- Email: [email protected]
Utilization of substrates by bacterial communities (biofilm) as they develop on stored chicken meat samples
Utilization of Substrates by Bacterial Communities (Biofilm) as They Develop on Stored Chicken Meat Samples1 D. D. HALE BOOTHE,* J. W. ARNOLD,*,2 and V. CHEW† USDA, *Richard B. Russell Agricultural Research Center, P.O. Box 5677, Athens, Georgia 30604, and †Biometrical Services, P.O. Box 14565, Gainesville, Florida 32604 ABSTRACT Understanding and controlling the metabolic processes of microorganisms associated with chicken meat can lead to safer poultry products with a longer shelf life. The objective of the present study was threefold: 1) to determine the feasibility of using 96-well Biolog GN microtiter plates to assess substrate utilization profiles of bacterial communities (biofilm) as they develop on poultry products, 2) to identify substrates metabolized by microbial populations associated with stored chicken meat, and 3) to compare the substrate utilization profiles of biofilm communities as they develop on meat stored at 4 C (refrigeration temperature) for up to 5 d or at 13 C (a temperature common in poultry processing areas) for 2 d. The protocol used herein for preparing inocula for microplates was acceptable for the collection of optical density values (590 nm) in microplate wells as an indicator of microbial substrate utilization over time. Data from treatment of chicken meat samples using this protocol indicate that most of the 95 substrates tested were metabolized by microbial communities present as
early as 1 d after storage at 4 or at 13 C. However, the rapidity (incubation time required for initial substrate utilization) and frequency (percentage of plates positive for transformation of an individual substrate) of metabolism of the substrates by the biofilm communities varied from 4 to 164 h of plate incubation and from 17 to 100% of microplates, respectively. At 13 C, polymers were the most rapidly metabolized substrate group, followed by carbohydrates, carboxylic acids, miscellaneous or amino acids, and amides or amines. Initial utilization of these substrate groups at 4 C occurred within a consistently shorter period (24 h of plate incubation). The frequency of metabolism of each individual substrate group varied only 3 to 16% between samples stored at 4 and 13 C. However, a greater difference in frequency of utilization of some individual substrates was noted. Such divergences may be useful in characterizing biofilm communities implicated in pathogenicity or affecting food quality of poultry products.
(Key words: poultry, biofilms, bacteria, substrate utilization, chicken meat) 1999 Poultry Science 78:1801–1809
INTRODUCTION The safety of our food supply can be immediately improved by reducing the potential for foodborne illnesses caused by bacterial contamination of food products. Growth of bacteria on meat, surrounding fluid, and inanimate surfaces of the processing plants leads to product contamination and spoilage. Substantial bacterial contamination of poultry products in the processing and storage environment is composed of many different species of microbes in a biofilm community (Arnold, 1998). Members of the biofilm community share and compete for
Received for publication April 5, 1999. Accepted for publication September 1, 1999. 1 This article is a US Government work, and, as such, is in the public domain in the United States of America. Reference to a company name or product does not imply endorsement by the US Department of Agriculture. 2 To whom correspondence should be addressed: email: jarnold@ ars.usda.gov. 3 Biolog, Inc., Hayward, CA 94545.
surface attachment sites, light, and carbon and energy sources. Although traditional microbiological studies have been based on pure cultures of one bacterial species, the behavior of a single species can change when mixed with other bacteria with different properties (Arnold and Shimkets, 1988). Recently, methods have been developed to assess the biological and chemical properties of biofilm communities as a whole. Garland and Mills (1991) introduced the concept of community-level physiological profiles using Biolog3 GN microtiter plates for characterizing microbial communities in environmental samples. Several researchers have applied this approach to the analysis of microbial communities in a variety of environments, including soil, water, and wastewater (Garland and Mills, 1991; Haack et al., 1994; Guckert et al., 1996; Victorio et al., 1996; Buyer and
Abbreviation Key: ctemp = chicken meat storage temperature; itemp = microtiter plate incubation temperature; den = density of inoculum on the microtiter plate.
1801
1802
HALE BOOTHE ET AL.
Drinkwater, 1997). In this method, the metabolic capabilities of mixed populations of microorganisms are determined by their ability to utilize each of 95 substrates at predetermined concentrations commonly used to identify Gram-negative bacteria. Buyer and Drinkwater (1997) recently categorized these substrates into six major groups: carbohydrates (30 substrates), carboxylic acids (24 substrates), polymers (5 substrates), amines or amides (6 substrates), amino acids (20 substrates), and miscellaneous (10 substrates, including phosphorylated compounds). The ability of microbial communities to utilize these 95 substrates on Biolog3 GN microplates is assessed using the following procedure: 1) preparation of a suspension of the microorganisms from the sample of interest, 2) inoculation of the organisms onto a 96-well microtiter plate containing a different dehydrated substrate in each of 95 of the wells but no substrate in the control well, 3) incubation of the plate at an appropriate temperature, and 4) determination of the optical density at 590 nm in each of the wells of the plate. The last step quantitates reduction of the tetrazolium violet dye included in each well of the microplate. The reduced dye, which absorbs at 590 nm, is a purple color and is produced by substrate transformation (transfer of reducing equivalents from the substrate to the dye). The pattern of purple color development on the microtiter plate can subsequently be used to compare the metabolic capabilities of mixed populations of microorganisms. In the present study, the concept of community-level physiological profiles was applied to microbial populations in foods. Previously, Biolog3 GN microtiter plates have been used solely in food microbiology for the purpose of identification of individual organisms (Fernandez et al., 1993; Johnson and Lattuada, 1993; Ellender et al., 1994; Russell et al., 1996; Praphailong et al., 1997) rather than for the characterization of microbial populations associated with pathogenicity or food quality. The objectives in this research were threefold: 1) to determine the feasibility of using 96-well microtiter plates to assess substrate utilization profiles of bacterial communities (biofilm) as they develop on poultry products, 2) to identify substrates utilized by microbial populations from stored chicken meat, and 3) to compare the substrate utilization patterns of biofilm communities as they develop on meat stored up to 2 or 5 d at 4 C (the recommended storage temperature for processed, raw meats prior to shipping or further processing) or 2 d at 13 C (the maximum allowable temperature in poultry processing rooms). Meat samples were stored for 5 d at 4 C to simulate the approximate maximum time that a product would be maintained under ideal conditions before consumption or for 2 d at 13 C to simulate improper storage of product after processing.
MATERIALS AND METHODS Sample Collection and Storage Poultry meat samples were collected aseptically on a weekly basis for 6 wk from a local processing plant. Skin-
less chicken breasts were removed from the processing line, placed into resealable plastic bags, and stored on ice for transportation to the laboratory. These samples were subsequently stored at either 4 C for 1 to 5 d or at 13 C for 1 or 2 d prior to analysis. Each of the three trials of the experiment included two meat samples for Day 0 (no storage), two meat samples at each temperature on each of Days 1 and 2 of storage, and two meat samples at 4 C on each of Days 3, 4, and 5 of storage.
Preparation of Microtiter Plates Containing Dehydrated Substrates The inocula for microtiter plates were prepared as follows. After storage for the specified time periods, meat samples were weighed (average = 243 g, range = 169 to 316 g, n = 42) and were shaken for 1 min with 35 mL of sterile normal saline (0.85% NaCl) in a resealable plastic bag. The volumes of rinses recovered were measured prior to dilution of approximately 0.5 to 2.0 mL of each rinse into 18-mL sterile normal saline (in a 20 × 150 mm borosilicate test tube). Final dilution volumes varied depending upon the rinse sample. From each rinse, triplicate saline tubes were prepared at 60 and 90% transmittance on a Biolog3 turbidimeter at 590 nm for use as inocula on microtiter plates. Two inoculum densities were used to ensure adequate optical density change (purple color development) at 590 nm within an acceptable time frame for analysis and to maintain approximately equal numbers of microorganisms (at any given transmittance value) per well on the microtiter plates. The 60% transmittance value was utilized because it is used for identification of pure cultures of Gram-negative bacteria. The 90% transmittance value was arbitrarily chosen as a lower inoculum density to ensure that color development was not too rapid. Microtiter plates containing the 95 substrates listed in Table 1 were inoculated and incubated according to the following procedure. Each rinse sample diluted in saline to 60 or 90% transmittance at 590 nm was poured into its respective sterile reservoir. An eight-channel repeating pipetter was used to dispense 150 µL of the inoculum into each of the 96 wells (8 rows of 12 wells with a different dehydrated substrate but no additional nutrients or growth factors in each well) of a microtiter plate. Plates were covered and placed into either a 4 or 13 C refrigerator for a time (96 h minimum) sufficient for adequate color development (determined as optical density at 590 nm) in at least one well other than the control well. The first well in each plate, containing rinse dilution and indicator dye but no added substrate, served as the control well. Data from this well was used to account for color development from substances in the rinse dilution. The change in optical density in each well, indicative of substrate utilization, varied with the substrate of interest and was determined by the Biolog software using the default parameters (i.e., seconds for reading the plate and mathematical determination of threshold for positive or negative reactions).
1803
SUBSTRATE USE BY BIOFILM COMMUNITIES ON CHICKEN MEAT TABLE 1. Utilization of substrates by microbial communities associated with meat samples stored at 4 or 13 C 4C Plate incubation (h)1 Days of meat storage: Test substrate Polymers (5)3 α-cyclodextrin dextrin glycogen tween 40 tween 80 Carbohydrates (28)3 N-acetyl-D-galactosamine N-acetyl-D-glucosamine adonitol L-arabinose D-arabitol cellobiose i-erythritol D-fructose L-fucose D-galactose gentiobiose α-D-glucose m-inositol α-D-lactose lactulose maltose D-mannitol D-mannose D-melibiose β-methyl-D-glucoside D-psicose D-raffinose L-rhamnose D-sorbitol sucrose D-trehalose turanose xylitol Carboxylic acids (27)3 methyl pyruvate mono-methyl succinate acetic acid cis-aconitic acid citric acid formic acid D-galactonic acid lactone D-galacturonic acid D-gluconic acid D-glucosaminic acid D-glucuronic acid α-hydroxybutyric acid β-hydroxybutyric acid γ-hydroxybutyric acid ρ-hydroxyphenylacetic acid itaconic acid α-ketobutyric acid α-ketoglutaric acid α-ketovaleric acid D,L-lactic acid malonic acid propionic acid quinic acid D-saccharic acid sebacic acid succinic acid bromosuccinic acid Amides/Amines (5)3 succinamic acid glucuronamide
1
2
13 C Plates positive2 (%) (n = 28) 1–2
(n = 70) 1–5
Plate incubation (h)1 1
Plates positive2 (%) (n = 24) 1–2
2
17 17 17 17 17
48 20 20 17 17
50 89 89 100 100
50 81 86 100 96
04 04 04 04 04
46 22 22 06 20
63 100 100 100 100
17 17 17 17 19 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17
19 17 17 17 23 17 45 17 17 17 17 17 17 17 17 20 17 17 17 17 17 17 17 17 17 17 17 17
96 96 79 89 71 71 54 100 100 100 93 100 100 89 89 100 100 93 86 96 96 100 100 96 100 100 93 89
80 87 77 84 67 61 47 99 97 99 89 97 93 80 83 97 91 79 71 91 90 94 93 91 97 96 88 86
04 44 68 43 51 51 144 04 04 04 04 04 04 04 43 43 04 43 51 04 04 04 04 04 04 43 04 04
29 22 44 29 46 29 120 20 24 22 29 20 46 24 20 22 24 22 29 22 22 24 29 24 20 22 20 46
88 100 50 100 92 100 21 100 100 100 92 96 96 92 29 100 100 100 96 100 96 100 100 100 96 100 96 42
19 17 17 17 19 17 17 17 17 17 17 17 17 17 19 17 17 17 17 17 17 17 17 17 17 19 17
20 45 17 20 20 17 20 17 17 20 20 20 20 24 17 20 20 41 17 20 20 17 17 20 20 45 41
93 71 79 100 96 96 96 100 100 96 96 93 89 71 57 86 68 100 71 100 86 96 100 100 79 82 96
81 49 50 91 96 90 90 91 96 91 90 83 87 63 56 80 69 84 69 97 77 89 97 100 76 84 80
44 164 48 04 04 04 04 04 04 04 48 04 04 19 04 44 43 43 144 43 51 44 43 43 43 44 43
24 120 44 22 22 06 22 22 22 46 20 46 24 46 48 29 46 22 120 20 46 46 22 29 48 29 22
92 17 42 100 96 71 100 100 100 79 100 67 96 83 79 92 54 100 21 96 75 54 100 100 25 96 96
17 17
20 17
82 79
84 70
43 51
29 29
96 63 (continued)
1804
HALE BOOTHE ET AL. TABLE 1. Continued 4C Plate incubation (h)1
Days of meat storage: Test substrate alaninamide phenylethylamine putrescine Amino acids (20)3 D-alanine L-alanine L-alanyl-glycine L-asparagine L-aspartic acid L-glutamic acid glycyl-L-aspartic acid glycyl-L-glutamic acid L-histidine hydroxy L-proline L-leucine L-ornithine L-phenylalanine L-proline L-pyroglutamic acid D-serine L-serine L-threonine D,L-carnitine γ-aminobutyric acid Miscellaneous (10)3 urocanic acid inosine uridine thymidine 2-aminoethanol 2,3-butanediol glycerol D,L-α-glycerol phosphate glucose-1-phosphate glucose-6-phosphate
1
2
13 C Plates positive2 (%) (n = 28) 1–2
(n = 70) 1–5
Plate incubation (h)1 1
2
Plates positive2 (%) (n = 24) 1–2
19 17 19
19 20 20
82 68 79
80 59 61
51 144 67
20 120 46
50 25 92
17 17 17 17 17 17 17 25 17 17 17 17 17 17 17 17 17 17 17 17
20 20 17 17 20 20 20 72 45 20 20 20 24 20 17 17 17 20 20 117
89 93 93 100 96 93 79 68 86 96 75 93 79 96 100 96 100 75 86 89
89 93 91 97 94 94 73 49 84 94 69 84 76 93 97 94 91 71 80 81
43 43 43 43 43 43 48 68 43 43 48 44 144 43 43 43 44 68 06 43
22 24 44 22 22 22 46 46 24 22 46 46 120 20 44 22 22 46 120 29
100 100 100 100 100 100 63 42 100 100 54 92 25 100 100 100 100 63 25 100
144 17 17 17 17 17 17 17 17 17
24 17 17 17 20 20 20 24 41 45
82 96 89 86 82 71 96 71 82 86
61 73 73 71 57 63 94 61 64 60
48 44 68 68 144 04 04 44 44 44
44 29 29 46 120 53 22 46 29 29
96 100 75 63 17 13 96 58 100 100
1
Plate incubation time (hours) for detection of initial substrate metabolism. Percentage of plates on which individual substrates were transformed. 3 Substrate group tested (number of individual substrates in group). 2
Reading of Microtiter Plates
Statistical Analysis
After incubation of microtiter plates for specified time periods, plates were read at 590 nm on the Molecular Devices4 Emax plate reader. Plates were read at 0 h and at least twice daily for a minimum of 4 d. In some cases, microplates incubated at 4 C were read twice daily for up to 2 wk (336 h incubation) to measure optical density at 590 nm, which is indicative of substrate utilization. Data for each plate were stored in the following three formats in DOS as 8 × 12 matrices: 1) raw optical density values, 2) percentage change in optical density, and 3) +/− for substrate utilization. The frequency of metabolism of a substrate was determined by counting the number of plates positive for utilization of that substrate and determining the percentage of the total number of plates that were positive.
Statistical analysis of data was performed after the following data conversions. Raw optical density data from each microtiter plate for its entire incubation period were retrieved as 1 × 96 matrices in ASCII format. These data for each plate from each reading time were imported into an Excel spreadsheet in which the following functions were performed. First, optical density values were corrected for background absorbance caused by extraneous substances in the rinses by subtracting the value for the corresponding control well. Subsequently, corrected values were integrated using the formula of Guckert et al. (1996) to obtain the area under the time curve. Integrated values were then imported into SAS威 for statistical analysis using the general linear models procedure (SAS Institute, 1996). Integrated curve values were selected for use in statistical analysis because they represented substrate transformation throughout the entire incubation period rather than that at a single time.
4
Molecular Devices Corp., Sunnyvale, CA 94089.
SUBSTRATE USE BY BIOFILM COMMUNITIES ON CHICKEN MEAT
A factorial ANOVA with the integrated values described above was performed with main effects and twofactor interactions only. The factors were chicken meat storage temperatures (4 and 13 C), days of storage (0, 1, 2, 3, 4, and 5), plate incubation temperatures (4 and 13 C), and inoculum densities (60 and 90% transmittance). The experimental design was an unbalanced, incomplete block design in three blocks (trials). Data from each of the three trials were used as replication in the statistical model (see Results and Discussion section).
RESULTS AND DISCUSSION Feasibility of Using Biolog3 Plates with Poultry Microbial Communities Results of the present work show that Biolog3 plates can be used to determine the potential of in vitro substrate utilization by bacterial communities (biofilm) removed from poultry products as they develop. Chicken meat samples used in this study were obtained immediately after cut-up of broiler carcasses that follows chlorine treatment in chiller tanks. Such treatment would be expected to yield a minimal bacterial population on the meat samples at the time of their collection (Day 0) but not necessarily after storage of the samples for several days. The rinsing protocol in this study was designed to remove samples of the more loosely attached bacterial communities that developed on the meat samples over time. Rinsing protocols for bacteriological analysis of chicken meat samples have been in use for 30 yr (Surkiewicz et al., 1969). The 35 mL of 0.85% normal saline used herein for rinsing each meat sample yielded a sufficient volume (mean = 26 mL) for subsequent dilution and inoculation of Biolog3 plates at 60 or 90% transmittance at 590 nm. However, the rinsing protocol for other types of meat samples may need to be altered depending upon the characteristics (i.e., fat content or weight) of the meat and the inoculum density requisite on the microtiter plate. An acceptable dilution of rinse solution for inoculation onto microtiter plates would contain a minimum of substances interfering with optical density values at 590 nm. Replicate samples were used in the present study to assess the possibility of such interferences by determining the incidence of false-positive wells (those in which the tetrazolium dye is reduced to a purple color after inoculation with an individual rinse dilution but in which substrate has not been transformed).
Utilization of Substrates by Biofilm Communities as They Develop on Meat Samples Table 1 summarizes the utilization of the tested substrates by biofilm communities from meat samples stored at either 4 C for up to 5 d or at 13 C for up to 2 d. Utilization was defined as the ability of the microflora to obtain reducing equivalents from the substrate to reduce the tetrazolium dye. This term is used interchangeably
1805
with the terms metabolism and transformation in the remainder of this paper and does not indicate the extent of use (i.e., partial or complete) of the substrate. Data in Table 1 are grouped according to the major substrate classes (polymers, carbohydrates, carboxylic acids, amines or amides, amino acids, and miscellaneous) on the microtiter plate. Two types of data are listed as an indicator of the rapidity and frequency, respectively, of utilization of the 95 substrates tested: 1) the plate incubation time (h) when substrate metabolism was initially detected and 2) the percentage of plates on which individual substrates were transformed. Data reported are from microplates with an inoculum density corresponding to a 90% transmittance value. All substrates were metabolized by microbial communities from meat samples stored 1 d at 4 or at 13 C. However, the length of incubation of these populations on microtiter plates prior to substrate metabolism ranged from 17 to 144 h at 4 C and from 4 to 164 h at 13 C. Biofilm communities associated with samples at 4 C utilized 98% (93) of the 95 substrates within 24 h of plate incubation and used one substrate after 25 and 144 h, respectively, of plate incubation. The time required for substrate utilization by biofilm communities associated with meat samples stored 1 d at 13 C was more variable than that at 4 C. Communities from meat stored 1 d at 13 C transformed 38% (36), 43% (41), and 13% (12) of the 95 substrates between 0 to 24, 25 to 48, and 49 to 72 h, respectively, of plate incubation at 13 C. Only 4 h of plate incubation were required for utilization of 34 (36%) of the 95 substrates. In contrast, 144 h of incubation were required for metabolism of five substrates (i-erythritol, α-ketovaleric acid, phenylethylamine, L-phenylalanine, and 2-aminoethanol), and 164 h of incubation were required for transformation of one substrate (mono-methyl succinate). All substrates were also utilized by biofilm communities from meat samples stored 2 d at 4 or 13 C, but the requisite length of plate incubation varied from 17 to 117 h at 4 C and from 6 to 120 h at 13 C. Transformation of most of the substrates by microbial populations associated with these meat samples began before 48 h of plate incubation. Microbial communities associated with samples at 4 C utilized 88% (84) of the 95 substrates within 24 h of plate incubation, an additional 9% (9) of the substrates by 48 h of incubation, and another 1% of the substrates at 72 and 117 h of plate incubation. Biofilm populations associated with 13 C samples metabolized 48% (46) of the substrates within 24 h of plate incubation, an additional 43% (41) of the substrates by 48 h of incubation, and another 1 and 7% at 53 and 120 h of plate incubation. Thus, the data indicate that microbial communities on chicken meat stored at 4 C for 2 d may potentially metabolize half the substrates more rapidly than the communities on chicken meat stored at 13 C for 2 d. However, communities associated with meat stored at 13 C for 2 rather than 1 d utilized more rapidly approximately one-third of the substrates. The rapidity of metabolism of the different substrate classes by microbial populations from meat samples
1806
HALE BOOTHE ET AL.
stored 1 d at 4 or at 13 C is summarized in Table 2. At 4 C, utilization of all substrate classes except the miscellaneous group occurred after 17 to 18 h of plate incubation. However, it should be noted that these data for 4 C samples represent only two samples, and data from four other meat samples stored under the same conditions indicate that a 42- to 66-h plate incubation period may be needed for the transformation of many of the substrates. Such a plate incubation period is comparable to that required for transformation of several substrate classes by meat samples stored 1 d at 13 C. At this temperature, polymers, transformed after 4 h of plate incubation, were more rapidly utilized than other substrate groups. Metabolism of the remaining substrate groups at 13 C required between 24 to 72 h and occurred in the following order: carbohydrates > carboxylic acids > miscellaneous and amino acids > amides or amines. The rapidity of metabolism of the different substrate classes by microbial populations from meat samples stored 2 d at 4 or 13 C is also summarized in Table 2. At 4 C, the order of metabolism of the substrate groups was carbohydrates and amides or amines > carboxylic acids and polymers and miscellaneous > amino acids. At 13 C, substrates were utilized in the following order: polymers > carbohydrates > carboxylic acids and amino acids > miscellaneous > amides or amines. The data indicate that carbohydrates and carboxylic acids may be more readily utilized than the other substrates by biofilm communities associated with meat samples stored 2 d at 4 or 13 C. The frequency of metabolism of individual substrates at 4 and at 13 C varied as follows (Table 1). The percentage of plates on which the individual substrates were metabolized ranged from 50 to 100% and from 47 to 100% for plates prepared from meat samples stored at 4 C for Days 1 and 2 and for Days 1 through 5, respectively. For plates prepared from meat samples at 13 C stored 1 and 2 d, the percentage of plates on which individual substrates were utilized varied from 17 to 100%. The frequency of utilization of 81 (85%) of the 95 substrates was similar between the microbial populations associated with meat samples stored 1 and 2 d at 4 or at 13 C. Storage of meat samples for up to 5 d at 4 C generally did not increase
the frequency of metabolism of substrates. The substrates whose frequency of metabolism varied the greatest during meat storage include cellobiose, lactulose, and xylitol in the carbohydrate group; mono-methyl succinate, αketovaleric acid, propionic acid, and sebacic acid in the carboxylic acid group; alaninamide, phenylethylamine, and putrescine in the amide or amine group; L-phenylalanine, D, L-carnitine in the amino acid group; and urocanic acid, 2-aminoethanol, 2,3-butanediol, glucose-1-phosphate, and glucose-6-phosphate in the miscellaneous group. Differences in utilization of these substrates might have been due to diverse metabolic capabilities of individual bacterial species or the interaction of multiple species at 4 and at 13 C. The frequency of metabolism of substrate groups by microbial communities associated with stored meat was comparable in most cases for plates incubated at 4 and 13 C (Table 3). Mean percentage of plates on which substrate groups were utilized by microbial populations ranged from 78 to 92% and from 68 to 86% for plates prepared from 4 C samples stored Days 1 and 2 and Days 1 through 5, respectively. Substrate groups were metabolized on 65 to 93% of plates prepared from meat samples stored 1 and 2 d at 13 C. Carbohydrates were most frequently utilized by samples at 4 C, whereas polymers were most frequently metabolized by samples at 13 C. Amides or amines and miscellaneous were the most infrequently utilized groups at the two temperatures. Because the difference in frequency of utilization of any individual substrate group by biofilm communities at 4 and 13 C was only 3 to 16%, metabolically similar bacteria may be present in meat stored at 4 C for up to 5 d and at 13 C for up to 2 d.
Statistical Comparison of Substrate Utilization A statistical analysis of the integrated microtiter plate data was performed to identify significant differences in substrate utilization profiles of biofilm communities from meat stored at 4 or 13 C. In particular, the effect of meat storage time (d) and temperature was investigated
TABLE 2. Average time (h) for initial utilization of substrate groups by biofilm communities associated with meat samples stored at 4 or 13 C Substrate group Storage conditions 4 C samples Day 11 Day 21 13 C samples Day 12 Day 23
Polymers
Carbohydrates
Carboxylic acids
Amides/ amines
Amino acids
Miscellaneous
17 24
17 18
17 23
18 19
17 28
30 (17)4 25
4 23
25 30
37 38
71 (53) 49 (31)
52 (48) 41
51 (41) 45
n (number of plates) = 14. n = 11. 3 n = 13. 4 Average incubation time excluding a single substrate requiring a lengthy incubation period (e.g., 144 h) for transformation. 1 2
1807
SUBSTRATE USE BY BIOFILM COMMUNITIES ON CHICKEN MEAT TABLE 3. Frequency of utilization (mean % positive plates) of substrate groups by biofilm communities associated with meat samples stored at 4 or 13 C Substrate group Storage conditions
Polymers
Carbohydrates
Carboxylic acids
Amides/ amines
Amino acids
Miscellaneous
4 C Samples Days 1–51 Days 1–22
83 86
86 92
82 89
71 78
85 89
68 84
13 C Samples Days 1–23
93
89
79
65
83
72
n (number of plates) = 70. n = 28. 3 n = 24. 1 2
through use of a factorial analysis of variance model. The following main effects and two-factor interactions were included in the SAS威 analysis. The main effects were 1) meat storage time (d), 2) (chicken) meat storage temperature (ctemp), 3) microtiter plate incubation temperature (itemp), and 4) density of inoculum on the microtiter plate (den). The two-factor interactions included 1) day by ctemp, 2) day by itemp, 3) day by den, and 4) itemp by den. The integrated data for each of the 95 substrates obtained using mixed microbial communities associated with meat samples are designated as y2 to y96 in the following model: proc glm; class trial day ctemp itemp den; model y2 to y96 = trial day ctemp itemp den day *ctemp day*itemp day*den itemp*den trial*day*ctemp*itemp*den; test h = day ctemp itemp den day*ctemp day*itemp day*den itemp*den e = trial *day*ctemp*itemp*den; means day ctemp itemp den day*ctemp day*itemp day*den itemp*den;. The model accounted for approximately 50 to 92% of the variation in the data for most of the 95 substrates. Main effects that were significant at P = 0.05 included plate incubation temperature (for 91 substrates), inoculum density (for 69 substrates), and the interaction of plate incubation temperature and inoculum density (for 92 substrates). Other researchers (Haack et al., 1995; Garland, 1996; Hitzl et al., 1997) have also noted the effects of incubation temperature and inoculum density on the utilization of most of the Biolog3 substrates. The P values for the main effects of plate incubation temperature and inoculum density and their interaction on selected substrates are included in Table 4. Other main effects were significant at P = 0.05 for fewer substrates than the main effects of plate incubation temperature or inoculum density and their interaction (Table 4). Meat incubation temperature was significant for one polymer (α-cyclodextrin), three carbohydrates (adonitol, i-erythritol, and D-psicose), four carboxylic acids (γ-hydroxybutyric acid, α-ketobutyric acid, α-ketovaleric acid, and sebacic acid) and one amine (phenylethylamine). Substrates affected by meat temperature at probability values between 0.06 and 0.15 include one carbohydrate, five carboxylic acids, one amide, and two amino acids (not all data shown). Metabolism of the carboxylic acids, carbohydrates, and amino acids included on the microtiter plate may thus be expected to be more variable in
poultry samples stored at 4 or 13 C than the metabolism of the polymers, amines, or amides. Day of meat storage was significant for only one substrate (phenylethylamine) at P = 0.05; the P values for four carboxylic acids, one amino acid, and one miscellaneous compound were between 0.06 and 0.15. Thus, distinct differences in metabolism of most of the substrates by microbial populations were not evident in this study, based on day of storage of the associated meat samples. The interactions of day and meat sample temperature, day and microtiter plate incubation temperature, and day and density were not significant (P = 0.05) as often as the interaction of incubation temperature and density. The interaction of day and chicken temperature was significant (P = 0.05) for only three carboxylic acids (γ-hydroxybutyric acid, α-ketobutyric acid, and sebacic acid) and one amine (phenylethylamine). For two carbohydrates and two carboxylic acids, probability values ranged between 0.08 and 0.12. The interaction of day and incubation temperature was significant (P = 0.05) for the same three carboxylic acids as the day and chicken temperature interaction. Other substrates for which this interaction was significant included one amine (phenylethylamine), one carbohydrate (i-erythritol), one amino acid (L-leucine), and one miscellaneous compound (2-aminoethanol). For two carbohydrates and one amino acid, the probability values were between 0.07 and 0.14. The interaction of day and density was significant for only 2-aminoethanol (miscellaneous group).
CONCLUSIONS The present study focused on assessing the metabolic processes of microorganisms growing on chicken meat surfaces with the ultimate goal of controlling these processes to yield safer poultry products with less potential for spoilage. Our results demonstrate that the substrate utilization profiles of bacterial communities (biofilm) as they develop on chicken meat surfaces may be assessed using an appropriate rinsing protocol and Biolog3 GN microtiter plates. Utilization of some of the 95 test substrates by these biofilm communities may begin as early as 1 d after meat storage at 4 or 13 C. However, the rapidity and frequency of substrate metabolism may vary
1808
HALE BOOTHE ET AL. 1
TABLE 4. Source of variation and significance2 for integrated substrate utilization data3 Source of variation and p value Test substrate
Day
Ctemp
Itemp
Den
Day*Ctemp
Day*Itemp
Day*Den
Itemp*Den
Polymers α-cyclodextrin
0.3448
0.0405
0.0126
0.6743
0.1948
0.1954
0.9556
0.0170
0.2864 0.3014 0.6356
0.0107 0.0435 0.0430
0.0001 0.0046 0.0001
0.0385 0.1422 0.1151
0.0923 0.0829 0.2379
0.0676 0.0370 0.1929
0.9813 0.9843 0.9350
0.0020 0.0059 0.0105
0.4446 0.7178 0.0609 0.0606 0.3707 0.0681
0.0979 0.7598 0.0009 0.0002 0.0167 0.0004
0.0001 0.1777 0.0015 0.0027 0.0354 0.0015
0.0721 0.2292 0.2435 0.0618 0.4595 0.1366
0.3862 0.0778 0.0188 0.0056 0.1244 0.0095
0.1982 0.4053 0.0176 0.0088 0.1649 0.0103
0.9468 0.9845 0.9566 0.9669 0.9907 0.9587
0.0042 0.0757 0.0196 0.0044 0.1235 0.0068
0.5007 0.0082
0.1091 0.0001
0.0002 0.0001
0.0724 0.5197
0.3183 0.0001
0.1623 0.0001
0.9872 0.8735
0.0018 0.0269
0.1184 0.6501
0.0657 0.0556
0.0001 0.0005
0.0184 0.1662
0.1762 0.2734
0.0328 0.1358
0.9543 0.9949
0.0001 0.0324
0.1491
0.3840
0.0002
0.9570
0.3746
0.0116
0.0001
0.0006
Carbohydrates adonitol i-erythritol D-psicose Carboxylic acids methyl pyruvate α-hydroxybutyric acid γ-hydroxybutyric acid α-ketobutyric acid α-ketovaleric acid sebacic acid Amines/amides alaninamide phenyethylamine Amino acids L-leucine L-threonine Miscellaneous 2-aminoethanol
Abbreviations for sources of variation are Ctemp = (chicken) meat temperature, Itemp = microtiter plate incubation temperature, Den = microtiter plate inoculum density, Day*Ctemp = interaction of meat storage time (d) and meat temperature, Day*Itemp = interaction of meat storage time and microtiter plate incubation temperature, Day*Den = interaction of meat storage time and microtiter plate inoculum density, and Itemp*Den = interaction of microtiter plate incubation temperature and inoculum density. 2 P values are for means of integrated data for substrate utilization determined as optical density at 590 nm. 3 Means of main effects and interactions for all 95 substrates may be obtained by writing to the senior author. 1
significantly, as evidenced in the present study by the 4 to 164 h of microtiter plate incubation required for utilization of some substrates and the range of 17 to 100% of microplates on which specific substrates were transformed. Utilization of chemically related substrates was less variable at 4 than at 13 C. At the latter temperature, the following order was observed for the rapidity and frequency of initial transformation of the respective substrate groups: polymers > carbohydrates > carboxylic acids > miscellaneous or amino acids > amides or amines. Data on substrate utilization obtained in this research may be useful in characterizing and controlling biofilm communities implicated in pathogenicity or affecting food quality of chicken meat samples. For example, reduction or removal of rapidly metabolized substrates in the poultry processing environment may reduce biofilm formation and persistence. Substrates may be removed by chemical sequestration or change in the pH or redox potential of fluids prevalent in poultry-processing plants. Candidate substrates may include some of the individual substrates (i.e., cellobiose in the carbohydrate group or glucose-6-phosphate in the miscellaneous group) identified in this study whose frequency of utilization differed the most between biofilm communities associated with meat samples stored at 4 or at 13 C. Removal or reduction of these substrates from the processing environment may reduce the metabolic activity of microbial populations on chicken meat samples shifted from storage at 4 to 13 C or vice versa. Other substrates commonly present in the
poultry processing environment may be tested in future work.
ACKNOWLEDGMENTS The authors thank the following individuals of USDAARS, PPMQ: Kurt Lawrence and Gavin Poole for help in SAS威 input and Manju V. Amin and Ruth A. Lebo for technical support and assistance in data collection.
REFERENCES Arnold, J. W., 1998. Development of bacterial biofilms during poultry processing. Poult. Avian Biol. Rev. 9:1–9. Arnold, J. W., and L. J. Shimkets, 1988. Inhibition of cell-cell interactions in Myxococcus xanthus by Congo red. J. Bacteriol. 170:5765–5770. Buyer, J. S., and L. E. Drinkwater, 1997. Comparison of substrate utilization assay and fatty acid analysis of soil microbial communities. J. Microbiol. Methods 30:3–11. Ellender, R. D., L. Huang, S. L. Sharp, and R. P. Tettleton, 1994. Isolation, enumeration, and identification of Gram-positive cocci from frozen crabmeat. J. Food Prot. 58:853–857. Fernandez, J. L., W. J. Simpson, and T. M. Dowhanick, 1993. Enumeration of Obesumbacterium proteus in brewery yeasts and characterization of isolated strains using Biolog GN MicroPlates and protein fingerprinting. Lett. Appl. Microbiol. 17:292–296. Garland, J. L., 1996. Analytical approaches to the characterization of samples of microbial communities using patterns of potential C source utilization. Soil Biol. Biochem. 28:213– 221.
SUBSTRATE USE BY BIOFILM COMMUNITIES ON CHICKEN MEAT Garland, J. L., and A. L. Mills, 1991. Classification and characterization of heterotrophic microbial communities on the basis of patterns of community-level-sole-carbon-sourceutilization. Appl. Environ. Microbiol. 57:2351–2359. Guckert, J. B., G. J. Carr, T. D. Johnson, B. G. Hamm, D. H. Davidson, and Y. Kumagai, 1996. Community analysis by BIOLOG: Curve integration for statistical analysis of activated sludge microbial habitats. J. Microbiol. Methods 27:183–197. Haack, S. K., H. Garchow, M. J. Klug, and L. J. Forney, 1995. Analysis of factors affecting the accuracy, reproducibility and interpretation of microbial community carbon source utilization patterns. Appl. Environ. Microbiol. 61:1458– 1468. Haack, S. K., H. Garchow, W. Sobczak, L. J. Forney, and M. J. Klug, 1994. Analysis of changes in stream biofilm community structure and function in response to environmental perturbation using whole-community fatty acid methyl ester profiles and carbon-substrate utilization patterns. Page 352 in: Abstracts of the 94th General Meeting of the American Society for Microbiology, Las Vegas, NV.
1809
Hitzl, W., A. Rangger, S. Sharma, and H. Insam, 1997. Separation power of the 95 substrates of the BIOLOG system determined in various soils. FEMS Microbiol. Ecol. 22:167–174. Johnson, J. L., and C. P. Lattuada, 1993. Comparison of nucleic acid hybridization assays and biochemical characterization tests for the confirmation of Listeria monocytogenes. J. Food Prot. 56:834–840. Praphailong, W., M. Vangestel, G. H. Fleet, and G. M. Heard, 1997. Evaluation of the Biolog system for the identification of food and beverage yeasts. Lett. Appl. Microbiol. 24:455–459. Russell, S. M., D. L. Fletcher, and N. A. Cox, 1996. Spoilage bacteria of fresh broiler chicken carcasses. Poultry Sci. 75:2041–2047. Surkiewicz, B. F., R. W. Johnston, A. B. Moran, and G. W. Krumm, 1969. A bacteriological survey of chicken eviscerating plants. Food Technol. 23:80–85. SAS Institute, 1996. SAS威 Release 6.12 for Windows. SAS Institute Inc., Cary, NC. Victorio, L., K. A. Gilbride, D. G. Allen, and S. N. Liss, 1996. Phenotypic fingerprinting of microbial communities in wastewater treatment systems. Water Res. 30:1077–1086.
early as 1 d after storage at 4 or at 13 C. However, the rapidity (incubation time required for initial substrate utilization) and frequency (percentage of plates positive for transformation of an individual substrate) of metabolism of the substrates by the biofilm communities varied from 4 to 164 h of plate incubation and from 17 to 100% of microplates, respectively. At 13 C, polymers were the most rapidly metabolized substrate group, followed by carbohydrates, carboxylic acids, miscellaneous or amino acids, and amides or amines. Initial utilization of these substrate groups at 4 C occurred within a consistently shorter period (24 h of plate incubation). The frequency of metabolism of each individual substrate group varied only 3 to 16% between samples stored at 4 and 13 C. However, a greater difference in frequency of utilization of some individual substrates was noted. Such divergences may be useful in characterizing biofilm communities implicated in pathogenicity or affecting food quality of poultry products.
(Key words: poultry, biofilms, bacteria, substrate utilization, chicken meat) 1999 Poultry Science 78:1801–1809
INTRODUCTION The safety of our food supply can be immediately improved by reducing the potential for foodborne illnesses caused by bacterial contamination of food products. Growth of bacteria on meat, surrounding fluid, and inanimate surfaces of the processing plants leads to product contamination and spoilage. Substantial bacterial contamination of poultry products in the processing and storage environment is composed of many different species of microbes in a biofilm community (Arnold, 1998). Members of the biofilm community share and compete for
Received for publication April 5, 1999. Accepted for publication September 1, 1999. 1 This article is a US Government work, and, as such, is in the public domain in the United States of America. Reference to a company name or product does not imply endorsement by the US Department of Agriculture. 2 To whom correspondence should be addressed: email: jarnold@ ars.usda.gov. 3 Biolog, Inc., Hayward, CA 94545.
surface attachment sites, light, and carbon and energy sources. Although traditional microbiological studies have been based on pure cultures of one bacterial species, the behavior of a single species can change when mixed with other bacteria with different properties (Arnold and Shimkets, 1988). Recently, methods have been developed to assess the biological and chemical properties of biofilm communities as a whole. Garland and Mills (1991) introduced the concept of community-level physiological profiles using Biolog3 GN microtiter plates for characterizing microbial communities in environmental samples. Several researchers have applied this approach to the analysis of microbial communities in a variety of environments, including soil, water, and wastewater (Garland and Mills, 1991; Haack et al., 1994; Guckert et al., 1996; Victorio et al., 1996; Buyer and
Abbreviation Key: ctemp = chicken meat storage temperature; itemp = microtiter plate incubation temperature; den = density of inoculum on the microtiter plate.
1801
1802
HALE BOOTHE ET AL.
Drinkwater, 1997). In this method, the metabolic capabilities of mixed populations of microorganisms are determined by their ability to utilize each of 95 substrates at predetermined concentrations commonly used to identify Gram-negative bacteria. Buyer and Drinkwater (1997) recently categorized these substrates into six major groups: carbohydrates (30 substrates), carboxylic acids (24 substrates), polymers (5 substrates), amines or amides (6 substrates), amino acids (20 substrates), and miscellaneous (10 substrates, including phosphorylated compounds). The ability of microbial communities to utilize these 95 substrates on Biolog3 GN microplates is assessed using the following procedure: 1) preparation of a suspension of the microorganisms from the sample of interest, 2) inoculation of the organisms onto a 96-well microtiter plate containing a different dehydrated substrate in each of 95 of the wells but no substrate in the control well, 3) incubation of the plate at an appropriate temperature, and 4) determination of the optical density at 590 nm in each of the wells of the plate. The last step quantitates reduction of the tetrazolium violet dye included in each well of the microplate. The reduced dye, which absorbs at 590 nm, is a purple color and is produced by substrate transformation (transfer of reducing equivalents from the substrate to the dye). The pattern of purple color development on the microtiter plate can subsequently be used to compare the metabolic capabilities of mixed populations of microorganisms. In the present study, the concept of community-level physiological profiles was applied to microbial populations in foods. Previously, Biolog3 GN microtiter plates have been used solely in food microbiology for the purpose of identification of individual organisms (Fernandez et al., 1993; Johnson and Lattuada, 1993; Ellender et al., 1994; Russell et al., 1996; Praphailong et al., 1997) rather than for the characterization of microbial populations associated with pathogenicity or food quality. The objectives in this research were threefold: 1) to determine the feasibility of using 96-well microtiter plates to assess substrate utilization profiles of bacterial communities (biofilm) as they develop on poultry products, 2) to identify substrates utilized by microbial populations from stored chicken meat, and 3) to compare the substrate utilization patterns of biofilm communities as they develop on meat stored up to 2 or 5 d at 4 C (the recommended storage temperature for processed, raw meats prior to shipping or further processing) or 2 d at 13 C (the maximum allowable temperature in poultry processing rooms). Meat samples were stored for 5 d at 4 C to simulate the approximate maximum time that a product would be maintained under ideal conditions before consumption or for 2 d at 13 C to simulate improper storage of product after processing.
MATERIALS AND METHODS Sample Collection and Storage Poultry meat samples were collected aseptically on a weekly basis for 6 wk from a local processing plant. Skin-
less chicken breasts were removed from the processing line, placed into resealable plastic bags, and stored on ice for transportation to the laboratory. These samples were subsequently stored at either 4 C for 1 to 5 d or at 13 C for 1 or 2 d prior to analysis. Each of the three trials of the experiment included two meat samples for Day 0 (no storage), two meat samples at each temperature on each of Days 1 and 2 of storage, and two meat samples at 4 C on each of Days 3, 4, and 5 of storage.
Preparation of Microtiter Plates Containing Dehydrated Substrates The inocula for microtiter plates were prepared as follows. After storage for the specified time periods, meat samples were weighed (average = 243 g, range = 169 to 316 g, n = 42) and were shaken for 1 min with 35 mL of sterile normal saline (0.85% NaCl) in a resealable plastic bag. The volumes of rinses recovered were measured prior to dilution of approximately 0.5 to 2.0 mL of each rinse into 18-mL sterile normal saline (in a 20 × 150 mm borosilicate test tube). Final dilution volumes varied depending upon the rinse sample. From each rinse, triplicate saline tubes were prepared at 60 and 90% transmittance on a Biolog3 turbidimeter at 590 nm for use as inocula on microtiter plates. Two inoculum densities were used to ensure adequate optical density change (purple color development) at 590 nm within an acceptable time frame for analysis and to maintain approximately equal numbers of microorganisms (at any given transmittance value) per well on the microtiter plates. The 60% transmittance value was utilized because it is used for identification of pure cultures of Gram-negative bacteria. The 90% transmittance value was arbitrarily chosen as a lower inoculum density to ensure that color development was not too rapid. Microtiter plates containing the 95 substrates listed in Table 1 were inoculated and incubated according to the following procedure. Each rinse sample diluted in saline to 60 or 90% transmittance at 590 nm was poured into its respective sterile reservoir. An eight-channel repeating pipetter was used to dispense 150 µL of the inoculum into each of the 96 wells (8 rows of 12 wells with a different dehydrated substrate but no additional nutrients or growth factors in each well) of a microtiter plate. Plates were covered and placed into either a 4 or 13 C refrigerator for a time (96 h minimum) sufficient for adequate color development (determined as optical density at 590 nm) in at least one well other than the control well. The first well in each plate, containing rinse dilution and indicator dye but no added substrate, served as the control well. Data from this well was used to account for color development from substances in the rinse dilution. The change in optical density in each well, indicative of substrate utilization, varied with the substrate of interest and was determined by the Biolog software using the default parameters (i.e., seconds for reading the plate and mathematical determination of threshold for positive or negative reactions).
1803
SUBSTRATE USE BY BIOFILM COMMUNITIES ON CHICKEN MEAT TABLE 1. Utilization of substrates by microbial communities associated with meat samples stored at 4 or 13 C 4C Plate incubation (h)1 Days of meat storage: Test substrate Polymers (5)3 α-cyclodextrin dextrin glycogen tween 40 tween 80 Carbohydrates (28)3 N-acetyl-D-galactosamine N-acetyl-D-glucosamine adonitol L-arabinose D-arabitol cellobiose i-erythritol D-fructose L-fucose D-galactose gentiobiose α-D-glucose m-inositol α-D-lactose lactulose maltose D-mannitol D-mannose D-melibiose β-methyl-D-glucoside D-psicose D-raffinose L-rhamnose D-sorbitol sucrose D-trehalose turanose xylitol Carboxylic acids (27)3 methyl pyruvate mono-methyl succinate acetic acid cis-aconitic acid citric acid formic acid D-galactonic acid lactone D-galacturonic acid D-gluconic acid D-glucosaminic acid D-glucuronic acid α-hydroxybutyric acid β-hydroxybutyric acid γ-hydroxybutyric acid ρ-hydroxyphenylacetic acid itaconic acid α-ketobutyric acid α-ketoglutaric acid α-ketovaleric acid D,L-lactic acid malonic acid propionic acid quinic acid D-saccharic acid sebacic acid succinic acid bromosuccinic acid Amides/Amines (5)3 succinamic acid glucuronamide
1
2
13 C Plates positive2 (%) (n = 28) 1–2
(n = 70) 1–5
Plate incubation (h)1 1
Plates positive2 (%) (n = 24) 1–2
2
17 17 17 17 17
48 20 20 17 17
50 89 89 100 100
50 81 86 100 96
04 04 04 04 04
46 22 22 06 20
63 100 100 100 100
17 17 17 17 19 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17 17
19 17 17 17 23 17 45 17 17 17 17 17 17 17 17 20 17 17 17 17 17 17 17 17 17 17 17 17
96 96 79 89 71 71 54 100 100 100 93 100 100 89 89 100 100 93 86 96 96 100 100 96 100 100 93 89
80 87 77 84 67 61 47 99 97 99 89 97 93 80 83 97 91 79 71 91 90 94 93 91 97 96 88 86
04 44 68 43 51 51 144 04 04 04 04 04 04 04 43 43 04 43 51 04 04 04 04 04 04 43 04 04
29 22 44 29 46 29 120 20 24 22 29 20 46 24 20 22 24 22 29 22 22 24 29 24 20 22 20 46
88 100 50 100 92 100 21 100 100 100 92 96 96 92 29 100 100 100 96 100 96 100 100 100 96 100 96 42
19 17 17 17 19 17 17 17 17 17 17 17 17 17 19 17 17 17 17 17 17 17 17 17 17 19 17
20 45 17 20 20 17 20 17 17 20 20 20 20 24 17 20 20 41 17 20 20 17 17 20 20 45 41
93 71 79 100 96 96 96 100 100 96 96 93 89 71 57 86 68 100 71 100 86 96 100 100 79 82 96
81 49 50 91 96 90 90 91 96 91 90 83 87 63 56 80 69 84 69 97 77 89 97 100 76 84 80
44 164 48 04 04 04 04 04 04 04 48 04 04 19 04 44 43 43 144 43 51 44 43 43 43 44 43
24 120 44 22 22 06 22 22 22 46 20 46 24 46 48 29 46 22 120 20 46 46 22 29 48 29 22
92 17 42 100 96 71 100 100 100 79 100 67 96 83 79 92 54 100 21 96 75 54 100 100 25 96 96
17 17
20 17
82 79
84 70
43 51
29 29
96 63 (continued)
1804
HALE BOOTHE ET AL. TABLE 1. Continued 4C Plate incubation (h)1
Days of meat storage: Test substrate alaninamide phenylethylamine putrescine Amino acids (20)3 D-alanine L-alanine L-alanyl-glycine L-asparagine L-aspartic acid L-glutamic acid glycyl-L-aspartic acid glycyl-L-glutamic acid L-histidine hydroxy L-proline L-leucine L-ornithine L-phenylalanine L-proline L-pyroglutamic acid D-serine L-serine L-threonine D,L-carnitine γ-aminobutyric acid Miscellaneous (10)3 urocanic acid inosine uridine thymidine 2-aminoethanol 2,3-butanediol glycerol D,L-α-glycerol phosphate glucose-1-phosphate glucose-6-phosphate
1
2
13 C Plates positive2 (%) (n = 28) 1–2
(n = 70) 1–5
Plate incubation (h)1 1
2
Plates positive2 (%) (n = 24) 1–2
19 17 19
19 20 20
82 68 79
80 59 61
51 144 67
20 120 46
50 25 92
17 17 17 17 17 17 17 25 17 17 17 17 17 17 17 17 17 17 17 17
20 20 17 17 20 20 20 72 45 20 20 20 24 20 17 17 17 20 20 117
89 93 93 100 96 93 79 68 86 96 75 93 79 96 100 96 100 75 86 89
89 93 91 97 94 94 73 49 84 94 69 84 76 93 97 94 91 71 80 81
43 43 43 43 43 43 48 68 43 43 48 44 144 43 43 43 44 68 06 43
22 24 44 22 22 22 46 46 24 22 46 46 120 20 44 22 22 46 120 29
100 100 100 100 100 100 63 42 100 100 54 92 25 100 100 100 100 63 25 100
144 17 17 17 17 17 17 17 17 17
24 17 17 17 20 20 20 24 41 45
82 96 89 86 82 71 96 71 82 86
61 73 73 71 57 63 94 61 64 60
48 44 68 68 144 04 04 44 44 44
44 29 29 46 120 53 22 46 29 29
96 100 75 63 17 13 96 58 100 100
1
Plate incubation time (hours) for detection of initial substrate metabolism. Percentage of plates on which individual substrates were transformed. 3 Substrate group tested (number of individual substrates in group). 2
Reading of Microtiter Plates
Statistical Analysis
After incubation of microtiter plates for specified time periods, plates were read at 590 nm on the Molecular Devices4 Emax plate reader. Plates were read at 0 h and at least twice daily for a minimum of 4 d. In some cases, microplates incubated at 4 C were read twice daily for up to 2 wk (336 h incubation) to measure optical density at 590 nm, which is indicative of substrate utilization. Data for each plate were stored in the following three formats in DOS as 8 × 12 matrices: 1) raw optical density values, 2) percentage change in optical density, and 3) +/− for substrate utilization. The frequency of metabolism of a substrate was determined by counting the number of plates positive for utilization of that substrate and determining the percentage of the total number of plates that were positive.
Statistical analysis of data was performed after the following data conversions. Raw optical density data from each microtiter plate for its entire incubation period were retrieved as 1 × 96 matrices in ASCII format. These data for each plate from each reading time were imported into an Excel spreadsheet in which the following functions were performed. First, optical density values were corrected for background absorbance caused by extraneous substances in the rinses by subtracting the value for the corresponding control well. Subsequently, corrected values were integrated using the formula of Guckert et al. (1996) to obtain the area under the time curve. Integrated values were then imported into SAS威 for statistical analysis using the general linear models procedure (SAS Institute, 1996). Integrated curve values were selected for use in statistical analysis because they represented substrate transformation throughout the entire incubation period rather than that at a single time.
4
Molecular Devices Corp., Sunnyvale, CA 94089.
SUBSTRATE USE BY BIOFILM COMMUNITIES ON CHICKEN MEAT
A factorial ANOVA with the integrated values described above was performed with main effects and twofactor interactions only. The factors were chicken meat storage temperatures (4 and 13 C), days of storage (0, 1, 2, 3, 4, and 5), plate incubation temperatures (4 and 13 C), and inoculum densities (60 and 90% transmittance). The experimental design was an unbalanced, incomplete block design in three blocks (trials). Data from each of the three trials were used as replication in the statistical model (see Results and Discussion section).
RESULTS AND DISCUSSION Feasibility of Using Biolog3 Plates with Poultry Microbial Communities Results of the present work show that Biolog3 plates can be used to determine the potential of in vitro substrate utilization by bacterial communities (biofilm) removed from poultry products as they develop. Chicken meat samples used in this study were obtained immediately after cut-up of broiler carcasses that follows chlorine treatment in chiller tanks. Such treatment would be expected to yield a minimal bacterial population on the meat samples at the time of their collection (Day 0) but not necessarily after storage of the samples for several days. The rinsing protocol in this study was designed to remove samples of the more loosely attached bacterial communities that developed on the meat samples over time. Rinsing protocols for bacteriological analysis of chicken meat samples have been in use for 30 yr (Surkiewicz et al., 1969). The 35 mL of 0.85% normal saline used herein for rinsing each meat sample yielded a sufficient volume (mean = 26 mL) for subsequent dilution and inoculation of Biolog3 plates at 60 or 90% transmittance at 590 nm. However, the rinsing protocol for other types of meat samples may need to be altered depending upon the characteristics (i.e., fat content or weight) of the meat and the inoculum density requisite on the microtiter plate. An acceptable dilution of rinse solution for inoculation onto microtiter plates would contain a minimum of substances interfering with optical density values at 590 nm. Replicate samples were used in the present study to assess the possibility of such interferences by determining the incidence of false-positive wells (those in which the tetrazolium dye is reduced to a purple color after inoculation with an individual rinse dilution but in which substrate has not been transformed).
Utilization of Substrates by Biofilm Communities as They Develop on Meat Samples Table 1 summarizes the utilization of the tested substrates by biofilm communities from meat samples stored at either 4 C for up to 5 d or at 13 C for up to 2 d. Utilization was defined as the ability of the microflora to obtain reducing equivalents from the substrate to reduce the tetrazolium dye. This term is used interchangeably
1805
with the terms metabolism and transformation in the remainder of this paper and does not indicate the extent of use (i.e., partial or complete) of the substrate. Data in Table 1 are grouped according to the major substrate classes (polymers, carbohydrates, carboxylic acids, amines or amides, amino acids, and miscellaneous) on the microtiter plate. Two types of data are listed as an indicator of the rapidity and frequency, respectively, of utilization of the 95 substrates tested: 1) the plate incubation time (h) when substrate metabolism was initially detected and 2) the percentage of plates on which individual substrates were transformed. Data reported are from microplates with an inoculum density corresponding to a 90% transmittance value. All substrates were metabolized by microbial communities from meat samples stored 1 d at 4 or at 13 C. However, the length of incubation of these populations on microtiter plates prior to substrate metabolism ranged from 17 to 144 h at 4 C and from 4 to 164 h at 13 C. Biofilm communities associated with samples at 4 C utilized 98% (93) of the 95 substrates within 24 h of plate incubation and used one substrate after 25 and 144 h, respectively, of plate incubation. The time required for substrate utilization by biofilm communities associated with meat samples stored 1 d at 13 C was more variable than that at 4 C. Communities from meat stored 1 d at 13 C transformed 38% (36), 43% (41), and 13% (12) of the 95 substrates between 0 to 24, 25 to 48, and 49 to 72 h, respectively, of plate incubation at 13 C. Only 4 h of plate incubation were required for utilization of 34 (36%) of the 95 substrates. In contrast, 144 h of incubation were required for metabolism of five substrates (i-erythritol, α-ketovaleric acid, phenylethylamine, L-phenylalanine, and 2-aminoethanol), and 164 h of incubation were required for transformation of one substrate (mono-methyl succinate). All substrates were also utilized by biofilm communities from meat samples stored 2 d at 4 or 13 C, but the requisite length of plate incubation varied from 17 to 117 h at 4 C and from 6 to 120 h at 13 C. Transformation of most of the substrates by microbial populations associated with these meat samples began before 48 h of plate incubation. Microbial communities associated with samples at 4 C utilized 88% (84) of the 95 substrates within 24 h of plate incubation, an additional 9% (9) of the substrates by 48 h of incubation, and another 1% of the substrates at 72 and 117 h of plate incubation. Biofilm populations associated with 13 C samples metabolized 48% (46) of the substrates within 24 h of plate incubation, an additional 43% (41) of the substrates by 48 h of incubation, and another 1 and 7% at 53 and 120 h of plate incubation. Thus, the data indicate that microbial communities on chicken meat stored at 4 C for 2 d may potentially metabolize half the substrates more rapidly than the communities on chicken meat stored at 13 C for 2 d. However, communities associated with meat stored at 13 C for 2 rather than 1 d utilized more rapidly approximately one-third of the substrates. The rapidity of metabolism of the different substrate classes by microbial populations from meat samples
1806
HALE BOOTHE ET AL.
stored 1 d at 4 or at 13 C is summarized in Table 2. At 4 C, utilization of all substrate classes except the miscellaneous group occurred after 17 to 18 h of plate incubation. However, it should be noted that these data for 4 C samples represent only two samples, and data from four other meat samples stored under the same conditions indicate that a 42- to 66-h plate incubation period may be needed for the transformation of many of the substrates. Such a plate incubation period is comparable to that required for transformation of several substrate classes by meat samples stored 1 d at 13 C. At this temperature, polymers, transformed after 4 h of plate incubation, were more rapidly utilized than other substrate groups. Metabolism of the remaining substrate groups at 13 C required between 24 to 72 h and occurred in the following order: carbohydrates > carboxylic acids > miscellaneous and amino acids > amides or amines. The rapidity of metabolism of the different substrate classes by microbial populations from meat samples stored 2 d at 4 or 13 C is also summarized in Table 2. At 4 C, the order of metabolism of the substrate groups was carbohydrates and amides or amines > carboxylic acids and polymers and miscellaneous > amino acids. At 13 C, substrates were utilized in the following order: polymers > carbohydrates > carboxylic acids and amino acids > miscellaneous > amides or amines. The data indicate that carbohydrates and carboxylic acids may be more readily utilized than the other substrates by biofilm communities associated with meat samples stored 2 d at 4 or 13 C. The frequency of metabolism of individual substrates at 4 and at 13 C varied as follows (Table 1). The percentage of plates on which the individual substrates were metabolized ranged from 50 to 100% and from 47 to 100% for plates prepared from meat samples stored at 4 C for Days 1 and 2 and for Days 1 through 5, respectively. For plates prepared from meat samples at 13 C stored 1 and 2 d, the percentage of plates on which individual substrates were utilized varied from 17 to 100%. The frequency of utilization of 81 (85%) of the 95 substrates was similar between the microbial populations associated with meat samples stored 1 and 2 d at 4 or at 13 C. Storage of meat samples for up to 5 d at 4 C generally did not increase
the frequency of metabolism of substrates. The substrates whose frequency of metabolism varied the greatest during meat storage include cellobiose, lactulose, and xylitol in the carbohydrate group; mono-methyl succinate, αketovaleric acid, propionic acid, and sebacic acid in the carboxylic acid group; alaninamide, phenylethylamine, and putrescine in the amide or amine group; L-phenylalanine, D, L-carnitine in the amino acid group; and urocanic acid, 2-aminoethanol, 2,3-butanediol, glucose-1-phosphate, and glucose-6-phosphate in the miscellaneous group. Differences in utilization of these substrates might have been due to diverse metabolic capabilities of individual bacterial species or the interaction of multiple species at 4 and at 13 C. The frequency of metabolism of substrate groups by microbial communities associated with stored meat was comparable in most cases for plates incubated at 4 and 13 C (Table 3). Mean percentage of plates on which substrate groups were utilized by microbial populations ranged from 78 to 92% and from 68 to 86% for plates prepared from 4 C samples stored Days 1 and 2 and Days 1 through 5, respectively. Substrate groups were metabolized on 65 to 93% of plates prepared from meat samples stored 1 and 2 d at 13 C. Carbohydrates were most frequently utilized by samples at 4 C, whereas polymers were most frequently metabolized by samples at 13 C. Amides or amines and miscellaneous were the most infrequently utilized groups at the two temperatures. Because the difference in frequency of utilization of any individual substrate group by biofilm communities at 4 and 13 C was only 3 to 16%, metabolically similar bacteria may be present in meat stored at 4 C for up to 5 d and at 13 C for up to 2 d.
Statistical Comparison of Substrate Utilization A statistical analysis of the integrated microtiter plate data was performed to identify significant differences in substrate utilization profiles of biofilm communities from meat stored at 4 or 13 C. In particular, the effect of meat storage time (d) and temperature was investigated
TABLE 2. Average time (h) for initial utilization of substrate groups by biofilm communities associated with meat samples stored at 4 or 13 C Substrate group Storage conditions 4 C samples Day 11 Day 21 13 C samples Day 12 Day 23
Polymers
Carbohydrates
Carboxylic acids
Amides/ amines
Amino acids
Miscellaneous
17 24
17 18
17 23
18 19
17 28
30 (17)4 25
4 23
25 30
37 38
71 (53) 49 (31)
52 (48) 41
51 (41) 45
n (number of plates) = 14. n = 11. 3 n = 13. 4 Average incubation time excluding a single substrate requiring a lengthy incubation period (e.g., 144 h) for transformation. 1 2
1807
SUBSTRATE USE BY BIOFILM COMMUNITIES ON CHICKEN MEAT TABLE 3. Frequency of utilization (mean % positive plates) of substrate groups by biofilm communities associated with meat samples stored at 4 or 13 C Substrate group Storage conditions
Polymers
Carbohydrates
Carboxylic acids
Amides/ amines
Amino acids
Miscellaneous
4 C Samples Days 1–51 Days 1–22
83 86
86 92
82 89
71 78
85 89
68 84
13 C Samples Days 1–23
93
89
79
65
83
72
n (number of plates) = 70. n = 28. 3 n = 24. 1 2
through use of a factorial analysis of variance model. The following main effects and two-factor interactions were included in the SAS威 analysis. The main effects were 1) meat storage time (d), 2) (chicken) meat storage temperature (ctemp), 3) microtiter plate incubation temperature (itemp), and 4) density of inoculum on the microtiter plate (den). The two-factor interactions included 1) day by ctemp, 2) day by itemp, 3) day by den, and 4) itemp by den. The integrated data for each of the 95 substrates obtained using mixed microbial communities associated with meat samples are designated as y2 to y96 in the following model: proc glm; class trial day ctemp itemp den; model y2 to y96 = trial day ctemp itemp den day *ctemp day*itemp day*den itemp*den trial*day*ctemp*itemp*den; test h = day ctemp itemp den day*ctemp day*itemp day*den itemp*den e = trial *day*ctemp*itemp*den; means day ctemp itemp den day*ctemp day*itemp day*den itemp*den;. The model accounted for approximately 50 to 92% of the variation in the data for most of the 95 substrates. Main effects that were significant at P = 0.05 included plate incubation temperature (for 91 substrates), inoculum density (for 69 substrates), and the interaction of plate incubation temperature and inoculum density (for 92 substrates). Other researchers (Haack et al., 1995; Garland, 1996; Hitzl et al., 1997) have also noted the effects of incubation temperature and inoculum density on the utilization of most of the Biolog3 substrates. The P values for the main effects of plate incubation temperature and inoculum density and their interaction on selected substrates are included in Table 4. Other main effects were significant at P = 0.05 for fewer substrates than the main effects of plate incubation temperature or inoculum density and their interaction (Table 4). Meat incubation temperature was significant for one polymer (α-cyclodextrin), three carbohydrates (adonitol, i-erythritol, and D-psicose), four carboxylic acids (γ-hydroxybutyric acid, α-ketobutyric acid, α-ketovaleric acid, and sebacic acid) and one amine (phenylethylamine). Substrates affected by meat temperature at probability values between 0.06 and 0.15 include one carbohydrate, five carboxylic acids, one amide, and two amino acids (not all data shown). Metabolism of the carboxylic acids, carbohydrates, and amino acids included on the microtiter plate may thus be expected to be more variable in
poultry samples stored at 4 or 13 C than the metabolism of the polymers, amines, or amides. Day of meat storage was significant for only one substrate (phenylethylamine) at P = 0.05; the P values for four carboxylic acids, one amino acid, and one miscellaneous compound were between 0.06 and 0.15. Thus, distinct differences in metabolism of most of the substrates by microbial populations were not evident in this study, based on day of storage of the associated meat samples. The interactions of day and meat sample temperature, day and microtiter plate incubation temperature, and day and density were not significant (P = 0.05) as often as the interaction of incubation temperature and density. The interaction of day and chicken temperature was significant (P = 0.05) for only three carboxylic acids (γ-hydroxybutyric acid, α-ketobutyric acid, and sebacic acid) and one amine (phenylethylamine). For two carbohydrates and two carboxylic acids, probability values ranged between 0.08 and 0.12. The interaction of day and incubation temperature was significant (P = 0.05) for the same three carboxylic acids as the day and chicken temperature interaction. Other substrates for which this interaction was significant included one amine (phenylethylamine), one carbohydrate (i-erythritol), one amino acid (L-leucine), and one miscellaneous compound (2-aminoethanol). For two carbohydrates and one amino acid, the probability values were between 0.07 and 0.14. The interaction of day and density was significant for only 2-aminoethanol (miscellaneous group).
CONCLUSIONS The present study focused on assessing the metabolic processes of microorganisms growing on chicken meat surfaces with the ultimate goal of controlling these processes to yield safer poultry products with less potential for spoilage. Our results demonstrate that the substrate utilization profiles of bacterial communities (biofilm) as they develop on chicken meat surfaces may be assessed using an appropriate rinsing protocol and Biolog3 GN microtiter plates. Utilization of some of the 95 test substrates by these biofilm communities may begin as early as 1 d after meat storage at 4 or 13 C. However, the rapidity and frequency of substrate metabolism may vary
1808
HALE BOOTHE ET AL. 1
TABLE 4. Source of variation and significance2 for integrated substrate utilization data3 Source of variation and p value Test substrate
Day
Ctemp
Itemp
Den
Day*Ctemp
Day*Itemp
Day*Den
Itemp*Den
Polymers α-cyclodextrin
0.3448
0.0405
0.0126
0.6743
0.1948
0.1954
0.9556
0.0170
0.2864 0.3014 0.6356
0.0107 0.0435 0.0430
0.0001 0.0046 0.0001
0.0385 0.1422 0.1151
0.0923 0.0829 0.2379
0.0676 0.0370 0.1929
0.9813 0.9843 0.9350
0.0020 0.0059 0.0105
0.4446 0.7178 0.0609 0.0606 0.3707 0.0681
0.0979 0.7598 0.0009 0.0002 0.0167 0.0004
0.0001 0.1777 0.0015 0.0027 0.0354 0.0015
0.0721 0.2292 0.2435 0.0618 0.4595 0.1366
0.3862 0.0778 0.0188 0.0056 0.1244 0.0095
0.1982 0.4053 0.0176 0.0088 0.1649 0.0103
0.9468 0.9845 0.9566 0.9669 0.9907 0.9587
0.0042 0.0757 0.0196 0.0044 0.1235 0.0068
0.5007 0.0082
0.1091 0.0001
0.0002 0.0001
0.0724 0.5197
0.3183 0.0001
0.1623 0.0001
0.9872 0.8735
0.0018 0.0269
0.1184 0.6501
0.0657 0.0556
0.0001 0.0005
0.0184 0.1662
0.1762 0.2734
0.0328 0.1358
0.9543 0.9949
0.0001 0.0324
0.1491
0.3840
0.0002
0.9570
0.3746
0.0116
0.0001
0.0006
Carbohydrates adonitol i-erythritol D-psicose Carboxylic acids methyl pyruvate α-hydroxybutyric acid γ-hydroxybutyric acid α-ketobutyric acid α-ketovaleric acid sebacic acid Amines/amides alaninamide phenyethylamine Amino acids L-leucine L-threonine Miscellaneous 2-aminoethanol
Abbreviations for sources of variation are Ctemp = (chicken) meat temperature, Itemp = microtiter plate incubation temperature, Den = microtiter plate inoculum density, Day*Ctemp = interaction of meat storage time (d) and meat temperature, Day*Itemp = interaction of meat storage time and microtiter plate incubation temperature, Day*Den = interaction of meat storage time and microtiter plate inoculum density, and Itemp*Den = interaction of microtiter plate incubation temperature and inoculum density. 2 P values are for means of integrated data for substrate utilization determined as optical density at 590 nm. 3 Means of main effects and interactions for all 95 substrates may be obtained by writing to the senior author. 1
significantly, as evidenced in the present study by the 4 to 164 h of microtiter plate incubation required for utilization of some substrates and the range of 17 to 100% of microplates on which specific substrates were transformed. Utilization of chemically related substrates was less variable at 4 than at 13 C. At the latter temperature, the following order was observed for the rapidity and frequency of initial transformation of the respective substrate groups: polymers > carbohydrates > carboxylic acids > miscellaneous or amino acids > amides or amines. Data on substrate utilization obtained in this research may be useful in characterizing and controlling biofilm communities implicated in pathogenicity or affecting food quality of chicken meat samples. For example, reduction or removal of rapidly metabolized substrates in the poultry processing environment may reduce biofilm formation and persistence. Substrates may be removed by chemical sequestration or change in the pH or redox potential of fluids prevalent in poultry-processing plants. Candidate substrates may include some of the individual substrates (i.e., cellobiose in the carbohydrate group or glucose-6-phosphate in the miscellaneous group) identified in this study whose frequency of utilization differed the most between biofilm communities associated with meat samples stored at 4 or at 13 C. Removal or reduction of these substrates from the processing environment may reduce the metabolic activity of microbial populations on chicken meat samples shifted from storage at 4 to 13 C or vice versa. Other substrates commonly present in the
poultry processing environment may be tested in future work.
ACKNOWLEDGMENTS The authors thank the following individuals of USDAARS, PPMQ: Kurt Lawrence and Gavin Poole for help in SAS威 input and Manju V. Amin and Ruth A. Lebo for technical support and assistance in data collection.
REFERENCES Arnold, J. W., 1998. Development of bacterial biofilms during poultry processing. Poult. Avian Biol. Rev. 9:1–9. Arnold, J. W., and L. J. Shimkets, 1988. Inhibition of cell-cell interactions in Myxococcus xanthus by Congo red. J. Bacteriol. 170:5765–5770. Buyer, J. S., and L. E. Drinkwater, 1997. Comparison of substrate utilization assay and fatty acid analysis of soil microbial communities. J. Microbiol. Methods 30:3–11. Ellender, R. D., L. Huang, S. L. Sharp, and R. P. Tettleton, 1994. Isolation, enumeration, and identification of Gram-positive cocci from frozen crabmeat. J. Food Prot. 58:853–857. Fernandez, J. L., W. J. Simpson, and T. M. Dowhanick, 1993. Enumeration of Obesumbacterium proteus in brewery yeasts and characterization of isolated strains using Biolog GN MicroPlates and protein fingerprinting. Lett. Appl. Microbiol. 17:292–296. Garland, J. L., 1996. Analytical approaches to the characterization of samples of microbial communities using patterns of potential C source utilization. Soil Biol. Biochem. 28:213– 221.
SUBSTRATE USE BY BIOFILM COMMUNITIES ON CHICKEN MEAT Garland, J. L., and A. L. Mills, 1991. Classification and characterization of heterotrophic microbial communities on the basis of patterns of community-level-sole-carbon-sourceutilization. Appl. Environ. Microbiol. 57:2351–2359. Guckert, J. B., G. J. Carr, T. D. Johnson, B. G. Hamm, D. H. Davidson, and Y. Kumagai, 1996. Community analysis by BIOLOG: Curve integration for statistical analysis of activated sludge microbial habitats. J. Microbiol. Methods 27:183–197. Haack, S. K., H. Garchow, M. J. Klug, and L. J. Forney, 1995. Analysis of factors affecting the accuracy, reproducibility and interpretation of microbial community carbon source utilization patterns. Appl. Environ. Microbiol. 61:1458– 1468. Haack, S. K., H. Garchow, W. Sobczak, L. J. Forney, and M. J. Klug, 1994. Analysis of changes in stream biofilm community structure and function in response to environmental perturbation using whole-community fatty acid methyl ester profiles and carbon-substrate utilization patterns. Page 352 in: Abstracts of the 94th General Meeting of the American Society for Microbiology, Las Vegas, NV.
1809
Hitzl, W., A. Rangger, S. Sharma, and H. Insam, 1997. Separation power of the 95 substrates of the BIOLOG system determined in various soils. FEMS Microbiol. Ecol. 22:167–174. Johnson, J. L., and C. P. Lattuada, 1993. Comparison of nucleic acid hybridization assays and biochemical characterization tests for the confirmation of Listeria monocytogenes. J. Food Prot. 56:834–840. Praphailong, W., M. Vangestel, G. H. Fleet, and G. M. Heard, 1997. Evaluation of the Biolog system for the identification of food and beverage yeasts. Lett. Appl. Microbiol. 24:455–459. Russell, S. M., D. L. Fletcher, and N. A. Cox, 1996. Spoilage bacteria of fresh broiler chicken carcasses. Poultry Sci. 75:2041–2047. Surkiewicz, B. F., R. W. Johnston, A. B. Moran, and G. W. Krumm, 1969. A bacteriological survey of chicken eviscerating plants. Food Technol. 23:80–85. SAS Institute, 1996. SAS威 Release 6.12 for Windows. SAS Institute Inc., Cary, NC. Victorio, L., K. A. Gilbride, D. G. Allen, and S. N. Liss, 1996. Phenotypic fingerprinting of microbial communities in wastewater treatment systems. Water Res. 30:1077–1086.