Comparison of four methods to enumerate probiotic bifidobacteria in a fermented food product

ARTICLE IN PRESS FOOD MICROBIOLOGY Food Microbiology 23 (2006) 571–577 www.elsevier.com/locate/fm

Comparison of four methods to enumerate probiotic bifidobacteria in a fermented food product Sampo J. Lahtinen, Miguel Gueimonde, Arthur C. Ouwehand, Johanna P. Reinikainen, Seppo J. Salminen Functional Foods Forum, Department of Food Chemistry and Biochemistry, University of Turku, 20520 Turku, Finland Received 18 May 2005; received in revised form 9 September 2005; accepted 14 September 2005 Available online 26 October 2005

Summary Four methods of enumeration were compared by monitoring levels of probiotic bifidobacteria in fermented oat drink during storage. Strains of Bifidobacterium longum and B. lactis were quantified by plate counts, fluorescent in situ hybridization (FISH), quantitative TM real-time PCR and commercial LIVE/DEADs BacLight bacterial viability kit, and the methods were further developed to suit the enumeration of bifidobacteria in fermented foods. Plate counts of both B. lactis and B. longum were lower than the PCR and FISH counts. The LIVE/DEAD counts of B. lactis were comparable to PCR and FISH counts. The plate counts of B. lactis were slightly but significantly lower than LIVE/DEAD counts, suggesting that the cells that were not able to grow on plates may have become dormant. The plate counts of B. longum were several log units lower than LIVE/DEAD counts, suggesting that a remarkable part of the cells were dormant. Real-time PCR and FISH were shown to suit the quantification of the total amount of probiotic bifidobacteria in foods. Plate counts and LIVE/DEAD counts provided conflicting information on viability especially in the case of B. longum. We conclude that the choice of enumeration method for probiotic bacteria may have significant effect on the results of the analysis. The strain-specific properties and the objects of the analysis should be taken into account when enumeration methods for different probiotic strains are chosen. r 2005 Elsevier Ltd. All rights reserved. Keywords: Dormant bacteria; Enumeration methods; Probiotic bacteria; Survival; Viability

1. Introduction Reliable determination of viability of bacteria in probiotic products is important as the definition of probiotics calls for viable microbes. Many studies have shown that viability of bacteria is not a simple question of cell being dead or alive (Barer et al., 2000; Bloomfield et al., 1998; Colwell, 2000; Dodd et al., 1997; Kell et al., 1998; Nystrom, 2001; Yamamoto, 2000). Plate count method has traditionally been used for determination of viability of bacteria, but there are obvious disadvantages. First, plate count requires long incubation times. Plate count method is often hampered by technical difficulties such as clumping Corresponding author. Tel.: +358 2 3336861; fax: +358 2 3336862.

E-mail address: sajola@utu.fi (S.J. Lahtinen). 0740-0020/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.fm.2005.09.001

and inhibition by neighbouring cells. The choice of enumeration medium and incubation conditions for specific species may also be challenging. For many species a suitable growth medium is not known. Furthermore, for fastidious micro-organisms such as Bifidobacterium species of intestinal origin, it may be difficult to find an optimal growth medium for reliable enumeration. Such bacteria have unique nutritional and environmental requirements for optimal growth, and plate counts for certain strains may vary by several log units when grown on different nutrient-rich media (Martineau, 1999; Payne et al., 1999). Screening for optimal growth medium is a laborious task, and it is not often feasible to test all potential growth media. The difficulties described above may lead to underestimation of the bacterial counts. Another major problem of the plate count method is that so-called

ARTICLE IN PRESS 572

S.J. Lahtinen et al. / Food Microbiology 23 (2006) 571–577

‘dormant’ bacteria (Kell et al., 1998) are unable to grow on conventional growth media, but may nevertheless be measured as viable using cytological viability assays. Bunthof and Abee (2002) reported that such dormant population may exist in probiotic products and dairy starters, and similar population occurring in bile acid stressed bifidobacteria was later demonstrated by Ben Amor et al. (2002). Data from our laboratory suggests that probiotic bacteria in fermented product may become dormant during prolonged storage (Lahtinen et al., 2005). To determine viability of bifidobacteria in a probiotic product during storage, we compared four different methods of enumeration. These included plate counting, quantitative real-time PCR, fluorescent in situ hybridization (FISH), and commercial LIVE/DEADs BacLightTM viability assay (L/D). Fermented oat drink (e.g. Molin et al., 1990) was used as a pilot food application. Fermented vegetable-based products such as oat or soy drinks are an emerging group of foods with a healthy image, which offer alternatives to dairy products for those intolerant to milk (Molin, 2001). Probiotic bifidobacteria have received growing interest in food industry, due to their proposed health effects (Ballongue, 2004). In this study three Bifidobacterium strains were used as probiotic model organisms. The aim of this proof-of-principle study was to assess the suitability of different enumeration assays for probiotic bifidobacteria, and to compare the results obtained with these methods. 2. Methods 2.1. Probiotic products Fermented products were prepared by fermenting commercial non-fermented oat drink (‘Oatly’, Ceba Foods AB, Lund, Sweden) with either a commercially available probiotic strain Bifidobacterium lactis Bb-12 (Chr. Hansen A/S, Hørsholm, Denmark), or a combination of two potential probiotic strains, B. longum 2C (DSM 14579) and B. longum 46 (DSM 14583), isolated from faeces of healthy elderly subjects (Laine et al., 2003). Bacteria suspended in 10 mM phosphate buffered saline (PBS; pH 7.2) were inoculated at 1% into oat medium prior to fermentation. The cultures were protected from light, and fermented anaerobically at 37 1C for 22 h. Four replicate fermentations of B. longum and three replicate fermentations of B. lactis were prepared, each fermentation in duplicate. Fermented products were covered from light and stored at 4 1C until use. 2.2. Plate counts A range of nutrient-rich commercial media suitable for anaerobic cultivations were tested to optimize the growth of the bifidobacteria on plates (data not shown). Of these, reinforced clostridial medium (RCM; Oxoid Ltd., Hampshire, UK) was chosen as the enumeration media for

bifidobacteria. Samples of the fermented products were diluted with PBS and plated on RCM plates supplemented with 1.5% agar (Pronadisa, Madrid, Spain). The plates were incubated in anaerobic chamber at 37 1C and enumerated after 3 days. 2.3. FISH FISH was performed as described by Langendijk et al. (1995) with slight modifications. In short, samples of the probiotic products were diluted in PBS and homogenized. Cells were subsequently fixed in a 4% (v/v) paraformaldehyde solution overnight at 4 1C, washed twice, and resuspended in 1 ml of PBS:ethanol (1:1, v/v). A portion of the cell suspension was hybridized overnight at 50 1C in hybridization buffer (HB; 10 mM Tris–HCl, 0.9 M NaCl) with a Cy3 indocarbocyanin-labeled genus-specific probe BIF164 (‘5-CATCCGGCATTACCACCC). Total cell numbers were determined using 40 ,6-diamidino-2-phenylindole (DAPI). Cells were washed with HB, applied to a 0.2 mm polycarbonate filter (Millipore Corporation, Bedford, USA) and mounted on a glass slide. Fifteen microscopic fields were counted per assay. 2.4. Quantitative real-time PCR The quantitative real-time PCR procedure described by Gueimonde et al. (2004) was used for the study of quantification of bacteria in fermented products. The two sets of oligonucleotide primers and probes (Thermo Biosciences, Ulm, Germany) used to quantify B. longum strains and B. lactis are presented in Table 1. Specificity of the set of oligonucleotides used for B. longum has been previously shown (Gueimonde et al., 2004). The specificity of B. lactis oligonucleotides was tested in silico by comparing their sequences with BLAST database search program (www.ncbi.nlm.nih.gov/BLAST) (Altschul et al., 1997) and by using the Probe Match application at the Ribosomal Database Project II (www.rdp.cme.msu.edu/html) (Cole et al., 2003). The oligonucleotide primers were tested by qualitative PCR against an array of different intestinal and food microorganisms (Table 2) as previously described (Gueimonde et al., 2004). For the quantitative real-time PCR analyses 200 ml of the products were used for the DNA extraction by using the QIAamp DNA stool Mini kit (Qiagen, Hilden, Germany). Samples (1 ml) were analysed in 50 ml amplification reactions consisting of 1
PCR buffer II, 3.5 mM MgCl2, 0.2 mM of each primer, 200 mM of each dNTP, 0.016 mM europium-labeled probe, 0.166 mM quencher probe and 1.25 U of AmpliTaq Gold DNA polymerase. All reactions were performed on MicroAmp Optical Plates sealed with MicroAmp Optical Caps (Applied Biosystems). Thermal cycling (iCycler) for the quantification of B. longum strains consisted of an initial cycle of 95 1C 10 min, 40 cycles of 95 1C 15 s, 60 1C 1 min, 61 1C 45 s and 35 1C 15 s

ARTICLE IN PRESS S.J. Lahtinen et al. / Food Microbiology 23 (2006) 571–577

573

Table 1 The oligonucleotides used for the quantification of B. longum (Bifido set) and B. lactis (Animalis set) by real-time PCR Name

Sequence (50 –30 )a

Reference

Bifido50 Bifido30 Bifidoprobe Bifido quencher

GATTCTGGCTCAGGATGAACGC CTGATAGGACGCGACCCCAT CATCCGGCATTACCACCCGTTTCCTCb GTGGTAATGCCGGATGc

Gueimonde et al. (2004)

Animalis5 Animalis3 Aniprobe Aniquencher

ACCAACCTGCCCTGTGCACCG CCATCACCCCGCCAACAAGCT ACCATGCGATGGAGCGGAGCATCCGGTTAb GCTCCATCGCATGGTc

Lahtinen et al. (2005)

a

Bold letters indicate bases that are not complementary to the target. 2,20 ,200 ,2000 {{6,60 -{400 -[2-(4-isothiocyanatophenyl)ethyl]-1H-pyrazole-100 ,300 -diyl}bis(pyridine)-2,20 -diyl}bis(methylenenitrilo)}tetrakis(acetato) europium (III). c Dabcyl. b

Table 2 The strains against which the B. lactis primers were tested Bifidobacterium

Bacteroides Clostridium Enterobacter T

adolescentis JCM 1275T angulatum JCM 7096T animalis JCM 1190T bifidum JCM 1254T breve JCM 1192T catenulatum JCM 7130 denticolens DSM 10105T dentium DSM 20436T gallicum DSM 20093T infantis DSM 20088T lactis Bb-12 longum JCM 1217T pseudocatenulatum JCM 1200T vulgatus DSM 1447T butyricum DSM 10702T coccoides DSM 935T aerogenes DSM 30053T

Enterococcus Escherichia Eubacterium Lactobacillus

Peptostreptococcus Anaerococcus Ruminococcus Streptococcus Veillonella

faecalis DSM 20478T faecium Gaio coli K12 cylindroides DSM 3983T halii DSM 3353T acidophilus La-5 gasseri DSM 20077 jensenii DSM 20557T Casei DSM 20011T paracasei DSM 20244 plantarum 299v rhamnosus GG anaerobius DSM2949T prevotii DSM 20548T hansenii DSM 20583T thermophilus DSM 20617T dispar DSM 20735T

Type strain.

(Gueimonde et al., 2004). For the products containing B. lactis, quantification thermal cycling (iCycler) consisted of the following time and temperature profile: an initial cycle of 95 1C 10 min, 40 cycles of 95 1C 15 s, 67 1C 1 min, 68 1C 45 s and 35 1C 15 s. Europium fluorescence measurements were performed in real-time at the end of each cycle. Calibration curves were determined by comparing the C t values with plate counts of various dilutions of fresh products, and cell counts of the products were determined by comparing their C t values with the calibration curve. Plate counts of fresh samples did not differ from the FISH counts, suggesting that plate counts were suitable for determining the standard curve for PCR assay. Samples were analysed in duplicate in two independent PCR runs. 2.5. Live/dead BacLightTM bacterial viability assay Fermented products were centrifuged at 800g at 4 1C for 7 min in order to separate a less turbid upper fraction

containing bacterial cells. Duplicate samples of the fraction were stained with LIVE/DEADs BacLightTM bacterial viability kit (L/D; Molecular Probes, Leiden, The Netherlands). Green fluorescence of the samples was analysed with a Victor2 multi-label counter using a 515 nm filter (30 nm bandwidth, 25 mm diameter) and 96-well plates (NUNC MaxiSorp, Nunc A/S, Roskilde, Denmark). Samples were also enumerated simultaneously by plate counts. To measure the green fluorescence of dead cells, duplicate control samples were incubated in water bath at 981C for 5 min followed by L/D assay. Microscopic observation revealed that the heat-treatment did not result in cell lysis. In addition to heat treatment, dead cells were also obtained by treating bacteria with acetone at 20 1C as described by Armstrong and He (2001). The fluorescence of intact and heat-treated cells was always measured simultaneously. Heat-killed cells showed no growth on RCM agar plates. The average background fluorescence originating from oat medium was measured by acidifying samples of sterile non-fermented product to the same pH

ARTICLE IN PRESS S.J. Lahtinen et al. / Food Microbiology 23 (2006) 571–577

574

value as the fermented samples, and subjecting the samples to L/D assay. Background fluorescence was determined separately for both heat-treated and untreated samples. To establish relationship between fluorescence and plate counts of fresh samples, a calibration curve was obtained by using suspensions of viable and heat-killed cells mixed in different ratios. A strictly linear relationship (r2 ¼ 0:999) was established between green fluorescence and plate counts of a fresh sample. The FISH counts and plate counts of the fresh samples correlated well, indicating that in the end of the fermentation virtually all cells were viable and culturable. Thus, for the fresh samples, the plate counts represented the amount of living cells in a suspension. It was estimated that the detection limit for green fluorescence of living cells was 5% of the initial green fluorescence of fresh samples. The following equation was formulated for estimating the number of living cells in a sample using L/D assay: cx ¼ ½ðax  bÞ=ða0  bÞ
c0 , in which cx is number of living cells per millilitre after x days of storage; ax is green fluorescence of cells after x days; a0 is green fluorescence of cells after 0 days of storage; b is background fluorescence; and c0 is cfu/ml after 0 days.

storage time and measurable variables such as cfu/ml. Regression analysis (r2) was used as a goodness-of-fit measure of linear models. Paired t-test was used to compare the means of the measured variables. Statistical analysis was performed by using SPSS program ver. 11 (SPSS Inc., Chicago, USA).

2.6. Statistical analysis

FISH counts of both B. lactis and B. longum did not change significantly during storage (Tables 3 and 4). FISH counts were statistically not different from the plate counts of freshly fermented products, confirming that virtually all

Pearson correlation was used to determine the significance and direction of a possible correlation between

3. Results 3.1. Plate counts There was a rapid reduction of 7.4 log units in cfu/ml of the B. longum strains during the storage (Table 3). The correlation between time and loss of culturability was statistically significant (Po0:001). The plate counts for B. lactis decreased slowly and significantly (P ¼ 0:008) by 0.9 log units during storage (Table 4). The magnitude of the loss of culturability was far greater for the B. longum strains than for B. lactis. 3.2. Cell enumeration by fish and quantitative real-time PCR

Table 3 The changes in total numbers of B. longum 2C and B. longum 46 during storage enumerated with four different methods Time (days)

0 7 14 20 25 30

Plate count

PCR

FISH

L/D

Mean

Std. dev.

Mean

Std. dev.

Mean

Std. dev.

Mean

Std. dev.

7.8 6.0 3.4 2.2 1.2 0.4a

0.2 0.5 0.6 0.7 1.4 0.8

7.8 7.7 7.8 7.9 7.7 8.0

0.3 0.2 0.3 0.1 0.0 0.2

7.8 7.9 7.9 8.0 7.8 8.0

0.6 0.2 0.3 0.2 0.0 0.0

7.8 7.6 7.2 7.2 6.9 7.2a

0.2 0.3 0.3 0.2 0.2 0.2

The results are means of four replicate fermentations, each performed in duplicate, and expressed as log cfu/ml. a Statistically significant downward trend (Pearson correlation).

Table 4 The changes in numbers of B. lactis during storage enumerated with four different methods Time (days)

0 7 14 20 25 30

Plate count

PCR

FISH

L/D

Mean

Std. dev.

Mean

Std. dev.

Mean

Std. dev.

Mean

Std. dev.

7.2 6.8 6.8 6.7 6.7 6.3a

0.3 0.7 0.9 1.0 0.6 0.7

7.2 7.2 7.2 7.5 7.2 7.4

0.2 0.0 0.1 0.0 0.1 0.1

7.2 7.2 6.9 7.3 7.1 6.8

0.1 0.0 0.1 0.1 0.0 0.0

7.2 7.1 7.4 7.8 7.8 7.0

0.3 0.5 0.2 0.0 0.0 1.0

The results are means of three replicate fermentations, each performed in duplicate. a Statistically significant downward trend (Pearson correlation).

ARTICLE IN PRESS S.J. Lahtinen et al. / Food Microbiology 23 (2006) 571–577

575

Table 5 The significances of the differences between different enumeration assays (paired Student’s t-test) Strain

B. longum

B. lactis

Method

Plates

PCR

Plates PCR FISH L/D

— P ¼ 0:015 P ¼ 0:016 P ¼ 0:016

— n.s. P ¼ 0:014

FISH

— P ¼ 0:032

L/D

Plates

PCR

FISH

L/D

—

— P ¼ 0:021 P ¼ 0:019 P ¼ 0:015

— n.s. n.s.

— n.s.

—

n.s. ¼ not significant.

bacteria in the fresh samples were viable and culturable. Total bacteria counts obtained by DAPI staining were similar to the counts obtained with probe BIF164, proving that all bacteria in the samples belonged to the Bifidobacterium genus. Real-time PCR counts remained also constant during storage (Tables 3 and 4), indicating that the DNA of the bifidobacteria in the products was intact. No significant difference was observed between FISH counts and real-time PCR counts for B. lactis or for B. longum (Table 5). For B. longum, FISH and PCR results differed significantly from plate counts and L/D assay results, but for B. lactis these counts differed significantly only from plate counts (Table 5). 3.3. Cell viability by live/dead BacLightTM assay Viability results based on L/D assay of the products containing B. longum differed significantly from the results obtained by plate counting, PCR and FISH (Table 5). The green fluorescence of the cells decreased gradually and significantly (P ¼ 0:016) during storage, but the reduction was slow compared to plate counts (Table 3). After 1 month the number of viable cells determined by the L/D assay decreased by 0.6 log units. The green fluorescence of the oat product containing B. lactis did not decrease significantly during storage (Table 4). The L/D counts of B. lactis differed significantly from the plate counts but not from the PCR and FISH counts (Table 5). The average fluorescence for the fixed cells was 3.5% of the initial fluorescence of the fresh cells. Heat-treated cells and acetone-treated cells did not differ from each other statistically. Throughout the study the fluorescence of the untreated cells remained above detection limit and was always higher than the fluorescence of the heat-killed cells. 4. Discussion The target of this study was to assess the suitability of four different methods for the quantification of probiotic bifidobacteria in foods. Traditional plate count was compared with three cytological methods, namely quantitative real-time PCR, fluorescent in situ hybridization (FISH), and LIVE/DEAD BacLightTM (L/D) viability assay. Fermented oat drink was used as a pilot food

application, and strains of Bifidobacterium longum and B. lactis were used as model probiotic organisms. The methods described here are applicable to various food products, provided that the extraction of bacteria and/or bacterial DNA, or labeling of bacterial 16S rRNA can be accomplished. Quantitative real-time PCR is based on quantification of bacterial DNA. It is generally accepted that the DNA levels are not associated with viability, as dead cells may also retain significant amounts of DNA. In comparison, FISH is based on detection of 16S rRNA of the cells. The half-life of rRNA is shorter than that of DNA, making rRNA a potential target molecule for viability assays. Suitability of FISH for viability assays depends on rRNA decay after cell death. The factors affecting FISH ‘positive’ signal include the amount of target 16S rRNA present in cell, accessibility of probe for rRNA, and fluorescence intensity of the bound probe (Smith et al., 2004). The rRNA decay following cell death depends, among other factors, on RNase levels inside and outside the cell, and cell membrane integrity and permeability. McKillip et al. (1998) studied rRNA stability of heat-killed and UV-irradiated Staphylococcus aureus and Escherichia coli, and demonstrated that bacteria retained detectable amounts of rRNA for 48 h after moderate heat treatment but not after severe heat inactivation. Dead cells may possess residual rRNA, and although the intensity of the fluorescent signal of these cells is likely to be decreased after death, the signal may still be strong enough for the cells to be visually counted. In this study, neither the FISH counts nor the real-time PCR counts changed significantly during storage (Tables 3 and 4), but the results of both assays differed significantly from plate counts of both B. lactis and B. longum (Table 5). In addition, FISH and PCR counts of B. longum were also significantly higher than respective L/D counts. These results indicate that real-time PCR counts and FISH counts determined by microscopy were not associated with cell viability. Degradation of DNA had not occurred, and rRNA levels remained high enough for the cells to be detected. Although quantitative real-time PCR and FISH provide only limited information on viability, this study shows that both methods are suitable for the quantification of total cells. There is little variation between the measurements

ARTICLE IN PRESS 576

S.J. Lahtinen et al. / Food Microbiology 23 (2006) 571–577

(Tables 3 and 4), and the results of these two methods are comparable with each other (Table 5). Of the two, real-time PCR is faster but requires more expensive reagents. Downside of this method is that a standard curve (determined for example plate count or FISH) is required. The drawback of FISH is that the method is relatively laborious and time-consuming compared to real-time PCR. The detection limit for quantitative real-time PCR analysis of bifidobacteria is about 104 cells/ml, and for FISH combined with fluorescent microscopy and visual counting it is around 106 cells/ml. FISH may also be combined with flow cytometry, enabling high-throughput and low detection limit (Rigottier-Gois et al., 2003). Both FISH and quantitative real-time PCR are suitable for liquid and solid food applications, provided that tested food samples are homogenous or can be easily homogenized. Foods in which the cells are entrapped in fat may cause problems, as the cells may not suspend readily into aqueous phase during homogenization. In such cases the homogenization buffers can be fortified with a mild detergent or emulsifier to enhance the release of bacteria from food. The L/D assay is based on two fluorescent dyes; the redfluorescent nucleic acid dye propidium iodine and the green-fluorescent nucleic acid stain SYTOs 9. SYTO 9 stains all bacteria regardless of their cell membrane condition, whereas propidium iodine stains only bacteria with damaged membranes (dead bacteria). It is well established that, by maintaining optimal internal conditions for cell metabolism and energy transduction, the cell membrane has a vital function in bacterial survival (Konings et al., 2002). However, this does not necessarily imply that a cell with an intact cell membrane is viable or culturable. In this work the L/D assay was used in combination with a microplate reader. Although this method is not as sensitive as flow cytometric method, it is more affordable and more accessible to many food microbiologists. The advantage of the L/D assay is that it is more rapid than plate count, quantitative real-time PCR, or FISH. It also provides viability information determined by cell membrane condition, unobtainable by other methods. The L/D assay combined with a multi-label counter requires a standard curve to be determined using another method such as plate count or FISH. The use of L/D assay requires the cells to be extracted from food matrix undamaged. For liquid samples this can be easily achieved, although turbidity remaining solid particles may cause problems. In this work a suitable centrifugation procedure was used to overcome the problem. For thick liquids such as yoghurts and for solid foods such as cheeses, the extraction of the bacteria requires the samples to be diluted in buffer and homogenized. The homogenization process has to be thorough enough to release bacteria from food matrix, but gentle enough to keep them alive. In this study, the L/D assay results for B. lactis differed significantly from plate count results. The difference between L/D counts and plate counts was 0.7 log units after 1 month of storage. This suggests that the cell

population which had lost culturability on nutrient agar may still have been viable, or dormant as described by Kell et al. (1998). The L/D counts of B. lactis were comparable to FISH and quantitative real-time PCR results. These results suggest that all B. lactis cells had retained a functional cell membrane and were either culturable or dormant. In the case of B. longum, the difference between plate counts and L/D counts at the end of the storage was substantial (6.8 log units), suggesting that virtually all B. longum cells had lost their culturability, but a remarkable part of the population had retained a functional cell membrane. The L/D counts of B. longum differed from FISH and quantitative PCR counts, but the magnitude of the difference was only 0.8 log units. The cells that were counted viable by neither plate counts nor L/D assay were considered dead. Thus, at the end of the storage, B. longum population consisted of dead cells, dormant cells, and a small number of culturable cells. Quality of a probiotic product is often determined by the level and viability of the probiotic cells it contains. It has been proposed that for a probiotic product to have beneficial effects on human health, it should contain 108–109 living cells per serving. There is a growing need for regulation and legislation for the quality of probiotic products. Before official requirements for the microbial quality of probiotic products are set, how to determine cell culturability and viability in these products should be discussed first. Plate count has maintained its position as the most common method to monitor bacterial numbers in foods, including probiotic products. For easily culturable strains plate count is a justified and reliable assay of viability. However, this study shows that for certain bacteria plate counts may not provide reliable information on the status of the bacteria, as some bacteria may enter a dormant state during storage. The clinical impact of dormant probiotic cells has not been determined. Such studies are urgently needed for assessing the efficacy of probiotics in food products during storage. In conclusion, the choice of the enumeration method for probiotic bacteria should depend on the properties of the probiotic strains, as well as the objectives of the analysis. This study provides a comparison between potential methods of enumeration for probiotic bacteria, and describes the methodology required for these assays. The results of this work confirm findings of the recent pilot study by Lahtinen et al. (2005), and emphasizes the need for further assessment of the methods applied to determine viability of probiotic products. The possibility of probiotic cells being dormant and not readily culturable plate counting should be considered when suitable enumeration methods are selected. Acknowledgements We thank Riikka Parhiala for her skillful assistance. S. Lahtinen was supported by the Applied Biosciences Graduate School.

ARTICLE IN PRESS S.J. Lahtinen et al. / Food Microbiology 23 (2006) 571–577

References Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller, W., Lipman, D.J., 1997. Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25, 3389–3402. Armstrong, D.W., He, L., 2001. Determination of cell viability in single or mixed samples using capillary electrophoresis laser-induced fluorescence microfluidic systems. Anal. Chem. 73, 4551–4557. Ballongue, J., 2004. Bifidobacteria and probiotic action. In: Salminen, S., Von Wright, A., Ouwehand, A. (Eds.), Lactic Acid Bacteria— Microbiological and Functional Aspects. Marcel Dekker, Inc., New York, pp. 67–125. Barer, M.R., Smith, R.J., Cooney, R.P., Kimmitt, P.T., 2000. Relationships between culturability, activity and virulence in pathogenic bacteria. J. Infect. Chemother. 6, 108–111. Ben Amor, K., Breeuwer, P., Verbaarschot, P., Rombouts, F.M., Akkermans, A.D., de Vos, W.M., Abee, T., 2002. Multiparametric flow cytometry and cell sorting for the assessment of viable, injured, and dead Bifidobacterium cells during bile salt stress. Appl. Environ. Microbiol. 68, 5209–5216. Bloomfield, S.F., Stewart, G.S., Dodd, C.E., Booth, I.R., Power, E.G., 1998. The viable but non-culturable phenomenon explained? Microbiology 144, 1–3. Bunthof, C.J., Abee, T., 2002. Development of a flow cytometric method to analyze subpopulations of bacteria in probiotic products and dairy starters. Appl. Environ. Microbiol. 68, 2934–2942. Cole, J.R., Chai, B., Marsh, T.L., Farris, R.J., Wang, Q., Kulam, S.A., Chandra, S., McGarrell, D.M., Schmidt, T.M., Garrity, G.M., Tiedje, J.M., 2003. The Ribosomal Database Project (RDP-II): previewing a new autoaligner that allows regular updates and the new prokaryotic taxonomy. Nucleic Acids Res. 31, 442–443. Colwell, R.R., 2000. Viable but nonculturable bacteria: a survival strategy. J. Infect. Chemother. 6, 121–125. Dodd, C.E., Sharman, R.L., Bloomfield, S.F., Booth, I.R., Stewart, G.S., 1997. Inimical process: Bacterial self-destruction and sub-lethal injury. Trends Food Sci. Technol. 8, 238–241. Gueimonde, M., To¨lkko¨, S., Korpima¨ki, T., Salminen, S., 2004. New real-time quantitative PCR procedure of quantification of Bifidobacterium in human fecal samples. Appl. Environ. Microbiol. 70, 4165–4169. Kell, D.B., Kaprelyants, A.S., Weichart, D.H., Harwood, C.R., Barer, M.R., 1998. Viability and activity in readily culturable bacteria: a

577

review and discussion of the practical issues. Antonie von Leeuwenhoek 73, 169–187. Konings, W.N., Albers, S.V., Koning, S., Driessen, A.J., 2002. The cell membrane plays a crucial role in survival of bacteria and archaea in extreme environments. Antonie von Leeuwenhoek 81, 61–72. Lahtinen, S.J., Gueimonde, M., Ouwehand, A., Reinikainen, J.P., Salminen, S., 2005. Probiotic bacteria may become dormant during storage. Appl. Environ. Microbiol. 71, 1662–1663. Laine, R., Salminen, S., Benno, Y., Ouwehand, A.C., 2003. Performance of bifidobacteria in oat-based media. Int. J. Food Microbiol. 83, 105–109. Langendijk, P.S., Schut, F., Jansen, G.J., Raangs, G.C., Kamphuis, G.R., Wilkinson, M.H., Welling, G.W., 1995. Quantitative fluorescence in situ hybridization of Bifidobacterium spp. with genus-specific 16S rRNA-targeted probes and its application in fecal samples. Appl. Environ. Microbiol. 61, 3069–3075. Martineau, B., 1999. Comparison of four media for the selection of bifidobacteria in dog faecal samples. Anaerobe 5, 123–127. McKillip, J.L., Jaykus, L.A., Drake, M., 1998. rRNA stability in heatkilled and UV-irradiated enterotoxigenic Staphylococcus aureus and E. coli O157:H7. Appl. Environ. Microbiol. 64, 4264–4268. Molin, G., 2001. Probiotics in foods not containing milk or milk constituents, with special reference to Lactobacillus plantarum 299v. Am. J. Clin. Nutr. 73, 380S–385S. Molin, N., Albertsson, C.E., Bengmark, S., Larsson, K., 1990. US Patent 5,190,755. Nystrom, T., 2001. Not quite dead enough: on bacterial life, culturability, senescence, and death. Arch. Microbiol. 176, 159–164. Payne, J.F., Morris, A.E., Beers, P., 1999. Note: evaluation of selective media for the enumeration of Bifidobacterium sp. in milk. J. Appl. Microbiol. 86, 353–358. Rigottier-Gois, L., Le Bourhis, A.-G., Gramet, G., Rochet, V., Dore, J., 2003. Fluorescent hybridisation combined with flow cytometry and hybridisation of total RNA to analyse the composition of microbial communities in human faeces using 16S rRNA probes. FEMS Microbiol. Ecol. 43, 237–245. Smith, J.J., Gunasekera, T.S., Barardi, C.R., Veal, D., Vesey, G., 2004. Determination of Cryptosporidium parvum oocyst viability by fluorescence in situ hybridization using a ribosomal RNA-directed probe. J. Appl. Microbiol. 96, 409–417. Yamamoto, H., 2000. Viable but nonculturable state as a general phenomenon of non-spore-forming bacteria, and its modeling. J. Infect. Chemother. 6, 112–114.