above [14].
It is believed that identifying lactic acid bacteria species correctly that are isolated from fermented products would be a guideline to chose correct starter culture for the production of these products. Use of right culture will lead to production of goods that meet quality specifications and improvement of public health and nutrition. In this study, it is aimed to determine and define morphological, metabolical and physiological characteristics of bacterial isolates with standard type strains, and to classify them into homogenous groups by using fundamental principles of numerical taxonomy.
94 units of lactic acid bacteria, which were isolated from packaged frozen meat and meat products, were classified based on 96 phenotype characteristics by using 59 reference strains of Brochothrix, Lactobacillus, Leuconostoc, Pediococcus and Streptococcus. All of the microorganisms were classified into 23 groups that had at least 2 or more members with 84% similarities [15]. 61 lactic acid bacteria isolated from deteriorated vacuum packaged Vienna sausages and 15 reference strains were tested for 72 phenotype properties. Identification diagrams and computer data were used to determine identities, and it was observed that there was (86.9%) high correlation between these two methods on level of genus. This ratio was 54.8% on base of species. 60 units of strains were classified into 6 groups (89% similarity) by numerical taxonomy [13].
MATERIALS and METHODS Materials Isolates (OTU: Operational Taxonomy Unit) used in numerical taxonomy research includes isolated specimens from different food products and internationally accepted type strains and their duplicates. In this study, total of 32 units of OTU isolated from cheese, sausage (sucuk), sour dough and pickle were used. And also, total of 10 type strains collected from international culture collection were used in these tests. To control reliability of tests 7 (16.7% of samples) duplicates were chosen randomly.
Identification and classification of 249 lactic acid bacteria which were isolated from freshwater fish and their environment was conducted. Approximately 90% (226 strains) of isolates were not able to grow up on acetate agar. Furthermore, they were assumed as Carnobacterium based on their homogenous phenotype. 22 strains were Lactobacillus, Enterococcus, Lactococcus or Vagococcus. One isolate was not able to be identified at species level. Results indicated that it was possible to determine distinctions between Carnobacteriums and Lactobacilli successfully by numerical taxonomy, yet in species level, identification of Carnobacterium was not clear enough. 6 clusters similarity ratios were found to be 86% when phenotype properties were used for numerical taxonomy. 13 clusters were created after UPGMA clustering process was applied. Among these 13 clusters, 7 of them included at least 3 or more isolates [16].
Three different indicator bacteria, important in food products, were used to determine anti-microbial activity spectrum. M17 and MRS broth with MRS agar were used to preserve and grow lactic bacteria, for other indicators nutrient broth and nutrient agar were used for same purposes like preserving and growth [17]. The lactic acid bacteria were preserved in skim milk (Oxoid L31) that contains 15% glycerin in frozen form at -20°C [19, 20], it was also kept as submerged culture into MRS agar at +4°C and as lyophilized cultures [17, 20, 21].
Methods Numerical Taxonomy Tests conducted during numerical taxonomy study were grouped under 4 main titles: carbohydrate fermentation, heavy metals resistance, antibiotics sensitivity and biochemical tests. Tests explaining 176 units’ characters of OTU were performed (Table 1). For each of this OTU, test method was optimized and standardized (method of incubation, period, and temperature with inoculation dose and life stage).
It is possible to encounter with important obstacles when lactic acid bacteria are identified by conventional carbohydrate fermentation tests. Finally, in a research by Çon [17], 20 units (20%) of isolates that were used were able to be identified based on their biochemical and physiological features at species level weren’t identified. That’s why, it is necessary to take into consideration not only carbohydrate tests result but also other features when it is needed to identify the relationships among species and under species level. Related to this issue, it was decided to use numerical taxonomy to classify lactic acid bacteria because it was published that it gave stable taxonomic results when all
Statistical Analysis 176 unit character tests applied on OTUs results were saved in X-Taxon program in format of +/-. Positive test results were coded as (+) and negative test results were coded as (-). After converting data to 1/0 format, a statistical packet program (NTSys-pc) was used for
13
Ö. Ertekin, A.H. Çon Akademik Gıda 9(5) (2011) 11-20
indicator microorganisms and performed as described by Schillinger and Lücke [24], Pepe et al. [25] and Dal Bello et al. [26]. Gram staining Şimşek [27] and Kostinek et al. [28], catalase test Kim et al. [29] and Carr et al. [30], gas production from glucose Çon [17] and Kostinek et al. [28], hydrolyze of arginine Şimşek et al. [31], growth in different pH values Papamanoli et al. [32] and G-Allegria et al. [33], growth in different salt concentrations Kask et al. [21] and Randazzo et al. [34], growth in different temperatures Şimşek et al. [31] and Kostinek et al. [28], resistance to alcohol G-Allegria et al. [33], β- Galactosidase test Hébert et al. [35], resistance to bile salt Cebeci and Gürakan [36] and Papamanoli et al. [32] were conducted.
analysis purposes as Wishart described directions. To calculate test error, individual test variance (Si2) between duplicate specimens was calculated, and for each of the test, test error ratio (Pi) was also measured [22]. To calculate similarity of OTUs, SSM (Simple matching coefficients, [23]) coefficient was used, and to explain relative relationship connections AverageLinkage (UPGMA: Unweighted Pair Group Method with Arithmetic Averages [1]) clustering method was chosen.
Microbiological Analysis Biochemical Tests Antimicrobial activity spectrum L.sake Lb790, L.monocytogenes Li1 and E.faecium were used as
Table 1. Used unit characters in numerical taxonomy research A. Carbohydrate Fermentation (used API50CH Strip) Control Inositol Glycerol D-Mannitol Erythritol D-Sorbitol D-Arabinose Methyl-αD-Mannopyranoside L-Arabinose Methyl-αD-Glucopyranoside D-Ribose N-Acetylglucosamine D-Xylose Amygdalin L-Xylose Arbutin D-Adonitol Esculin(ferric citrate) Methyl-βD-Xylopyranoside Salicin D-Galactose D-Cellobiose D-Glucose D-Maltose D-Fructose D-Lactose D-Mannose D-Melibiose L-Sorbose D-Saccharose L-Rhamnose D-Trehalose Dulcitol Inulin B. Resistance to Heavy Metals (10, 25, 50, 150, 300 and 500 ppm) CrN3O9.9H2O CuSO4.5H2O N2NiO6.6H2O MgSO4.7H2O Al(NO3)3.9H2O MnSO4.2H2O ZnSO4.7H2O CoN2O6.6H2O AgNO3 CdN2O6.4H2O C. Sensitivity to Antibiotics Neomycin (30 µg) Streptomycin (10 µg) Polymixin B (300 U) Penicillin G (10 U) Novobiocin (30 µg) Tetracycline (30 µg) Ampicillin (10 µg) Rifampicin (30 µg) Bacitracin (10 U) Erytromycin (15 µg) Kanamycin (30 µg) Chloramphenicol (30 µg) D. Biochemical and Phsylogical Tests D.1. Gas production from glucose D.5. β- Galactosidase D.2. Growth in temperature (°C) D.6. Resistance to alcohol (%) (10, 15, 45) (10.0, 12.0, 15.0) D.3. Resistance to salt (%) D.7. Resistance to ox-bile (3.0, 4.0, 5.0, 6.5, 8.0, 9.0) (3, 5, 9% in agar and 9% in broth) D.4. Growth in different pH (3.0, 3.5, 9.6)
D-Melezitose D-Raffinose Amidon (starch) Glycogen Xylitol Gentiobiose D-Turanose D-Lyxose D-Tagatose D-Fucose L-Fucose D-Arabitol L-Arabitol Potassium Gluconate Potassium2-Ketogluconate Potassium 5- Ketogluconate Pb(NO3)2 Fe2 (SO4)3. H2O FeSO4.7H2O NaNO2 Novobiocin (5 µg) Rifamycin (30 µg) Vancomycin (30 µg) Gentamicin (10 µg)
D.8. Hydrolyze of arginine D.9. Morphology D.10. Antimicrobial activity (Against L.sake Lb790, E. faecium and L.monocytogenes Li1)
Carbohydrate Fermentation Tests
Resistance to Heavy Metals Tests
Ability of isolates to use various carbohydrate sources was determined by using API 50CH test kit (bioMeriux sa 69280 Marcy I’Etoile France) under directions of the producer company [25, 28, 31, 37-39].
Resistance of isolates to heavy metals was performed according to methods by Hassen et al. [42] and Gülcan [43].
RESULTS and DISCUSSION
Sensitivity to Antibiotics Tests
To provide enough reliable numerical taxonomy results, all of the stages such as selection of strains, selection of tests and collection of data and coding of these data
Sensitivity of isolates to antibiotics was performed according to methods by Charteris et al. [40], Kelly et al. [41] with Cebeci and Gürakan [36].
14
Ö. Ertekin, A.H. Çon Akademik Gıda 9(5) (2011) 11-20
were performed precisely with extra attention as described above section namely material and method. In this study main data matrix contained a total of 49 OTUs (32 tests, 10 types and 7 duplicates) with 176 units characters. On the other hand, final data matrix was obtained by subtracting duplicates and also subtracting some test results that had completely positive or negative results, for 42 OTUs, and the remaining 122 unit characters were involved in final matrix.
this research is pretty high when it is compared with previous assumption. Similarity coefficient is used to control whether the data are qualified enough to use in hierarchy clustering at numerical taxonomy [1, 4]. Similarity values between 60%-95% in numerical classification are good indicator of success of taxonomic classification [44]. In this study, according to SSM-UPGMA analysis, 89.4% similarity level was defined in 42 OTUs. And the first cluster included 7 members, second cluster included 15 members, the third cluster included 3 members and 17 clusters have only one member. Depending on the results, clusters were grouped as follows: 25 units (59.52%) of total 42 isolates have multiple members, 17 units have (40.48%) one member. Dendrogram that belongs to these clusters is given in Figure 1, and cluster members and isolation sources are given in Table 2.
Test variance (Si2) between individual duplicates and for each of test, test error ratio (Pi) was calculated [22]. Test error was calculated as 1.11% after studying final data matrix by numerical taxonomy. According to Sneath and Johnson [22] 10% test error can be accepted as normal. Test error found in this study showed that reliability of
Figure 1. Dendrogram shows relationship that was determined by using SSM-UPGMA analysis between type strains and isolates
Cluster 1
cluster 1 members, carbohydrate fermentation tests showed glycerol, erythritol, D-arabinose, L-xylose, Dadonitol, methyl-ßD-xylopyranoside, L-sorbose, Lrhamnose, dulcitol, inositol, D-mannitol, D-sorbitol, methyl-αD-glucopyranoside, D-maltose, D-lactose, D-
This cluster includes 7 isolates and has 89.4% similarity level (Table 3). This high similarity ratio in this cluster indicates that taxonomic classification is acceptable. For
15
Ö. Ertekin, A.H. Çon Akademik Gıda 9(5) (2011) 11-20
melibiose, inulin, D-melezitose, D-raffinose, amidon, glycogen, xylitol, D-turanose, D-lyxose, D-fucose, Lfucose, D-arabitol, L-arabitol, potassium gluconate, potassium 2-ketagluconate, potassium 5-ketogluconate were not used, L-arabinose, D-ribose, D-xylose, Dgalactose, D-glucose, D-fructose, D-mannose, Nacetylglucosamine, amygdaline, arbutin, esculin, salicin, D-celiobiose, D-saccharose, D-trehalose, gentiobiose, D-tagatose were used; from heavy metals CrN3O9.9H2O for 150, 50, 25 and 10 ppm; N2NiO6.6H2O for 150, 50,
25 and 10 ppm; Al(NO3)3.9H2O for 150, 50, 25 and 10 ppm; ZnSO4.7H2O for 50, 25 and 10 ppm; CoN2O6.6H2O for 500, 300, 150, 50, 25 and 10 ppm; Pb(NO3)2 for 500, 300, 150, 50, 25 and 10 ppm; CuSO4.5H2O for 10 ppm; MgSO4.7H2O for 500, 300, 150, 50, 25 and 10 ppm; MnSO4.2H2O for 50, 25 and 10 ppm; Fe2(SO4)3.H2O for 50, 25 and 10 ppm; FeSO4.7H2O for 150, 50, 25 and 10 ppm and NaNO2 and for also 300, 150, 50, 25 and 10 ppm were found resistant to the concentrations.
Table 2. OTUs distribution and sources resulting from SSM-UPGMA analysis based on multi member clusters Cluster No Isolate No Specimen Name Source P.pentosaceus O24 LO160 Sausage (sucuk) P.pentosaceus O25 LO161 Sausage (sucuk) LO162 P.pentosaceus O27 Sausage (sucuk) P.pentosaceus O29 LO163 Sausage (sucuk) *1 P.pentosaceus O31 LO164 Sausage (sucuk) P.pentosaceus O57 LO165 Sausage (sucuk) LO125 Unidentified O28 Sausage (sucuk) L.plantarum O12 LO140 Mixed pickle L.plantarum O13 LO141 Mixed pickle L.plantarum O14 LO142 Sourdough L.plantarum O19 LO143 Cucumber pickle L.plantarum O20 LO144 Cheese L.plantarum O21 LO145 Cheese L.plantarum O23 LO147 Pickle 2 L.plantarum O26 LO148 Sausage (sucuk) LO149 L.plantarum O30 Sausage (sucuk) L.plantarum O33 LO150 Cheese L.plantarum O41 LO151 Cheese L.plantarum O58 LO153 Cheese L.plantarum 1 (DSM20174) LP155 Type strain L. plantarum O22b LO156 Cheese L. plantarum O44b LO157 Sourdough LO180 L.helveticus O16 Sourdough 3 LO121 Unidentified O10 Sourdough LO123 Unidentified O17 Sourdough L.lactis ssp. lactis O3 4 LO100 Cheese L.lactis 7 (DSM20481) 5 LL102 Type strain L.pentosus O8a 6 LO110 Sourdough L.brevis O11 7 LO130 Sourdough L.brevis 2 (DSM20054) 8 LB131 Type strain P.pentosaceus 10 (DSM20336) Type strain 9 PP167 L.curvatus O32 10 LO170 Cheese L.curvatus 3 (DSM20019) 11 LC171 Type strain L.paracasei ssp. paracasei O54 Cheese 12 LO190 L.paracasei 5 (DSM5622) 13 LR192 Type strain L.salivarius 4 (DSM20555) 14 LS201 Type strain L.acidophilus 6 (DSM20079) 15 LA210 Type strain L.mesenteroides 9 (DSM20343) Type strain 16 LM220 P.acidilactici 11 (DSM20284) 17 PA230 Type strain 18 LD120 Unidentified O1 Cheese 19 LO122 Unidentified O15 Sourdough 20 LO124 Unidentified O18 Sourdough For antibiotic sensitivity test, they were not sensitive against bacitracin, vancomycin and gentamicin. On the contrary, they were sensitive against novobiocin (5 and 30µg), tetracycline, rifampicin, erythromycin, chloramphenicol and rifamycin. Biochemical tests such as gas production from glucose and hydrolyze of
arginine were negative; there was no growth at 3.0 pH and 15% alcohol; there was growth in 15°C temperature, 3.0, 4.0, 5.0, 6.5 and 8.0% NaCl, 3.5 and 9.6 pH, 10% alcohol, 3.0%, 5.0% and 9.0% ox-bile added MRS agar and 9.0% ox-bile added MRS broth. It was also found that they showed antimicrobial activity
16
Ö. Ertekin, A.H. Çon Akademik Gıda 9(5) (2011) 11-20
against E.faecium and L.monocytogenes Li1, also they
were β-Galactosidase positive.
Table 3. Isolate numbers and similarity level of clusters that are distributed according to SSM-UPGMA analysis Cluster No Cluster 1 Cluster 2 Cluster 3
No of Specimens 7 15 3
SSM-UPGMA Level (%) 89.4 89.4 90.7
right to say; for L.plantarum carbohydrate tests and numerical definition ended in parallel results.
All food isolates which are defined as P.pentosaceus according to carbohydrate fermentation tests, were included in this cluster. However, P.pentosaceus type strain was not included in this cluster which was not expected (included in cluster 9). P.acidilactici type strain was not included in this cluster, and it was in cluster 17, which has only one member. This was an expected result.
Cluster 3 Among multi member cluster, this cluster has the highest similarity ratio (90.7%). This high similarity ratio shows that in this cluster taxonomic structure is acceptable. Cluster included 3 isolates. According to carbohydrate fermentation tests, glycerol, erythritol, Darabinose, L-arabinose, D-ribose, D-xylose, L-xylose, Dadonitol, methyl-βD-xylopyranoside, L-sorbose, Lrhamnose, dulcitol, inositol, D-mannitol, D-sorbitol, methyl-αD-mannopyranoside, methyl-αDglucopyranoside, amygdaline, arbutin, esculin, salicin, D-celiobiose, D-melibiose, D-saccharose, D-trehalose, inulin, D-melezitose, D-raffinose, amidon, glycogen, xylitol, gentiobiose, D-turanose, D-lyxose, D-tagatose, D-fucose, L-fucose, D-arabitol, L-arabitol, potassium gluconate, potassium 2-ketogluconate, potassium 5ketogluconate were not used by member of cluster 3. It was found that D-galactose, D-glucose, D-fructose, Dmannose, N-acetylglucosamine, D-maltose, D-lactose were used. 5 different concentrations were used for heavy metal resistance tests. Concentrations that all cluster members were resisted CrN3O9.9H2O for 50, 25 and 10 ppm; N2NiO6.6H2O for 50, 25 and 10 ppm; Al(NO3)3.9H2O for 150, 50, 25 and 10 ppm; ZnSO4.7H2O for 10 ppm; CoN2O6.6H2O for 50, 25 and 10 ppm; Pb(NO3)2 for 500, 300, 150, 50, 25 and 10 ppm; CuSO4.5H2O for 10 ppm; MgSO4.7H2O for 500, 300, 150, 50, 25 and 10 ppm; MnSO4.2H2O for 500, 300, 150, 50, 25 and 10 ppm; Fe2(SO4)3.H2O for 50, 25 and 10 ppm; FeSO4.7H2O for 150, 50, 25 and 10 ppm and NaNO2 and for also 50, 25 and 10 ppm. In antibiotic sensitivity tests they were not sensitive to neomycin, streptomycin, vancomycin and gentamicin; they were sensitive against polymixin B, novobiocin (5 and 30µg), tetracycline, rifampicin, erythromycin, chloramphenicol and rifamycin. When biochemical tests were evaluated, there was no gas production from glucose. There was no growth at 45°C temperature, 3.0 and 3.5 pH. There was growth at 10°C and 15°C temperature, 3.0, 4.0, 5.0, 6.5, 8.0, 9.0% NaCl, 9.6 pH, 10.0%, 12.0% alcohol, MRS agar+3.0% ox-bile, MRS agar+5.0% ox-bile, MRS agar+9.0% ox-bile and in MRS broth+9.0% ox-bile. They showed β-galactosidase activity and antimicrobial activity against E.faecium.
Cluster 2 This cluster includes 15 isolates and has 89.4% similarity level. This high similarity ratio in cluster is an indicator of an acceptable taxonomic classification. glycerol, erythritol, D-arabinose, D-xylose, L-xylose, Dadonitol, methyl-βD-xylopyranoside, L-sorbose, dulcitol, inositol, amidon, glycogen, xylitol, D-lyxose, D-tagatose, D-fucose, L-fucose, L-arabitol, potassium 2ketogluconate, potassium 5-ketogluconate were not used in these isolates carbohydrate fermentation. Dribose, D-galactose, D-glucose, D-fructose, D-mannose, D-mannitol, N-acetylglucosamine, amygdaline, arbutin, esculin, salicin, D-celiobiose, D-maltose, gentiobiose were used; from resistance to heavy metals tests; CrN3O9.9H2O for 150, 50, 25 and 10 ppm; N2NiO6.6H2O for 150, 50, 25 and 10 ppm; Al(NO3)3.9H2O for 150, 50, 25 and 10 ppm; ZnSO4.7H2O for 50, 25 and 10 ppm; CoN2O6.6H2O for 50, 25 and 10 ppm; Pb(NO3)2 for 500, 300, 150, 50, 25 and 10 ppm; CuSO4.5H2O for 25 and 10 ppm; MgSO4.7H2O for 500, 300, 150, 50, 25 and 10 ppm; MnSO4.2H2O for 500, 300, 150, 50, 25 and 10 ppm; Fe2(SO4)3.H2O for 150, 50, 25 and 10 ppm; FeSO4.7H2O for 300, 150, 50, 25 and 10 ppm and NaNO2 and for also 300, 150, 50, 25 and 10 ppm, they were resistant to concentrations given above. When these isolates’ sensitivity to antibiotics was analyzed, they were not sensitive against bacitracin and gentamicin, yet it was found that they were sensitive against tetracycline, rifampicin, chloramphenicol and rifamycin. For biochemical tests; it was observed that there were no gas production from glucose and there was no growth at arginine test medium. 15°C temperature, 3.0%, 4.0%, %5.0%, 6.5% NaCl, 9.6 and 3.5 pH, MRS agar+3.0% ox-bile, MRS agar+5.0% oxbile, MRS agar+9.0% ox-bile and MRS broth+9.0% oxbile, there was growth under preceding conditions. They showed antimicrobial activity against L.monocytogenes Li1, and as result of Gram staining, it was determined that they were Gram (+) bacillus.
Unidentified 2 units of isolates and food isolate that is defined as L.helveticus according to carbohydrate fermentation tests were in this cluster. As it was expected, type strains that belong to other species were not included in this cluster. This case points out reliability of result, and it is possible to classify some
All food isolates which were identified as L.plantarum according to Carbohydrate fermentation tests and all types of L.plantarum strains were included in this cluster as it was expected previously. In this case, it would be
17
Ö. Ertekin, A.H. Çon Akademik Gıda 9(5) (2011) 11-20
CONCLUSION
isolates that cannot be identified by carbohydrate test by numerical taxonomic system.
Lactobacillaceae family is one of microorganism group members, which is commonly used in production of fermented foods. These members show important differences and cause very difficult definition problems depending upon isolated materiel used during biochemical tests. In this study, it is aimed to overcome problems encountered when traditional carbohydrate fermentation tests, which are not reliable and not enough to make definitions, are applied on Lactobacillaceae family members by using numerical taxonomy method in which all of the obtained test results that belong to distinctive properties are taken into consideration. To reach this target, morphological, metabolic and physiological characteristics of isolates were evaluated with type strains and duplicates, than they were classified into homogenous groups by using fundamental principles of numerical taxonomy. In this research 32 test isolates were used and 10 type strains were defined at 89.4% similarity levels according to SSMUPGMA analysis and test error was 1.11%. These obtained results are good indicator of numerical taxonomy which yielded in resulted with good results and also it is reliable, the identification level determined by numerical taxonomy for bacteria was not satisfactory and advantageous.
Single Member Clusters Other isolates and type strains formed 17 single member clusters. In these clusters, L.lactis ssp. lactis O3 with L.lactis 7; L.brevis O11 with L.brevis 2; L.curvatus O32 with L.curvatus 3, L.paracasei ssp. paracasei O54 with L.paracasei 5 were in a different cluster which was unexpected. Undefined O1, O15, and O18 isolates which could not be defined by using carbohydrate fermentation tests (API 50CH) were not included in numerical taxonomical studies any of the clusters. When all results were evaluated, other than L.plantarum isolates, it can be said that there are some differences in classification when either numerical taxonomy or carbohydrate fermentation tests were applied on isolates. Main reasons for these differences were that inter species differences were more obvious due to the fact that numerous characters were taken into consideration in numerical taxonomy and that numerous features which were used made it possible to determine distinctions in higher rations. As a result of SSM-UPGMA analysis, when 3 multiple member clusters were compared to each other, following similarity ratios were found: 85% similarity between multi member clusters 1 and 2, 80% similarities between clusters 1 and 3, and 78% similarities between clusters 2 and 3.
ACKNOWLEDGMENTS This work was financially supported by the Pamukkale University Scientific Research Council (Project Contract No: 2006FBE011). We would like to thank Dr. Nevzat Şahin and Dr. Kamil Işık for their assistance in the numerical calculations.
Determinative Properties of Used Tests
REFERENCES
When 176 tests used in numerical taxonomy were evaluated among themselves for determinative properties; 49 carbohydrates were evaluated with carbohydrate fermentation tests. 15 of those were negative, 1 of them had all positive result. As a consequence, they were not included in evaluation matrix, and 33 of them were determinative for numerical taxonomy (67.35%).
[1] Sneath, P.H.A., Sokal, R.R., 1973. Numerical Taxonomy: The Principles and Practise of Numerical Classification. W. H. Freeman, Baltimore. [2] Goodfellow, M., Dickinson, C.H., 1985. Delineation and Description of Microbial Populations Using Numerical Methods. In Computer-Assisted Bacterial Systematics, Edited by M. Goodfellow, D. Jones, F.G. Priest, Academic Press, London, pp. 165-225. [3] Mac Donell, M.J., Colwell, R.R., 1985. The Contribution of Numerical Taxonomy to the Systematics of Gram-Negative Bacteria. In Computer-Assisted Bacterial Systematics, Edited by M. Goodfellow, D. Jones, F.G. Priest, Academic Press, London, pp. 107-135. [4] Sackin, M.J., Jones, D., 1993. Computer-Assisted Classification. In Handbook of New Bacterial Systematics, Edited by M. Goodfellow, A.G. Q’Donnell, Academic Press, London, pp. 281-313. [5] Johansson, C.B., 1999. Classification of Bacteria, In Basic and Clinic Microbiology, Edited by Ş. Ustaçelebi, G. Mutlu, T., İmir, A.T., Cengiz, E. Tümbay, Ö. Mete, Güneş Kitabevi, Ankara, pp. 1339 (in Turkish) [6] Priest, F., Austin, B., 1993. Modern Bacterial Taxonomy. Chapman & Hall, London.
In resistance to heavy metals tests, 14 metals and 6 different concentrations for each metal that is total of 84 charters were tested. Among those characters 32 of them were negative in all isolates, 4 of them were positive in all isolates. As a result, 36 of these metal concentrations were deleted from evaluation matrix. Determining result of heavy metal tests for numerical taxonomy was 57.14%. In antibiotic sensitivity test, only 1 isolate among 16 type antibiotics had a positive result. For this reason, only one antibiotic (RF30) was not added to evaluation matrix. Moreover, its determinative ratio was 93.75%. On the other hand, in biochemical tests, there were no test characters that any isolates had either all negative or all positive results. So, it was assumed that determinative ratio is 100%.
18
Ö. Ertekin, A.H. Çon Akademik Gıda 9(5) (2011) 11-20
[7] Williams, S.T., Goodfellow, M., Alderson, G., Wellington, E.M.H., Sneath, P.H.A., Sackin, M.J., 1983. Numerical classification of streptomyces and related genera. Journal of General Microbiology 129: 1743-1813. [8] Temmerman, R., Huys, G., Swings, J., 2004. Identification of lactic acid bacteria: culturedependent and culture-independent methods. Trends in Food Science & Technology 15: 348-359. [9] Johansson, M-L., Sanni, A., Lönner, C., Molin, G., 1995. Phenotypically based taxonomy using API 50CH of Lactobacilli from Nigerian Ogi and the occurence of starch fermenting strains. International Journal of Food Microbiology 25(2): 159-168. [10] Bringel, F., Quenee, P., Tailliez, P., 2001. Polyphasic investigation of the diversity within Lactobacillus plantarum related strains revealed two L. plantarum subgroups. Systematic and Applied Microbiology 24(4): 561-571. [11] Lönner C., Johansson M.L. 2005. Pro equoä-a liquid oat product fermented with Lactobacillus plantarum JI:1 (DSM 11520)- a new probiotic product for horses. http://www.proequo.se/docs/ProEquovet.sammanfattnin-050222.doc, pp.6. [12] De Angelis M., Corsetti A., Tosti N., Rossi J., Corbo M.R., Gobbetti, M. 2001. Characterization of NonStarter Lactic Acid Bacteria from Italian Ewe Cheeses Bases on Phenotypic, Genotypic, and Cell Wall Protein Analyses. Applied and Environmental Microbiology 67(5): 2011-2020. [13] Dykes, G.A., Britz, J.J., Holy, A., 1994. Numerical taxonomy and identification of lactic acid bacteria from spoiled, vacuum-packaged Vienna sausages. Journal of Applied Bacteriology 76(3): 246-52. [14] Grant, I., Patterson, M.F., 1991. A numerical taxonomic study of lactic acid bacteria isolated from irradiated pork and chicken packaged under various gas atmospheres. Journal of Applied Bacteriology 70: 302-307. [15] Borch, E., Molin, G., 1988. Numerical taxonomy of psychrotrophic lactic acid bacteria from prepacked meat and meat products. Antonie Van Leeuwenhoek 54(4): 301. [16] González, C.J., Encinas, J.P., García-López, M.L., Otero, A., 2000. Characterization and identification of lactic acid bacteria from freshwater fishes. Food Microbiology 17(4): 383-391. [17] Çon, A.H., 1995. Isolation and Identification of Bacteriocin and/or Bacteriocin-like Metabolities Producing Lactic Acid Bacteria From Sucuk Samples and Antagonistic Activities of These Bacteria Against to Some Food Spoilage and Food-Borne Pathogenic Bacteria. Ph.D Thesis. Atatürk University Graduate School of Natural and Applied Sciences, Erzurum (in Turkish). [18] Şahin, N., 1995. Selective isolation, characterisation and classification of novel thermotolerant Streptomycetes. Ph.D Thesis, Newcastle University, England. [19] Sánchez, M.M., Delgado, T., Alonso, L., Baltasayar, M., 2000. Phenotypic and genetic characterization of a selected set of Lactococcus lactis strains isolated from a starter-free farmhouse cheese. Food Microbiology 17: 449-460.
19
[20] Yüksekdağ, Z.N., Beyatlı, Y., Aslım, B., 2004. Determination of some characteristics coccoid forms of lactic acid bacteria isolated from Turkish kefirs with natural probiotic. Lebensmittel-Wissenschaft undTechnologie 37(6): 663-667. [21] Kask, S., Adamberg, K., Orlowski, A., Vogensen, F.K., Moller, P.L., Ardö, Y., Paalme, T., 2003. Physiological properties of Lactobacillus paracasei, L. danicus and L. curvatus strains isolated from Estonian semi-hard cheese. Food Research International 36: 1037-1046 [22] Sneath, P.H.A., Johnson, R., 1972. The influence on numerical taxonomic similarities of errors in microbiological tests. Journal of General Microbiology 72: 377-392. [23] Sokal, R.R., Michener, C.D.,1958. A statistical method for evaluating systematic relationship. Kansas University Science Bulletin 38: 1409-1438. [24] Schillinger, U., Lucke, F.K., 1989. Identification of Lactobacilli from meat and meat products. Applied and Environmental Microbiology 4: 199-209. [25] Pepe, O., Blaiotta, G., Anastasio, M., Moschetti, G., Ercolini, D., Villani, F., 2004. Technological and molecular diversity of Lactobacillus plantarum strains isolated from naturally fermented sourdoughs. Systematic and Applied Microbiology 27: 443-453. [26] Dal Bello, F., Clarke, C.I., Ryan, L.A.M., Ulmer, H., Schober, I.J., Ström, K., Sjögren, J., van Sinderen, D., Schnürer, J., Arendt, E.K., 2000. Improvement of the quality and shelf life of wheat bread by fermentation with the antifungal strain Lactobacillus plantarum FST 1.7. Journal of Cereal Science 45: 309-318. [27] Şimşek, Ö., 2003. Identification of Lactic Acid Bacteria Having Antimicrobial Activity and Determination of Their Some Metabolic Properties Isolated from Uşak Sourdoughs. Master of Science Thesis. Pamukkale Univ. Graduate School of Natural and Applied Sciences, Denizli, Turkey (in Turkish). [28] Kostinek, M., Specht, I., Edward, V.A., Pinto, C., Egounlety, M., Sossa, C., Mbugua, S., Dortu, C., Thonart, P., Taljaard, L., Mengu, M., Franz, C.M.A.P., Holzapfel, W.H., 2007. Characterisation and biochemical properties of predominant lactic acid bacteria from fermenting cassava for selection as starter cultures. International Journal of Food Microbiology 114: 342-351. [29] Kim, S.W., Perl, L., Park, H.J., Tandianus, E.J., Dunn, W.N., 2001. Assessment of stress response of the probiotic Lactobacillus acidophilus. Current Microbiology 43: 346-350. [30] Carr, J.F., Chill, D., Miada, N., 2002. The lactic acid bacteria: a literature survey. Critical Reviews in Microbiology 28(4): 281-370. [31] Şimşek, Ö., Çon, A.H., Tulumoğlu, Ş., 2006. Isolating lactic starter cultures with antimicrobial activity for sourdough processes. Food Control 17: 263-270. [32] Papamanoli, E., Tzanetakis, N., LitopoulouTzanetaki, E., Kotzekidou, P., 2003. Characterization of lactic acid bacteria isolated from a Greek dryfermented sausage in respect of their technological and probiotic properties. Meat Science 65: 859-867. [33] G-Allegria, E., Lopez, I., Ruiz, I.J., Saenz, J., Fernandez, E., Zarazaga, M., Dizy, M., Torres, C.,
Ö. Ertekin, A.H. Çon Akademik Gıda 9(5) (2011) 11-20
Ruiz-Larrea, F. 2004. High tolerance of wild Lactobacillus plantarum and Oenococcus oeni strains to lyophilisation and stress environmental conditions of acid pH and ethanol. FEMS Microbiology Letters 230: 53-61. [34] Randazzo, C.L., Restuccia, C., Daniele Romano, A., Caggia, C., 2004. Lactobacillus casei, dominant species in naturally fermented Sicilian green olives. International Journal of Food Microbiology 90: 9-14. [35] Hébert, E.M., Raya, R.R., Tailliez, P., De Giori, G.S., 2000. Characterization of natural isolates of Lactobacillus strains to be used as starter cultures in dairy fermentation. International Journal of Food Microbiology 59: 19-27. [36] Cebeci, A., Gürakan, C., 2003. Properties of potential probiotic Lactobacillus plantarum strains. Food Microbiology 20: 511-518. [37] Badis, A., Guetarni, D., Moussa-Boudjemâ, B., Henni, D.E., Tornadijo, M.E., Kihal, M., 2004. Identification of cultivable lactic acid bacteria isolated from Algerian raw goat’s milk and evaluation of their technological properties. Food Microbiology 21: 343349. [38] Sánchez, I., Sesena, S., Poveda, J.M., Cabezas, L., Palop, L., 2005. Phenotypic and genotypic characterization of lactobacilli isolated from Spanish goat cheeses. International Journal of Food Microbiology 102: 355-362.
20
[39] Bujalance, C., Jiménez-Valera, M., Moreno, E., RuizBravo, A., 2006. A selective differential medium for Lactobacillus plantarum. Journal of Microbiological Methods 66: 572-575. [40] Charteris, P.W., Kelly, M.P., Morelli, L., Collins, K.J., 1998. Antibiotic susceptibility of potentially probiotic Lactobacillus species. Journal of Food Protection 61(12): 1636-1643. [41] Kelly, W.J., Davey, G.P., Ward, L.J.H., 1998. Characterization of Lactococci isolated from minimally processed fresh fruit and vegetables. International Journal of Food Microbiology 45: 85-92. [42] Hassen, A., Saidi, N., Cherif, M., Boudabous, A., 1998. Resistance of environmental bacteria to heavy metals. Bioresource Technology 64: 7-15. [43] Gülcan, S., 2006. Determination of Heavy Metal and Naphthalene Tolerance of Pseudomonas, Isolated From Different Sources. Master of Science Thesis. Pamukkale University Graduate School of Natural and Applied Sciences, Denizli, Turkey (in Turkish). [44] Doğramacı, M., 2005. Isolation and Characterisation of Members of the Streptomyces violaceusniger Clade from Rhizosphere. Master of Science Thesis. Ondokuz Mayıs University Graduate School of Natural and Applied Sciences, Samsun, Turkey (in Turkish).