Rapid estimation of spoilage bacterial load in ... - Wiley Online Library

Journal of Applied Microbiology 1997, 82, 359–364

Rapid estimation of spoilage bacterial load in aerobically stored meat by a quantitative polymerase chain reaction K.S. Venkitanarayanan, C. Faustman, J.F. Crivello1, M.I. Khan2, T.A. Hoagland and B.W. Berry3 Departments of Animal Science, 1Physiology and Neurobiology, and 2Pathobiology, University of Connecticut, Storrs, CT, and 3USDA, Agriculture Research Service, Beltsville, MD, USA 5774/05/96: received 24 May 1996, revised 18 August 1996 and accepted 20 August 1996 K .S . V E NK IT A NA RA Y AN AN , C. FA U ST MA N , J .F . CR IV E LL O, M .I . K H AN , T . A. HO A GL AN D AN D

We report a quantitative PCR which utilizes primers from a conserved 23S rDNA sequence identified in nine different spoilage bacteria commonly present in meat. The PCR detected the spoilage bacteria by amplifying a specific 207 bp sequence from their chromosomal DNA. Quantification of PCR product by electrochemiluminescence revealed that the concentration of the amplified product was dependent on cycle number and the initial number of bacteria present in the sample. Statistical analysis of the results indicated a correlation coefficient of 0·94 (P ³ 0·001) between aerobic plate count and QPCR luminosity units. B .W . B E RR Y. 1997.

INTRODUCTION

Meat is contaminated by spoilage bacteria during slaughter and subsequent processing operations. The growth of spoilage bacteria reduces shelf-life of meat and the annual economic loss from spoilage of fresh meat and meat products in the United States is estimated at approximately 5 billion dollars (Ray et al. 1992). In addition, high initial bacterial loads in fresh meat indicate lower standards of slaughter and plant hygiene. Spoilage is a function of the total number of bacteria present in meat and determination of total bacterial load is the most direct approach for predicting spoilage status of meat and its potential shelf-life. However, the traditional standard plate count method requires 48 h for results and the bacterial load in meat can increase considerably during this time. Development of rapid methods to determine bacterial load would help predict bacteriological quality and spoilage status of meat in a shorter time and would also serve to validate the efficacy of various antimicrobial procedures for minimizing or eliminating growth of bacteria, thus helping in the implementation of Sanitation Standard Operating Procedures in the meat industry. The polymerase chain reaction (PCR) is a powerful molecCorrespondence to: Dr C. Faustman, Department of Animal Science, Box U-40, University of Connecticut, 3636 Horsebarn Hill Road Ext., Storrs, CT 06269, USA. © 1997 The Society for Applied Bacteriology

ular technique used for the rapid and sensitive detection of micro-organisms (Candrian 1995; Hill 1996). Although PCR has been used for the specific and rapid detection of several food-borne pathogens, relatively little research has been performed in developing a PCR-based assay for the detection of spoilage bacteria. In addition, PCR has been primarily a qualitative technique; the ability to accurately quantify a target sequence would be advantageous. The most common bacteria typically associated with spoilage of aerobically stored meat include Pseudomonas fluorescens, Ps. putida, Ps. fragi, Ps. aureofaciens, Acinetobacter calcoaceticus, Enterobacter liquefaciens, Flavobacterium spp., Moraxella spp. and Brochothrix thermosphactum (Gill and Newton 1978; Nortje et al. 1990; Lambert et al. 1991). Recently, we developed a 23S rDNA-based PCR which detected all these spoilage bacteria (Venkitanarayanan et al. 1996). However, in order to predict the spoilage status of meat, an estimation of bacterial load is necessary. Quantitative PCR (QPCR) is a relatively new technique that detects and quantifies target nucleic acid sequences. The basic principle of QPCR involves the use of chemically labelled primers in the PCR and quantifying the labelled product resulting from subsequent amplification (Jessen-Eller et al. 1994). One of the primers is labelled with a molecule (e.g. biotin) which serves as a specific substrate for binding to another known compound (e.g. streptavidin) so that the amplified product can be readily con-

360 K .S . V E NK IT A NA RA Y AN AN E T A L.

centrated. The other primer is labelled with a chemiluminescent reporter molecule (tris (2,2?-bypyridine) ruthenium chelate, TBR). The PCR product labelled with both molecules is then captured and quantified by chemiluminescence. Sninsky and Kwok (1993) developed a QPCR for rapid quantification of human immunodeficiency virus (HIV) RNA in the plasma or sera of acquired immune deficiency syndrome (AIDS) patients, as a tool for monitoring the efficacy of therapeutic intervention. A similar approach was reported by Ferre et al. (1993), who developed a QPCR for the quantification of HIV-1 DNA in the blood cells of AIDS patients undergoing immunotherapeutic treatment. Recently, Secchiero et al. (1995) reported a QPCR-based approach for determining the degree of Herpes virus infection in human patients. The objective of the present study was to determine the potential of QPCR for rapid estimation of spoilage bacterial load in aerobically stored meat. MATERIALS AND METHODS Bacterial culture

Pseudomonas fluorescens, a common spoilage bacterial species found on aerobically stored meat (Nortje et al. 1990; Lambert et al. 1991), was used to ascertain the potential of QPCR for estimating bacterial load. A pure culture of Ps. fluorescens (NRRL B-10, National Centre for Agricultural Utilization, Peoria, IL, USA) was inoculated into 50 ml of sterile nutrient broth (Difco, Detroit, MI, USA) and incubated overnight for 18 h at 30°C on a rotary shaker at 250 rev min−1. After this, 1 ml of the culture was serially diluted in 9 ml of sterile 0·1% peptone water (diluent). The aerobic bacterial count of the culture was determined by a spread plate technique (Steinbrugge and Maxcy 1988) using plate count agar (Difco, Detroit, MI, USA). Volumes (25 ml) of the inoculum and each dilution were transferred into separate microfuge tubes and frozen immediately at −70°C. As needed, samples were thawed at 25°C, heated at 94°C for 7 min and then cooled rapidly in ice. Finally, the samples were subjected to PCR. Meat samples

Meat samples for the experiment included semimembranosus beef steaks obtained from a local retail market and stored at 4°C for 2, 4, 6 and 8 d. On each sampling day, a 25 cm2 surface area on each steak was swabbed twice with two separate, sterile cotton swabs. The swabs were subsequently washed into a tube containing 4·5 ml of sterile 0·1% peptone water (diluent). Samples (25 ml) of the inoculated peptone water were frozen immediately at −70°C. Prior to PCR, the samples were thawed at 25°C, heated at 94°C for 7 min, and subjected to PCR. The aerobic bacterial counts of the meat

samples were determined by spread plate technique (Steinbrugge and Maxcy 1988). On each day of the experiment, 25 ml volumes of a pure culture of Ps. fluorescens NRRL B-10 with known bacterial number were also subjected to PCR simultaneously with meat samples for use as an external standard. Sterile peptone water (0·1%) and calf thymus DNA were used as negative controls. Polymerase chain reaction

Primers which amplify a conserved 207 bp sequence in the 23S rDNA of major meat spoilage bacteria were used in the PCR (Venkitanarayanan et al. 1996). PCR was carried out as described by Saiki et al. (1988) using reagents from a GeneAmp PCR kit (Perkin Elmer Cetus, Norwalk, CT, USA). A reaction volume of 100 ml of PCR mixture contained 1 × PCR buffer; 200 mmol l−1 each of dATP, dCTP, dGTP and dTTP; 2·5 mmol l−1 MgCl; 0·3 mmol l−1 of each primer, and 1 unit of AmpliTaq DNA polymerase. PCR was performed in an automatic DNA thermal cycler (Model 9600, Perkin Elmer Cetus, Norwalk, CT, USA) under the conditions reported by Venkitanarayanan et al. (1996). Each cycle consisted of a melting temperature of 92°C for 30 s, annealing temperature of 56°C for 30 s, and extension temperature of 74°C for 45 s. Detection of PCR products

Gel electrophoresis was used to detect the amplified PCR products in each of the samples. The PCR products in 10 ml were separated on a 1·5% agarose gel, stained with ethidium bromide, and exposed to u.v. light to confirm the presence and size of the amplified DNA product. Quantification of PCR products

Quantification of PCR products was carried out according to Jessen-Eller et al. (1994). The forward primer was labelled with biotin and the reverse primer with TBR. The amplified product was then captured using streptavidin-coated magnetic beads and quantified by chemiluminescence of TBR. In the QPCR experiments with pure culture of Ps. fluorescens, a 5 ml volume of the sample was removed at cycles zero, 14, 18, 22, 26, 28, 30, 32 and 34, while with samples from meat, 5 ml were removed at cycles zero, 22, 24, 26 and 28. Each was preserved in 100 ml of 1 × PCR buffer. At the end of PCR, 20 ml of streptavidin-coated magnetic beads (Perkin Elmer Cetus) were added to each sample and incubated for 10 min at 25°C. The samples were transferred to separate tubes each containing 0·4 ml of QPCR assay buffer (Perkin Elmer Cetus) and analysed using a QPCR System 5000 (Perkin Elmer Cetus). The concentration of the PCR-amplified product was reported in luminosity units. The entire experiment was repeated three times.

© 1997 The Society for Applied Bacteriology, Journal of Applied Microbiology 82, 359–364

E ST IM A TI ON O F S PO I LA GE B AC TE R IA 361

STATISTICAL ANALYSIS

Data were analysed by analysis of variance using the General Linear Models procedure of SAS Institute, Inc. (1985). A linear regression was performed with bacterial load as the independent variable and luminosity units as the dependent variable and the correlation coefficient (r-value) tested for significance at P ³ 0·01. RESULTS

The forward primer and reverse primer used in the PCR directed the amplification of a 207 basepair sequence from the pure culture of Ps. fluorescens (Venkitanarayanan et al. 1996). No amplification was observed in 0·1% peptone water (negative control) and the mean luminosity units ranged from 31 to 45 from cycle zero to 34, respectively. Quantification of PCR-amplified product from serial dilutions of Ps. fluorescens culture revealed ‘sigmoid’-shaped curves typical of PCR samples (Fig. 1). It was also observed that PCR product concentration was a function of the initial number of bacterial cells in the sample (Fig. 1). Comparison of luminosity units of samples at any given cycle (above cycle zero) indicated that the luminosity units of samples containing higher bacterial loads were greater than those with lower bacterial populations. In addition, it was found that the PCR product concentration from samples containing 106 and 107 cfu per 25 ml plateaued after 28 cycles of amplification (Fig. 1).

The results of QPCR experiments involving meat samples stored at 4°C for 2, 4, 6 and 8 d are presented in Fig. 2. As with the Ps. fluorescens culture, for a given cycle, the luminosity units of PCR products from meat samples increased with increased aerobic plate count of the samples (Fig. 2). The mean luminosity units of the PCR products from calf thymus DNA (negative control) were low throughout amplification, ranging from 11 to 56 as the PCR progressed from cycle zero to cycle 28. It was also observed that PCR product concentrations from meat samples and Ps. fluorescens standards containing comparable bacterial loads were similar at any given cycle (Fig. 3). The luminosity units produced by the PCR products from meat stored for 2 d (2·95 × 10 cfu per 25 ml diluent) and 4 d (5·75 × 102 cfu per 25 ml diluent) were consistently lower than those of Ps. fluorescens standards (2·00 × 103 cfu per 25 ml diluent, 2·00 × 105 cfu per 25 ml diluent, and 2·00 × 106 cfu per 25 ml diluent) and meat samples containing higher bacterial loads. However, meat samples stored at 4°C for 6 d (1·00 × 105 cfu per 25 ml diluent) and 8 d (1·30 × 106 cfu per 25 ml diluent) demonstrated comparable luminosity units to Ps. fluorescens standards of 2·00 × 105 cfu per 25 ml diluent, and 2·00 × 106 cfu per 25 ml diluent, respectively. These results indicated that PCR product concentration at a given cycle was dependent on bacterial number

10 000

14 000

8000

Luminosity units

Luminosity units

12 000 10 000 8000 6000

6000

4000

2000

4000 0

2000 0

0

0

14

18

22 26 28 30 PCR cycle number

32

34

Fig. 1 Luminosity units of PCR products from serially diluted

Pseudomonas fluorescens culture. ž, Undiluted culture (3·16 × 107 cfu per 25 ml; Ž, 1:10 diluted culture; R, 1:102 diluted culture; E, 1:103 diluted culture; , 1:104 diluted culture; , 1:105 diluted culture; r, 1:106 diluted culture; E, sterile 0·1% peptone water (negative control)

22 24 PCR cycle number

26

28

Fig. 2 Luminosity units of PCR products from

semimembranosus beef steaks stored at 4°C for 2, 4, 6 and 8 d. ž, Day 2 meat (2·95 × 10 cfu per 25 ml diluent, 2·12 × 102 cfu cm−2); Ž, day 4 meat (5·75 × 102 cfu per 25 ml diluent, 4·14 × 103 cfu cm−2); E, day 6 meat (1·00 × 105 cfu per 25 ml diluent, 7·20 × 105 cfu cm−2); , day 8 meat (1·30 × 106 cfu per 25 ml diluent, 9·30 × 106 cfu cm−2); R, 25 ml diluent containing calf thymus DNA (negative control)



12 000

10 000

Luminosity units

8000

6000

4000

2000

0

0

22 24 PCR cycle number

26

28

Fig. 3 Comparison of luminosity units of PCR products from

semimembranosus beef steaks stored at 4°C for 2, 4, 6 and 8 d, and Pseudomonas fluorescens standards. ž, Day 2 meat (2·95 × 10 cfu per 25 ml diluent, 2·12 × 102 cfu cm−2); E, day 4 meat (5·75 × 102 cfu per 25 ml diluent, 4·14 × 103 cfu cm−2); R, day 6 meat (1·00 × 105 cfu per 25 ml diluent, 7·20 × 105 cfu cm−2); Ž, day 8 meat (1·30 × 106 cfu per 25 ml diluent, 9·30 × 106 cfu cm−2); E, Ps. fluorescens standard (2·00 × 103 cfu per 25 ml culture); r, Ps. fluorescens standard (2·00 × 105 cfu per 25 ml culture); , Ps. fluorescens standard (2·00 × 106 cfu per 25 ml culture)

in the samples, and was comparable for meat samples and standards containing similar bacterial loads. DISCUSSION

Traditionally, PCR products have been quantified by gel electrophoresis or by densitometry utilizing radioactive isotopes. However, these methods are semiquantitative at best. Electrochemiluminescence-based detection and quantification of PCR products is a relatively new method which permits sensitive measurement of PCR products during the exponential phase of PCR amplification (Wages et al. 1993). PCR-based detection of pathogenic bacteria in foods involves the use of primers specific for each pathogen. This specificity is critical to identify the pathogen from the standpoint of therapeutic treatment and epidemiological investigations of a food-borne disease outbreak. However, food spoilage may involve simultaneous growth of many spoilage bacteria; thus it is more logical to develop a PCR which quantitatively detects all or a majority of spoilage bacteria as a group. Therefore, we selected primers for the PCR from a conserved 23S rDNA sequence of Ps. aeruginosa (Festl et al. 1986). The

primers were previously shown to amplify a common 207 bp sequence from the DNA of nine spoilage bacteria common in aerobically stored meat (Venkitanarayanan et al. 1996). Since PCR involves exponential amplification of a target sequence, the amount of template in the starting material determines the rate at which product concentration reaches a plateau phase. PCR product from a sample with high bacterial load will plateau earlier than from a sample with lower bacterial load. In experiments with a pure culture of Ps. fluorescens, we found that product concentration from samples with high bacterial load (106 cfu per 25 ml and higher) showed a tendency to plateau after 28 cycles of amplification (Fig. 1). This was critical because accurate quantitative determination of product cannot be obtained by measuring the product concentration at ‘plateau’ (Lantz 1995). Therefore, a maximum of 28 cycles was used for PCR with meat samples. In addition, we also selected cycles zero, 22, 24, 26 and 28 as the sampling points for potential quantification of the product. Quantitative PCR is commonly carried out using standards for monitoring variability in the assay that may be attributed to differences in reaction components, protocols or instrumentation (Sninsky and Kwok 1993; Seccheiro et al. 1995). Pseudomonas fluorescens is the most common spoilage bacterial species present in aerobically stored meat and was used as an external standard. In experiments with meat samples, Ps. fluorescens standards with known bacterial load were included in every PCR analysis. This accounted for the potential dayto-day variations in the experiments and facilitated quantitative determination of PCR product concentration in the meat samples by comparing product concentration between samples and standards with comparable bacterial loads. Detectable meat spoilage generally occurs when the spoilage bacterial load reaches 107 cfu cm−2 (Ayres 1960; Perez de Castro et al. 1988). Furthermore, meat with a bacterial load of 106−107 cfu cm−2 has a limited shelf-life but could pass sensory tests. One potential application of QPCR would be to use it as a rapid cut-off method to accept or reject meat samples relative to a bacterial load of 106 cfu cm−2. In order to accomplish this, the luminosity units at a particular cycle corresponding to a bacterial load cut-off for accepting or rejecting meat would have to be selected. In the present study, luminosity units at cycle 24 would be appropriate since a plateauing of PCR product concentration was observed after cycle 24 in meat samples with −106 cfu per 25 ml diluent (Fig. 2). In addition, the luminosity units of meat samples with 106 cfu per 25 ml diluent and those of standards containing 106 cfu per 25 ml culture were comparable at cycle 24 within the exponential phase of the reaction (Fig. 3). The PCR product from meat samples stored for 8 d (1·3 × 106 cfu per 25 ml diluent) demonstrated a mean luminosity of 8955 2 826 units which was higher (P ³ 0·001) than that of meat samples with a bacterial load ³106 cfu per 25 ml diluent


E ST IM A TI ON O F S PO I LA GE B AC TE R IA 363

10 000

Army Natick RD & E Centre, MA, USA. Thanks are expressed to Drs Gary Shults, Frank DiLeo, Curtis Blodgett and Richard Worfel for their assistance. The authors also thank the Storrs Agricultural Experiment Station, College of Agriculture and Natural Resources, University of Connecticut. Scientific contribution no. 1709, Storrs Agricultural Experiment Station.

Luminosity units

r = 0·94

1000

REFERENCES

100

10 1

2

3

4 5 –2 Log cfu cm

6

7

8

Fig. 4 Relationship between luminosity units of PCR products

at cycle 24 and aerobic plate count of semimembranosus beef steaks

(Fig. 2). These results indicated that QPCR could potentially be used as a rapid method to identify meat samples with a bacterial load of 106 cfu cm−2 and higher. A QPCR approach for enumerating bacteria in meat could also examine loads in the range of 102−107 cfu cm−2. Depending upon the desired level of sensitivity for quantification, the QPCR can be optimized in terms of number of PCR cycles, sampling frequency for quantification, and use of standards containing bacterial loads comparable to the range of detection required. A linear regression involving the bacterial load of meat samples and luminosity units at cycle 24 yielded a correlation coefficient (r-value) of 0·94 (P ³ 0·01) (Fig. 4). Interest in rapid methods for detection and enumeration of food-borne micro-organisms has been increasing during the last decade (Fung 1995). QPCR requires 4–5 h for results and this study indicated that QPCR has potential use as a rapid method for estimation of spoilage bacterial load in aerobically stored meat. The high sensitivity coupled with its quantitative ability renders QPCR a promising method for rapid detection and quantification of food-borne microorganisms. This is especially significant in the case of foodborne pathogens, as there has been an increasing debate concerned with the use of risk-based approaches rather than zero tolerance for addressing the presence and number of pathogens in foods (Spalding 1995). This will necessitate a means for rapid quantitative detection of specific pathogens in foods. ACKNOWLEDGEMENTS

Support for this research was provided by the United States Department of Agriculture, ARS, Beltsville, MD, and US

Ayres, J.C. (1960) Temperature relationships and some other characteristics of the microflora developing on refrigerated beef. Journal of Food Research 25, 1–18. Candrian, U. (1995) Polymerase chain reaction in food microbiology. Journal of Microbiology Methods 23, 89–103. Ferre, F., Marchese, A.L., Griffin, S.L., Daigle, A.E., Richieri, S.P., Jensen, F.C. and Carlo, D.J. (1993) Development and validation of a polymerase chain reaction method for the precise quantitation of HIV-1 DNA in blood cells from subjects undergoing a 1-year immunotherapeutic treatment. AIDS 7, S21–S27. Festl, H., Ludwig, W. and Schleifer, K.H. (1986) DNA hybridization probe for Pseudomonas fluorescens group. Applied and Environmental Microbiology 52, 1190–1194. Fung, D.Y.C. (1995) What’s needed in rapid detection of foodborne pathogens. Food Technology 49, 64–67. Gill, C.O. and Newton, K.G. (1978) The ecology of bacterial spoilage of fresh meat chill temperatures. Meat Science 2, 207–217. Hill, W.E. (1996) The polymerase chain reaction: applications for the detection of foodborne pathogens. Critical Reviews in Food Science Nutrition 36, 123–173. Jessen-Eller, K., Picozza, E. and Crivello, J.F. (1994) Quantitation of metallothionein mRNA by RT-PCR and chemiluminescence. Biotechniques 17, 962–973. Lambert, A.D., Smith, J.P. and Dodds, K.L. (1991) Shelf-life extension and microbiological safety of fresh meat—a review. Food Microbiology 8, 267–297. Lantz, O. (1995) Chemiluminescent PCR-ELISA assays for quantitative PCR using a kinetic approach. Luminescence Topics 4, 4– 5. Nortje, G.L., Nel, L., Jordaan, E., Badenhorst, K., Goedhart, E. and Holzapfel, W.H. (1990) The aerobic psychrotrophic populations on meat and meat contact surfaces in a meat production system and on meat stored at chill temperatures. Journal of Applied Bacteriology 68, 335–344. Perez de Castro, B., Asensio, M., Sans, B. and Ordonez, J.A. (1988) A method to assess the bacterial content of refrigerated meat. Applied and Environmental Microbiology 54, 1462–1465. Ray, B., Kalchayanand, N. and Field, R.A. (1992) Meat spoilage bacteria: are we prepared to control them? National Provisioner 206, 22. Saiki, R.K., Gelfand, D.H., Staffle, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B. and Erlich, H.A. (1988) Primerdirected enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239, 487–491. SAS Institute (1985) SAS User’s Guide: Statistics. SAS Institute, Inc., Cary, NC, USA.



Secchiero, P., Zella, D., Crowley, R.W., Gallo, R.C. and Lusso, P. (1995) Quantitative PCR for human Herpesviruses 6 and 7. Clinical Microbiology 33, 2124–2130. Sninsky, J.J. and Kwok, S. (1993) The application of quantitative polymerase chain reaction to therapeutic monitoring. AIDS 7, S29–S34. Spalding, B.J. (1995) Better tests are needed to meet goal of safer seafood supply. American Society Microbiology News 61, 639–641.

Steinbrugge, E.G. and Maxcy, R.B. (1988) Nature and number of ground-beef microorganisms capable of growth at 25°C but not at 32°C. Journal of Food Protection 51, 176–178. Venkitanarayanan, K.S., Khan, M.I., Faustman, C. and Berry, B.W. (1996) Detection of meat spoilage bacteria by using the polymerase chain reaction. Journal of Food Protection 59, 845–848. Wages, J.M., Dolenga, L. and Fowler, A.K. (1993) Amplifications: a forum for PCR users. The Perkin Elmer Corporation 10, 1–7.
