Briefly, each strain was grown on sheep blood agar anaerobically at 30 C for 5 days.
The inoculum of each strain was prepared using SL broth according to the manufacturer's instructions. API CH50 kits were inoculated and mineral oil was added to each tube of the API CH50 kit to provide anaerobic conditions. The kits were incubated at 30 C aerobically for 4 days. The tests were read after 24hr, 48hr, 72hr, and 96hr of incubation. A positive test corresponded to acidification revealed by the bromocresol purple indicator, which was contained in SL, changing to yellow. For the Esculin test, a change from purple to black indicated a positive reaction.
Isolated strains were identified to dairy propionibacteria species by the method summarised below:
Identification of isolated Propionibacterium strains
Genus identification - Traditional identification method
Propionibacterium genus-specific PCR
Species identification L
Traditional identification method dairy propionibacteria species-specific PCR
Comparison of strain difference
SDS-PAGE profile of whole cell water-soluble protein
API 50 CH profile Presence of dairy propioaibacteria strains
Six dairy propionibacteria strains were isolated of which four were from milk and two were from the Swiss-cheese sample (Table 6). These six isolates were Gram positive, catalase positive, non-spore forming, irregular short rods. These six isolates did not hydrolyse gelatin and can therefore be included in dairy propionibacteria species.
No dairy propionibacteria strains were isolated from Parmesan cheese, Best Quality Gouda cheese, Grana Papano cheese and Jarlsberg cheese samples (results not shown). No dairy propionibacteria strains were isolated from 87 human gut biopsy samples (results not shown).
Table 6 Labels and origin of isolated strains
Isolates strains Source
702 Milk sample 1
801 Milk sam le 2
901 Milk sample 3
1001 Milk sample 4
Genus Idectification using Propiofzibacterium genus-specific PCR method
Figure 1 shows the agarose gel electrophoretic profiles of PCR products obtained for six isolated and six reference strains using Propionibacterium genus-specific primers PB1- PB2. Identification of isolated strains was performed by visual comparison of the electrophoretic profiles of PCR products of isolated strains with that of reference strains of dairy propionibacteria species. Only one PCR product was produced for all the tested strains.
Primer pair PB 1-PB2 has been proven to be specific for dairy propionibacteria and P. acnes (Rossi, 1999). P. acnes belongs to cutaneous propionibacteria (Cummins and Johnson, 1986).
Therefore, the six isolates can be identified as strains of genus Propionibacterium, but not necessarily as strains of dairy propionibacteria, using genus specific PCR with primer sets PB 1-PB2.
Species idefatification of isolated strains usifag traditional methods
The morphological and tested biochemical characteristics of the isolated Propiottibacteriunt strains are listed in Table 7. For the purpose of comparison, the morphological and tested biochemical characteristics of some reference strains are also listed in Table 7. All the reference strains, except for P. freudenreichii CSCC2206, gave expected results based on published data (Cummins and Johnson, 1986). Isolated strains were identified to species by the quick traditional identification scheme. P. freudenreichii CSCC2206 did not give expected results for maltose and sucrose fermentation of P. freudenreichii species (Cummins and Johnson, 1986).
Most of the strains of dairy propionibacteria will be catalase positive (Cummins and Johnson, 1986). Strain P. acidopropionici ATCC25562 showed negative for catalase tests, but this is not a false negative because only 40-90% ofstrains of P. acidopropionici have been reported to be positive for catalase (Cummins and Johnson, 1986).
There is variation in nitrate reduction amongst the P. freudenreichii strains. This indicates strain differences between the 9 identified P. freudenreichii strains. This variation is supported by the published data on nitrate reduction by P. freudenreichii strains (Cummins and Johnson, 1986), that is, that 11-89% of the strains will reduce nitrate.
Table 7 A summary of the different characteristics of tested strains
Strains g. 0 u u u u u t) ton
U U U U U U tests c c) c < ) o') c) rj
U U U U U H
SES 'S' <
K S ! S S C'5 : K i i i i - c < ) cSooo'-aQ-. , Ei-. CL, (c Pigmentation
U U U U U U U U U U U U U Catalase + + + + + Gelatin
4 t : S 3 hydrolysis
Maltose + + + + + Malase , Sucrose + + 13-haemolysis 'F, " k. F a t c
S k, N tri + + i te + + + reduction
Species c- : Symbols : +, positive reaction or pH below 5. 7 ;-, negative reaction or pH above 5.7.
Species identification of isolated strains using Propionibacterium species-specific PCR
Identification of isolated strains was performed by visual comparison of the electrophoretic profiles of species-specific PCR products of isolated strains with those of reference strains. PCR products were detected from the PCR reaction using primer sets PF- PB2, PJ-PB2, and PT3-PB2 (Figure 2, Figure 3, Figure 4); however, no PCR products were found from the PCR reaction using primer set PA-PB2 (results not shown).
Figure 2 shows the electrophoretic profiles of PCR products produced by primer set PF-PB2. Only one PCR product of the same size was produced for strains 201al, 201b, 801, 901, 1001, P. freudenreichia CSCC2200, P. freudenreichii CSCC2201, P. freudenreichii CSCC2207, and P. freudenreichii CSCC2216. However, strain 702 and P. freudenreichii CSCC2206 produced no PCR product. Primer set PF-PB2 has been found to be specific for P.
freudenreichii (Rossi et al, 1999). Therefore, strains 201al, 201b, 801,901 and 1001 were identified as strains of P. freudenreichii but not strain 702 and P. freudenreichii CSCC2206.
This result correlates with the results using traditional identification methods (Table 7). In addition, this may suggest that the identification status of P. freudenreichii CSCC2206 is in doubt. It may not belong to P. freudenreichii.
Figure 3 shows the agarose gel eletrophoretic profiles of strain 702 and reference strains, which were obtained with primer pair PJ-PB2. Only one PCR product of the same size was produced by strains 702, P. jensenii NCFB571 and NCFB572. No PCR products were produced by strains P. freudeireichii CSCC2206, CSCC2207, or P. acidopropionici ATCC25562. Primer pair PJ-PB2 has been found to be specific to P. jensenii (Rossi et al, 1999). Therefore, strain 702 was identified as a strain of P. jensenii.
Figure 4 shows the agarose gel electrophoresis profiles of strain 702 and P. thoenii ACM365, which were obtained with primer pairs PT3-PB2. P. thoenii ACM365 produced one PCR product of the same size. No PCR products were produced by strain 702. Primer pair PT3-PB2 has been found to be specific to P. thoenii (Rossi et al, 1999). Therefore, strain 702 was not a strain of P. thoenii.
Species ideratificatioya of isolated strains using RAPD-PCR
Figure 5 shows the agarose gel electrophoresis profiles obtained by RAPD-PCR with OPL-05. Identification of six isolated strains was performed by visual comparison of their electrophoretic profiles with those of reference strains. Strain 201al, 201b, 801, 901, 1001, and P. freudenreichii CSCC2206 produced only one PCR product of the same size; while strains P. acidopropionici ATCC25562 and 341 had patterns of multiple PCR products with some common bands. Strain 702, P. jensenii NCFB572 and strain P. freudenreichii CSCC2207 did not produce any products when tested under the same amplification program using OPL-05 (results not shown). Therefore, strain 201al, 201b, 801,901 and 1001 can be identified as strains of P. freudenreichii, while strain 702 could not be identified using this RAPD-PCR method.
SDS-PAGE profile of whole cell water-soluble proteins of isolated strains
Whole cell water-soluble proteins of six isolated strains and reference strains were analysed by SDS-PAGE. The electrophoretic profile of each isolated strain was compared with those of reference strains. SDS-PAGE band patterns of different strains were compared by the presence or the absence of typical bands but not their intensity, since intensity of the bands may vary for different batches of protein samples for the same bacteria (Baer, 1987).
Figure 6 and Figure 7 show the SDS-PAGE band patterns of one isolate, strain 702, and reference strains of four species of dairy propionibacteria. The SDS-PAGE band patterns of replicate samples of strain 702, P jensenii NCFB572, and P. thoesaii ACM365 were identical. This indicates the high reproducibility of the SDS-PAGE identification method.
The band pattern of strain 702 was almost identical to that of Pjensenii NCFB572, but different from that of P. thoenii ACM365, P. freudenreichii CSCC2207, P. acidopropionici ATCC25562 and L. acidophilus MJLA1. Therefore, strain 702 was closely related to Pjensenii NCFB572.
Figure 8 shows SDS-PAGE band patterns of six isolated strains and five reference strains of P. freuderzreichii. Isolated strain 702 had different patterns from those of the five strains of P. freudenreichii, while isolated strains 201al, 201b, 801,901 and 1001 had partly similar band patterns to those of five reference strains of P. freudenreichii.
The relatedness of isolated strains to five reference strains of P. freudenreichii is further revealed in Table 8 by homology analysis of SDS-PAGE protein banding patterns using Dice Coeffient (Quality One software package, Bio-Rad). The homology percentile indicates genetic relatedness, calculated using the number of shared genetic markers or bands.
The higher percentiles, the more closely related the strains. The matching percentages of strain 201al, 201b, 801,901 and 1001, with P. freudenreichii CSCC2207, were higher than with any of the other four P. freudenreichii reference strains. Strain 702 had a low matching percentage with the five reference P. freudenreichii strains. Strains 201al, 201b, 801,901 and 1001 were considered closely related to P. freudenreichii CSCC2207, while strain 702 was not closely related to any of the five P. freudenreichii reference strains.
Table 8 Correlation matrix: homology from SDS-PAGE of whole cell soluble proteins of Propionibacterium strains
Percent C :) CD CD CD . y0
C, 4 cq cq N cq i S coq cul
1-4 cq CD
C) CD 0 C :) 0 C :) CO CO CI) 0, CO cn P. freudenreichii 100.
CSCC2216 0 CSCC2207 73. 1 69. 6 57. 7 73. 2 73. 6 73. 9 46. 7 82. 8 51. 5 100.
CSCC2207 0 CSCC2206 1 1 1 0 CSCC2206 0 P. f reuderareichi. i 66. 9 67. 7 51. 3 69. 9 70. 2 74. 5 5, 100.
CSCC2201 0 P. freudefireiclzii CSCC2200 483 46. 4 32. 7 56. 5 56. 4 36. 6 1 0. 0 1001 61. 4 64. 5 66. 9 61. 6 63. 7 100. 0 901 ~ 70251. 8 50. 4 100. 0 801 74. 6 75. 5 44. 2 100. 0 702 51. 8 50. 4 100. 0 201b 86. 8 loo. o 201al 100. 0 *: homology derived using the Dice Coefficient, Quality One Software (Bio-Rad)
Table 9 API 50 CH results of tested strains
; s ! s S' . 3 s Carbohydrate-40
U : o m U w U U r 0. ; Z 0. Z 0. d d YU Ne ative control Glycerol+ + + + + + Glycerol + + + + + + fZf N $Z ;'f'Z D af-,, 3 ; fa ; ,'g-ff'Rf > ..-, > i, ~ a, , ;'fD Jr fa, P : f ; ; + Ribose + + + + + t D-Xylose L-X lose ,"' f Adonitol + + + + + + L-Xylose--f, ;,-ff a f,"E ft < f fiX ff S iLiSr 0 a S f, ( ; f ;,'fg S ; f 6 Methyl-xylose Methyl-xylose w ff if-^ < E 2gf ; 7 < E E E,, 2ff,., a, Galactose + + + + + + D-Glucose + + + + + + D-Fructose + + + + + + D-Mannose + + + + + + L-Sorbose ~ < fa-i t00 < ; f, S i g 3z af < . : fnt 5g"i-affffft L-Sorbose Rhamnose Inositol + + + +. JSafa fff f + Dulcitol Inositol + + + + Mannitol + + + + Sorbitol + + a Methy1-D-mannoside-+ ct Meth 1-D-lucoside ", + + aMethyl-D-mannoside fffL fa, gfR t aff < i f fi + 0 ; t i < Zf f-itf < fff'f bf f'., ;, f t + + ff zff A ffifaf-ffafff fe f Amygdaline w affffkni Xa^ ; ff f ; u ; a affff a < < f g af, o, E ffaff faf , f ; iffffffififaf + + + D-Raffinose ; f f ffif}'', f}'', ffiit if ffb Sff i ; Arnfidon i g if ff'f "''uf : Ea-$ ; f fif'fZ t Glycogene, a fCfXfi fff j f f ; f, E f : f f i. ~ f f, > },'ff t fff'f f'if z f i,,"~. fi ;-; fa ; tgf- > f f f u Xylitol + + f fa'+. + + Genitobiose + + ; ffal-f f ; f if-~ + + L-sucrose D-ArabitoI++++++ D-Lyxose--fi} 0 7 t fL f ; ; C Fff ; g} Bf W ? ; ; f D-Tagatose : 77 < 7, Tff ; f-ffE gf !} ; X t if, D-Fucose ; : ; ff t ; f i tV ; ff 0 ffA ifX < L-Fucose f i f-a ;, fT"; gZ. 0 fiAt ; f if 0 f ff ; t fT fa ff L-Arabitol++++++ Gluconate + + + + + + 2 ceto-gluconate 5 ceto-gluconate 2ceto-jluconate z i05fiXf ;--af^ ; j-f 0Ti 0 -; f, ,'f < , f^^ff f-fr f 0 fE fd 5 ceto-gluconate t ffS }-ft e---gRg gi : : ; f ~ t i$.
API 50 CH profile of dairy propionibacteria strains
API 50 CH strips were used to look at the difference of strain P. jensenii 702 from the reference strains of P. jensenii, P. jeiisenii NCFB 571 and NCFB572. As reaction control, reference strains of the other three dairy propionibacteria species, P. freudenreichii CSCC2207, P. acidopropionici ATCC25562, and P. thoenii ACM365, were also tested using API 50 CH strips. Table 9 presents the 49 carbohydrate fermentation patterns of those six strains using API 50 CH strips.
The API 50 CH profiles of all tested strains (Table 9) were compared to available published carbohydrate profiles of each species (Cummins and Johnson, 1986). The API 50 CH profiles of strain P. freudenreichii CSCC2207, P. acidopropionici ATCC25562, P. jensenii NCFB572 and P. thoenii ACM365 correlated with published data; while P. jensenii NCFB 571 had one different reaction, namely, the inability to ferment salicine, which has been reported positive for 90-100% strains of P. jensenii (Cummins and Johnson, 1986).
Isolated strain 702 had similar carbohydrate profiles to those reported for P. jensenii (Cummins and Johnson, 1986), except for its inability to ferment D-Raffinose, which has been reported positive for 90-100% strains of P. jensenii (Cummins and Johnson, 1986).
Discussion and Conclusions
Six dairy propionibacteria strains were isolated using YELA and were identified to species by traditional methods, Propionibacterium genus-specific PCR, Propionibacterium species-specific PCR, RAPD-PCR, and SDS-PAGE analysis of whole cell water-soluble protein. The relatedness of isolated strains to selected reference strains was also illustrated by SDS-PAGE analysis of whole cell water-soluble proteins. In addition, the different carbohydrate profiles for five reference and one isolate strain (702) were compared using API 50 CH carbohydrate fermentation profile.
The six dairy propionibacteria strains were found in Swiss cheese and raw milk samples but not in samples of Parmesan cheese, Best Quality Gouda cheese, Grana Papano cheese and Jarlsberg cheese. However, other studies have reported that Propionibacterium have been isolated from Grana, Parmesan and Pasta Filata cheese samples as well as raw milk samples (Rossi et al, 1998; Fessler et al, 1999). These results confirm the known habitats of Propiortibacteriurrz, which primarily are dairy products, human skin and silage (Jones and Collins, 1986; Grant and Salminen, 1998).
No dairy propionibacteria strains were isolated from the 87 biopsy samples in this study. The collection of biopsy samples was from patients at the Sydney Adventist Hospital.
This hospital has been established by the Sydney Seventh Day Adventist Church, which is well recognised for promoting vegetarianism. As a consequence, these biopsy samples may
have been collected from a predominantly vegetarian population, who would have had little exposure to dairy products containing dairy propionibacteria. The results of biopsy samples may indicate that differential vitamin B12 levels in certain vegetarians may be more a consequence of exposure to certain food microorganisms rather than dietary intake.
The six isolated strains were identified to dairy propionibacteria species by traditional identification methods, based on morphological and biochemical characteristics. Only four characteristics were adopted in this study to differentiate four species of dairy propionibacteria, including fermentation of sucrose and maltose; reduction of nitrate; B- hemolysis, and pigment colour. The isomer of diaminopimelic acid (DAP) in the cell wall, which has been recommended as one of the differentiation characteristics for the four dairy propionibacteria species, was not tested in this study.
The species status of isolated strains, which were identified by traditional methods, was confirmed by Propionibacterium genus-specific and species-specific PCR profiles. All reference strains of Propionibacterium and six isolated strains had the expected PCR product, which confirms the findings of Rossi, et al. (1999). Tested Propionibacterium strains produced PCR products (Figure 3, Figure 4) as described by Rossi, et al. (1999), using primer sets PJ-PB2, and PT3-PB2. Four out of five reference strains of Pfreudenreichii produced the same PCR products as described by Rossi, et al. (1999) using primer set PF-PB2 (Figure 2); however, P. freudenreichii CSCC2206 did not produce any PCR product (Figure 2). There was also no PCR product detected for both P. acidopropionici ATCC25562 and P. acidopropionici 341 (data not shown) when using PA-PB2, which is specific for P. acidopropionici (Rossi, 1999). P. freudenreichii CSCC2206, P. acidopropionici ATCC25562 and P. acidopropionici 341, have not been used in the study by Rossi, et al. (1999). This discrepancy may be due to the different preparation of genomic DNA or difference between laboratory conditions.
Like species-specific PCR, the RAPD-PCR analysis using primer OPL-05 demonstrates that this technique shows clear similarity between strains of P. freudenreichii and the difference between strains of P. freudenrech and P. acidoproponici. This is correlated with the results of Rossi (1998). There was no PCR product produced by P. freudenreichii CSCC2207, strain 702 and P. jensenii NCFB572. This may indicate the limitation of RAPD-PCR, which may not be applied to all strains of Propionbacterium.
SDS-PAGE protein analysis is a quick technique for species and strain differentiation of Propionibacteriurz strains (Baer, 1987; Fessler et al, 1999). The almost identical SDS- PAGE profiles of replicate samples of strain 702 (Figure 6, Figure 7), replicate samples of P. thoenii ACM365 (Figure 6) and replicate samples of P. jensenii NCFB572 (Figure 7) indicates the high reproducibility of SDS-PAGE analysis of bacterial soluble protein. The reproducibility of SDS-PAGE analysis of bacterial soluble protein has been reported to be 92-
98% (Costas et al, 1990; Costas, 1990), and 93-97% (Tsakalidou et al, 1994) for the identification of the species.
A 70% homology of SDS-PAGE profile cut off has been recommended for species differentiation (Fessler et al, 1999). This separates the five reference strains of P. freudenreichii used here into four species, with only P. freudenreichii CSCC2201 and CSCC2207 having greater than 70% homology (82. 8%) (Table 8). If 201al and 801 are to be considered the same species as P. freudenreichii CSCC2207, they should be the same species as P. freudenrecltii CSCC2201. Therefore, further studies may be needed to screen more reference strains and to identify their common bands for species differentiation rather than just comparing homology of patterns of all the bands.
The clearly different API profiles for different species of dairy propionibacteria can be used to separate species (Table 9). The variation of the API 50 CH profile with three strains of P. jensenii is also a useful tool to show strain differences, which is essential for the selection of certain strains as probiotics.
EXAMPLE 2 DAIRY PROPIONIBACTERIA SIMULATED UPPER GASTROINTESTINAL TRANSIT TOLERANCE AND ADHESION TO C2BBE1 CELLS IN VITRO
This study aimed to assess in vitro upper gastrointestinal tract transit tolerance and adhesion to human epithelial cells of some dairy propionibacteria strains, which are able to synthesize vitamin B12. For this purpose, the following were examined, (1) the viability of dairy propionibacteria strains in simulated gastric transit conditions (pH2, pH3 and pH4 gastric juices); (2) the influence of vegetarian food on the pH2 gastric transit tolerance of dairy propionibacteria strains; (3) the viability of dairy propionibacteria strains in simulated small intestinal transit conditions; (4) the growth of dairy propionibacteria strains in SLB containing 0.3% bile salt; and (5) the adhesion of dairy propionibacteria strains to cell line C2BBel, a clone of the Caco-2 cell line.
Materials and Methods Bacterial strains
Thirteen dairy propionibacteria strains were screened for their in vitro gastrointestinal transit tolerance and adhesion to cell line C2BBel. These strains were selected based on their vitamin B12 synthesis in Fermentation Medium (FM). The sources and vitamin B12 production levels of these strains are presented in Table 10.
Lactobacillus acidophilus MJLA1 and Bifidobacteriufn lactis BDBB2 were provided and described as Caco-2 cell adhesion positive and negative respectively by Gist Brocades, Australia. They were used as positive and negative controls respectively in the C2BBel adhesion assay.
The recovery ayad preservation of bacterial strairas
The recovery and preservation of Propionibacteriurm strains and L. acidophilus MJLA1 was performed as described in Example 1. B. lactis BDBB2 was recovered in Reinforced Clostridial Medium (RCM, Oxoid), and then streaked onto Reinforced Clostridial Agar (RCA, Oxoid) to establish purity. It was Gram stained, catalase tested, and examined microscopically for correct cell morphology.
Determination of vitamin BIZ production by dairy propionibacteria in Fernzentation Mediunz (FM)
The determination of vitamin B12 production by dairy propionibacteria strains in FM was adapted from the method of Quesada-Chanto, et al. 1998.
Briefly, strains of dairy propionibacteria were grown in SLB for 48 hr anaerobically.
The 48 hr culture (20AL) of each strain was inoculated into FM broth (20mL) and incubated at 37 C for 4 days. Bacteria cells in FM (10mL) were then collected by centrifugation (3 500 x g, 15 min) in a 10 mL sterile centrifuge tube. Bacteria pellets were then washed in 0.9 % NaCl (lOmL) twice. The bacteria pellets were then resuspended in 0. 1M pH 5.5 phosphate buffer (8mL), and 1% KCN (0.1 mL) (Chem supply) was added. The bacterial suspension was then autoclaved at 121 C for 10 min. After cooling to room temperature, the suspension was centrifuged at 3 500 x g for 20 min. The supernatant was collected, and stored at 4 C prior to vitamin B12 testing. Vitamin B12 levels were determined by Quantaphase Bl2/Folate assay (Bio-Rad) at the Sydney Adventist Hospital, Sydney, NSW, Australia.
Human gut epithelial cell line C2Bbel
Cell culture C2BBel was kindly provided by Dr. Matthias Ernst of the Ludwig Institute in Melbourne, Australia. C2BBel is a clone of cell culture Caco-2, which is a human colon epithelial cell line.
Subculturing of C2BBel cells
C2BBel cells were cultured in 25cm3 tissue culture flasks containing lOmL of Supplemented RPMI 1640 at 37 C, 95% air/5% C02 for 7 days or until confluent. Cells were fed with fresh Supplemented RPMI 1640 every second day.
Preservation of C2BBel cells
The C2BBel cells were grown in a 25cm3 tissue culture flask until confluent. The growth medium was then and a solution of 0. 25% trypsin, 0.2% EDTA (lmL) (Sigma) was added to the culture flask to rinse the cells. An additional 1. 5mL of 0.25% trypsin, 0.2% EDTA solution was then added and the flask was allowed to stand at 37 C, 95% air/5% COx for 10 min or until the cells detached. The detached C2BBel cells were resuspended in lOmL of RPMI 1640 broth (Trace Biosciences) with 5% DMSO (Sigma) using Pasteur pipettes.