J. Dairy Sci. 2008. 91:1218-1228. doi:10.3168/jds.2007-0697
© 2008 American Dairy Science Association ®
Tracking Heat-Resistant, Cold-Thriving Fluid Milk Spoilage Bacteria from Farm to Packaged Product
J. R. Huck,
M. Sonnen and
K. J. Boor1
Milk Quality Improvement Program, Department of Food Science, Cornell University, Ithaca, NY 14853
1 Corresponding author: kjb4{at}cornell.edu
 |
ABSTRACT
|
|---|
Control of psychrotolerant endospore-forming spoilage bacteria, particularly Bacillus and Paenibacillus spp., is economically important to the dairy industry. These microbes form endospores that can survive high-temperature, short-time pasteurization; hence, their presence in raw milk represents a major potential cause of milk spoilage. A previously developed culture-dependent selection strategy and an rpoB sequence-based subtyping method were applied to bacterial isolates obtained from environmental samples collected on a New York State dairy farm. A total of 54 different rpoB allelic types putatively identified as Bacillus (75% of isolates), Paenibacillus (24%), and Sporosarcina spp. (1%) were identified among 93 isolates. Assembly of a broader data set, including 93 dairy farm isolates, 57 raw milk tank truck isolates, 138 dairy plant storage silo isolates, and 336 pasteurized milk isolates, identified a total of 154 rpoB allelic types, representing an extensive diversity of Bacillus and Paenibacillus spp. Our molecular subtype data clearly showed that certain endospore-forming bacterial subtypes are present in the dairy farm environment as well as in the processing plant. The potential for entry of these ubiquitous heat-resistant spoilage organisms into milk production and processing systems, from the dairy farm to the processing plant, represents a considerable challenge that will require a comprehensive farm-to-table approach to fluid milk quality.
Key Words: milk • Bacillus • Paenibacillus • bacterial subtyping
|
INTRODUCTION
|
|---|
The presence of psychrotolerant endospore-forming spoilage bacteria (e.g., Bacillus spp. and Paenibacillus spp.) in HTST-pasteurized fluid milk products has emerged as the biological barrier currently limiting shelf-life extension of these products (Meer et al., 1991; Fromm and Boor, 2004; Durak et al., 2006; Huck et al., 2007a). These bacteria are capable of forming endospores that can survive HTST pasteurization (Collins, 1981) and are able to reproduce under refrigeration temperatures (Huck et al., 2007b; Washam et al., 1977). Previous work has identified these microbes in milk samples collected within the processing continuum from raw milk tanker trucks (Huck et al., 2007a) to pasteurized packaged products (Fromm and Boor, 2004; Huck et al., 2007b). Although molecular subtyping data from previous tracking studies provide evidence in support of pasteurization survival for these microbes, suggesting their entry in raw milk (Huck et al., 2007a,b), processing plant contamination sources also have been identified (te Giffel et al., 1997; Eneroth et al., 1998, 2001; Svensson et al., 2000; Huck et al., 2007a,b).
Identification of points of entry for these bacteria may allow the development of effective strategies for reducing or eliminating their presence in milk production systems. Therefore, the objective of this study was to determine whether the same strains of bacteria that have been isolated previously from HTST-pasteurized milk also can be isolated from the milk production environment (i.e., the dairy farm). To this end, we collected environmental samples from a New York State (NYS) dairy farm and used a previously developed culture-dependent selection strategy (Mikolajcik, 1978; Huck et al., 2007b) and rpoB sequence-based subtyping method (Durak et al., 2006; Huck et al., 2007a) to isolate and characterize psychrotolerant endospore-forming bacteria from these samples. To gain further insight into the ecology and potential transmission of psychrotolerant endospore-forming bacteria from the dairy farm to the processing plant, analyses were conducted on a larger data set that was assembled from rpoB subtype data obtained during this and previous related studies (Durak et al., 2006; Huck et al., 2007a,b) encompassing bacterial isolates collected from 4 NYS fluid milk-processing plants and a NYS dairy farm.
|
MATERIALS AND METHODS
|
|---|
Dairy Farm Sample Collection
To gain a better understanding of the ecology and transmission of psychrotolerant spore-forming bacterial contaminants on the dairy farm, environmental and raw milk samples were collected from a NYS dairy farm during the summer of 2006. During 2006, the farm produced approximately 9.8 million kg of milk from 700 cows milked 3 times per day in a double-14 stanchion parlor and housed in 2 free-stall barns. A total of 13,000 metric tons of grass and corn silage were harvested for feed on this 1,300-acre farm.
Overall, 11 environmental samples were collected from 5 locations on the farm including 1) feed, 2) bedding materials, 3) manure, 4) soil, and 5) milking parlor wash water. Specifically, 2 freshly blended TMR feed samples, representing 2 different production group feed formulations, were collected from the barn floor immediately following dispersal by the farm feed wagon. Approximately 1 to 2 kg of each type of feed was collected in grab samples from various points along the length of the feeders, mixed in clean plastic buckets, and placed in 2-L sample bags. Two bedding material samples (approximately 0.5 to 1 kg each), one representing the paper mill by-product bedding and one representing the kiln-dried sawdust bedding, were collected from the bedding materials storage shed. Two manure samples (approximately 1 kg each), one each from the north and south barns, were collected as the automatic manure scraper deposited one sweep of the barn alley scraper into a trench leading to the liquid manure storage lagoon. In addition, approximately 0.5 to 1 kg of material was collected from the manure or bedding pack used to house close-up dry cows in the milking barn. The bedding pack samples were mixed in clean plastic buckets and filled into 2-L sample bags. Three soil samples were obtained from the area within 3 m of each housing barn (i.e., north and south) and from the soil or gravel parking area outside the milking parlor. Representative soil samples were obtained by collecting 10 to 15 trowel samples from a depth of 0 to 8 cm (each trowel sample represented approximately 200 g). Samples were mixed in a clean plastic bucket and commingled into a 2-L sample bag. One 2-L water sample, representing chlorinated [1.5 x 10–6 Cl2/H2O (vol/vol)] on-farm reservoir water, was collected. Following surface disinfection of the hose sprayer with 70% ethanol, the sample was collected from the hose used to wash down the milking parlor. In addition to the 11 environmental samples, one 500-mL raw milk sample was collected from 1 of the farm’s 2 bulk tanks following the afternoon milking. All samples were stored on ice for transport to the laboratory.
Treatment and Microbiological Testing of Dairy Farm Environmental Samples
For isolation of psychrotolerant, endospore-forming bacteria, all samples were heat-treated, stored under refrigeration, and plated for bacterial enumeration throughout storage. All solid (i.e., not liquid) environmental samples were initially diluted (1:10 wt/wt) in PBS (Weber Scientific, Hamilton, NJ). Filtered liquid aliquots were transferred to a secondary sample bottle. Specifically, 30 g of each sample was transferred into stomacher bags (with filters; Weber Scientific) and diluted with 270 g of PBS. Diluted samples were manually blended for approximately 1 min, and 200-mL filtered aliquots were transferred into sterile 250-mL screw-capped Pyrex containers. Water samples were mixed, and 200-mL aliquots were transferred into sterile 250-mL screw-capped Pyrex containers.
All water samples and samples diluted as described above were then heated to 80°C for 12 min to reduce background flora (i.e., vegetative cells) and immediately cooled on ice to <6°C. All samples were stored at 6°C and total bacterial counts (TBC) were determined at d 1, 7, and 14 postcollection. Samples were serially diluted in PBS (Weber Scientific) and spread-plated onto brain heart infusion agar (BHI; Difco, BD Diagnostics, Franklin Lakes, NJ). In addition, on d 1 (and thereafter, if bacterial numbers were low) a 1-mL aliquot of each sample was spread-plated over 5 BHI plates to allow bacterial enumeration in samples with low bacterial numbers. Total bacterial counts were determined after incubation at 32°C for 48 h.
Microbiological Testing of Dairy Farm Raw Milk
To assess raw milk quality and the extent and diversity of raw milk contamination by psychrotolerant endospore-forming bacteria on this particular farm, a raw milk sample collected from 1 of the 2 farm bulk tanks was assessed for SPC, coliform count, laboratory-pasteurized count, psychrotrophic bacterial count, mesophilic spore count, and psychrotrophic spore count. In addition, a portion of this raw milk sample was heat-treated at 80°C for 12 min and stored at 6°C, and TBC were determined at 1, 7, and 14 d postcollection. The milk sample was serially diluted in PBS (Weber Scientific) and spread-plated onto BHI agar (Difco). In addition, on d 1 (and thereafter, if bacterial numbers were low) a 1-mL aliquot of sample was spread-plated over 5 BHI plates to allow enumeration of low bacterial numbers. Total bacterial counts were determined after incubation at 32°C for 48 h.
Bacterial Isolation
For each heat-treated 200-mL diluted sample, bacterial colonies present on the TBC plated on d 7 and 14 of cold storage were visually examined, and colonies representing each visually distinct morphology (ranging from 1 to 5 colonies per sample) were selected and streaked for purity on BHI agar. Purified isolates were characterized for Gram reaction by using a 3-step Gram stain kit (Becton, Dickinson and Co., Sparks, MD) and subsequently frozen at –80°C in 15% glycerol.
rpoB Sequencing
A total of 93 isolates were further characterized by molecular subtyping. All isolates were obtained from dairy farm samples (87 and 6 isolated from environmental and raw milk samples, respectively) plated at d 7 and 14 of cold storage.
Subtyping was performed by determining the DNA sequence for a 632-nucleotide fragment of the rpoB gene, as previously described by Huck et al. (2007b) and detailed by Durak et al. (2006), using primers described by Drancourt et al. (2004). This method was selected because it allows for phylogenetic characterization of isolates in addition to subtype identification, and because it is more economical than most banding pattern-based methods such as ribotyping or pulsed-field gel electrophoresis (Kabuki et al., 2004). The software DnaSP, version 4.0 (Rozas and Rozas, 1999) was used for rpoB allele assignment and different allelic types (AT) were assigned to gene sequences that differed from each other by one or more nucleotides. Isolates with 2 different allelic types were considered to represent 2 different molecular subtypes.
Cluster Analyses of Dairy Farm Isolate rpoB Sequencing Data
Identification of related rpoB allelic types and confirmation of genus and species for the isolates subtyped in this study was accomplished by aligning 54 rpoB sequences representing the 54 rpoB AT identified in this study with rpoB sequences representing the 113 rpoB AT previously characterized by Durak et al. (2006) and Huck et al. (2007a, b) [GenBank accession numbers EF156897 to EF156925, EF156999 to EF15702, and EF203109 to EF203119 (except EF203111) and EU147203 to EU147243]. The 113 previously identified rpoB AT that were used in this study to construct phylogenetic trees had previously been used to assign isolates to genus (Bacillus and Paenibacillus) and species where possible by using both rpoB and 16S rDNA sequence data (Huck et al., 2007a,b).
16S rDNA Sequencing
Although a number of isolates collected here could be characterized to genus and species through phylogenetic analyses by using previously reported rpoB sequence data (Durak et al., 2006; Huck et al., 2007a,b) as described above, 16S rDNA sequencing was used to confirm the identification of genus, species, or both of rpoB AT that had not been identified in any of our previous studies. Specifically, one isolate representing each newly identified Bacillus or Paenibacillus spp. rpoB AT was characterized by sequencing the 3' end of the 16S rDNA. In addition, 4 isolates that were identified as gram-positive cocci by Gram stain were similarly classified to genus, species, or both with 16S rDNA sequencing. The 16S rDNA PCR and sequencing were performed as previously described by Huck et al. (2007a). Final partial 3' 16S rDNA sequences were used for similarity searches against the National Center for Biotechnology Information (NCBI) nucleotide sequence database (nr) by using the Basic Local Alignment Search Tool (BLAST; McGinnis and Madden, 2004). Assignments of genus, species, or both for a specific 16S rDNA sequence were based on the top matches returned by the BLAST search. In cases in which the 16S rDNA sequence BLAST search could not adequately classify isolates within a given rpoB AT (i.e., the "Bacillus cereus group"), rpoB sequences for these AT were used for similarity searches against the NCBI nucleotide sequence database (nr) by using BLAST.
Ribotyping
A total of 10 isolates representing 8 rpoB allelic types were ribotyped to compare the subtype discrimination achieved against that of the rpoB DNA sequence-based subtyping method used in this and previous spore-former tracking studies (Durak et al., 2006; Huck et al., 2007a,b). Isolates were sent to the Cornell University Laboratory for Molecular Typing, and EcoRI automated ribotyping was performed as previously described (Bruce et al., 1995). Because the Riboprinter was unable to assign a DuPont ID (i.e., for a new pattern with <0.85 similarity to existing patterns in the DuPont database) for any of the Bacillus or Paenibacillus spp. isolates typed, we assigned a unique type designation based on the "ribogroup" that had been assigned by the instrument (e.g., ribogroup 116-1244-S-4). All DuPont IDs were confirmed by visual inspection. Ribotype patterns for isolates in this study are available for comparison through Pathogen Tracker (www.pathogentracker.net).
Data Curation
Isolate characteristics, sequence, and riboprint data, along with relevant isolation and sample information, can be accessed through the Pathogen Tracker database (www.pathogentracker.net; Fugett, 2006). GenBank accession numbers for 16S rDNA and rpoB sequences of isolates representing each rpoB AT identified in this study are available in a supplemental table online at http://jds.fass.org/content/vol91/issue3/.
Combined rpoB Sequencing Data Analysis
To gain further insight into the ecology and potential transmission of psychrotolerant endospore-forming bacteria from the dairy farm to the processing plant, an rpoB sequence data set was assembled by using data obtained during this and previous related studies (Durak et al., 2006; Huck et al., 2007a,b). This combined data set contained rpoB sequence data for 154 rpoB allelic types found among 624 bacterial isolates collected from 4 NYS fluid milk-processing plants and 1 NYS dairy farm. Samples were obtained from various sampling points at these processing plants, including 1) raw milk samples collected from incoming raw milk tank trucks and 2) raw milk storage silos; 3) pasteurized milk samples obtained inline directly after the HTST pasteurizer and 4) inline just prior to selected filling units; and 5) from pasteurized packaged products. All 624 bacterial isolates were categorized by sample location (i.e., dairy farm, raw milk tank trucks, processing plant storage silos, and packaged pasteurized milk product samples) for further analysis.
|
RESULTS
|
|---|
Microbiological Analysis of Dairy Farm Raw Milk
To determine the extent and diversity of raw milk contamination by spore-forming bacteria on the NYS farm surveyed, a raw milk sample was collected from 1 of the 2 farm bulk tanks on the same day as environmental sampling. The raw milk sample was determined to have an SPC of 2,080 cfu/mL, a coliform count of 45 cfu/mL, a laboratory-pasteurized count of 12 cfu/mL, a psychrotrophic bacterial count of 20 cfu/mL, and a mesophilic spore count of 10 cfu/mL and showed no growth when plated for psychrotrophic spore count. A portion of this raw milk sample was also heat-treated at 80°C for 12 min and stored at 6°C for up to 14 d. On d 1, bacteria were not detected in the sample (<1 cfu/ mL), whereas samples plated on d 7 and 14 exhibited TBC of 1.79 and 2.94 log cfu/mL, respectively, indicating growth of psychrotolerant spore-forming bacteria in the heat-treated raw milk tank sample.
Microbiological Analysis of Dairy Farm Environmental Samples
Total bacteria counts of the diluted and heat-treated environmental samples were plotted to assess bacterial numbers over storage at 6°C (Figure 1
). On d 1, TBC ranged from 2.57 to 6.08 log cfu/g, indicating the initial presence of large numbers of spore-forming bacteria in all environmental sample types. Day 1 TBC varied among sample types (i.e., cow feed, cow bedding materials, manure, soil, and wash water). Day 1 TBC were lowest in the pond water and sawdust bedding samples (2.57 and 2.82 log cfu/g, respectively) and highest in the manure samples, ranging from 5.88 to 6.00 cfu/g (average 5.93 cfu/g), followed by the soil samples, with TBC ranging from 5.56 to 6.08 cfu/g (average 5.88 cfu/ g; Figure 1).
View larger version (25K):
[in this window]
[in a new window]
|
Figure 1. Total bacterial counts obtained from environmental and raw milk samples collected at one New York State dairy farm, heat-treated at 80°C for 12 min, held at 6°C up to 14 d, and plated at d 1, 7, and 14 of cold storage.
|
|
Total bacteria counts increased over time under refrigerated storage at 6°C for some diluted and subsequently heat-treated environmental samples. Specifically, TBC of the 3 diluted manure samples increased by an average of 2.70 cfu/g from d 1 to 14, whereas TBC of 2 of the 3 diluted soil samples increased by an average of 1.75 cfu/g during the same period (Figure 1
). Total bacterial counts of the wash water sample increased slightly, by 0.51 cfu/g, from d 1 to 14, whereas diluted bedding and feed samples showed no increase in bacterial numbers over 14 d of storage at 6°C.
rpoB Sequencing of Dairy Farm Isolates
A total of 103 isolates were selected for rpoB sequencing from d-7 and d-14 TBC plates (representing 17, 17, 24, 26, 9, and 10 isolates from heat-treated bedding, feed, manure or bedding pack, soil, water, and milk samples, respectively). Among these 103 isolates, 93 yielded rpoB PCR products that could be sequenced. For 6 isolates, sufficient growth could not be obtained from cultures that had been frozen; therefore, subtyping was not performed. Four isolates that did not yield an rpoB PCR product were identified as gram-positive cocci and were characterized by sequencing the 3' end of the 16S rRNA gene.
Overall, 54 rpoB AT (i.e., subtypes based on different rpoB sequences) were differentiated among the 93 isolates (16, 17, 22, 24, 8, and 6 isolates from bedding, feed, manure or bedding pack, soil, water, and milk samples, respectively). Although no one rpoB AT was found in all sample types (i.e., in bedding, feed, soil, manure, water, and milk), 4 rpoB AT were found in more than 2 sample types, whereas 11 rpoB AT were found in more than 1 sample type. A total of 43 rpoB AT (representing 51% of all isolates subtyped) were found in only 1 sample type (Table 1).
View this table:
[in this window]
[in a new window]
|
Table 1. Contamination patterns by endospore-forming bacteria isolated from environmental and raw milk samples obtained from a New York State dairy farm
|
|
Cluster Analysis of rpoB DNA Sequence Data
The rpoB sequences were used in phylogenetic analyses to identify groups of related subtypes (i.e., clusters) and to confirm genus and species identification of isolates. Specifically, 54 rpoB sequences, representing the 54 rpoB AT identified in this study, were aligned with rpoB sequences for the 113 rpoB AT previously characterized by our group (Durak et al., 2006; Huck et al., 2007a,b). A total of 41 new rpoB AT, representing 62 isolates, were identified from dairy farm raw milk and environmental samples in this study. Thirteen rpoB AT, representing a total of 31 isolates (25 Bacillus and 6 Paenibacillus spp.) collected during this study, had been identified previously in fluid milk products (Durak et al., 2006; Huck et al., 2007a,b).
16S rDNA Sequencing
The 16S rDNA sequencing was performed on 45 isolates to provide or confirm taxonomic identification. Specifically, 41 isolates, representing each of the 41 newly identified rpoB AT (i.e., rpoB AT that had not been identified in previous studies, as described above) and the 4 isolates identified as gram-positive cocci were characterized by sequencing the 3' end of the 16S rDNA gene. Forty-five, 16, and 1 of the 62 isolates representing newly identified rpoB AT were identified as Bacillus spp., Paenibacillus spp., and Sporosarcina spp., respectively. The 4 gram-positive cocci were identified as Staphylococcus spp.
Ribotyping
A total of 10 isolates (2 Bacillus and 8 Paenibacillus spp.), representing 8 rpoB AT, were ribotyped to compare the subtype discrimination achieved by this method against that obtained from the rpoB DNA sequence-based subtyping method used in this and previous spore-former tracking studies (Durak et al., 2006; Huck et al., 2007a,b). Two isolates, H8-017 and H8-094 (rpoB AT27 and AT32), were each ribotyped twice. Imaging results confirmed the reproducibility of this method for typing this group of bacteria. Although ribo-type patterns for Bacillus and Paenibacillus spp. were very distinct, only subtle differences were observed among the ribotype patterns for the 8 Paenibacillus isolates (Figure 2). Although the 2 Paenibacillus isolates classified into rpoB AT15 represented 2 different ribtypes, the 3 isolates within rpoB AT27 and the 2 isolates within rpoB AT32 showed identical ribotype patterns. Thus, although EcoRI ribotyping discriminates subtypes within certain rpoB AT, the large number of weak bands in the EcoRI ribotype patterns makes pattern interpretation difficult, especially for analyses of large data sets. Our data thus suggest that although ribotyping can provide reproducible and interpretable subtype discrimination, rpoB sequence-based subtyping provides a more economical and efficient tool for molecular subtype analysis of large isolate sets of Bacillus and Paenibacillus spp. isolates.
View larger version (52K):
[in this window]
[in a new window]
|
Figure 2. Riboprint image patterns for 10 selected bacterial isolates representing 8 different rpoB allelic types (2 Bacillus spp. and 6 Paenibacillus spp.).
|
|
Ecology of Psychrotolerant Spore-Forming Bacteria in the Dairy Farm Environment
A total of 54 rpoB allelic types were identified among the 93 isolates that were obtained from the 11 dairy farm environmental samples (87 isolates) and 1 bulk tank raw milk sample (6 isolates). For the 2 clean bedding samples, 7 and 6 rpoB AT were identified among the 9 and 7 isolates found in the sawdust and shredded paper mill waste bedding, respectively; only AT1 (Bacillus licheniformis) was found in both bedding samples (Table 1). A total of 11 rpoB AT were identified from the 17 isolates obtained from the 2 TMR feed samples. Three of these AT [AT1, AT6 (B. licheniformis), and AT20 (Bacillus pumilus)] were found in both feed samples. For the 2 manure samples and 1 manure or bedding pack sample, 22 isolates were obtained, representing 12 rpoB AT; only AT123 (Paenibacillus spp.) and AT145 (Bacillus spp.) were identified in all 3 samples. The soil samples collected in the areas adjacent to the milking parlor and housing barns provided the greatest bacterial diversity among environmental samples, yielding 19 rpoB AT among the 24 isolates collected; only rpoB AT123 was found in more than one soil sample. Interestingly, AT123 was also identified in shredded paper bedding, all 3 manure or bedding pack samples, and the milking parlor wash water sample. The wash water sample, representing water pumped from the farm pond, yielded 8 rpoB AT among the 8 isolates collected. Of these 8 allelic types, AT20, AT123, AT128, and AT129 were also identified in soil samples, suggesting that the surrounding soil is the most likely contamination source of the milking parlor wash water, which is drawn from this storage pond.
One 500-mL raw milk sample was collected from 1 of the 2 farm bulk tanks on the same day as environmental sampling to determine the diversity of spore-forming bacteria present in the farm’s raw milk. A total of 6 spore-forming isolates (representing 6 rpoB AT) were obtained from the heat-treated raw milk samples. Allelic type 15 (Paenibacillus spp.) was identified in raw milk and one of the manure samples, suggesting manure as a possible source of raw milk contamination by spore-forming bacteria on this farm.
Ecology of Psychrotolerant Spore-Forming Bacteria from Farm to Packaged Product
To gain further insight into the ecology and potential transmission of psychrotolerant endospore-forming bacteria from the dairy farm to the processing plant, a larger data set was assembled using the Pathogen Tracker database and data obtained from this and previous related studies (Durak et al., 2006; Huck et al., 2007a,b). The data set contains rpoB sequence data for 154 rpoB subtypes, representing 624 bacterial isolates collected from 4 NYS fluid milk-processing plants and 1 NYS dairy farm.
Overall, the 624 isolates were identified as Bacillus spp. (37%), Paenibacillus spp. (62.6%), Oceanobacillus spp. (0.2%), and Sporosarcina spp. (0.2%). A higher proportion of Bacillus spp. was identified on the farm (87%) than from raw milk tank trucks (8.8%), processing plant raw milk silos (48.2%), and pasteurized milk samples (23.8%; Figure 3).
View larger version (40K):
[in this window]
[in a new window]
|
Figure 3. Distribution of Bacillus and Paenibacillus spp. isolates obtained from 4 sampling locations: (A) the dairy farm (n = 92), (B) raw milk tank trucks (n = 57), (C) dairy plant raw milk storage silos (n = 137), and (D) pasteurized milk (n = 336) collected from 4 New York State dairy plants (on d 12 of shelf life or beyond).
|
|
A total of 4 rpoB AT (AT1, B. licheniformis; AT15, Paenibacillus spp.; AT21, Paenibacillus spp.; AT27, Paenibacillus spp.) were found in samples collected from the dairy farm, raw milk tank trucks (plants M and C), processing plant raw milk storage silos (plants M and U), and HTST-pasteurized milk products (plants M, C, U, and Q; Figures 4 and 5). These 4 AT represent 27% of the 624 isolates included in this data set, with AT15 being the most frequently isolated (11.8% of isolates), followed by AT27, AT1, and AT21 (6.8, 5.8, and 2.6% of all isolates, respectively; Figure 3). Allelic types 15, 21, and 27 were isolated from raw milk tank truck samples at both plants M and C, whereas AT1, AT15, and AT27 were isolated from processing plant raw milk silo samples at both plants M and U. In addition, all 4 rpoB AT (AT1, AT15, AT21, and AT27) were identified on the NYS dairy farm sampled in this study as well as in the raw and pasteurized milk samples from all 4 plants previously sampled, suggesting the potential for these major contributors for fluid milk spoilage to be introduced into the raw milk supply at the farm. Thirteen of the 54 rpoB AT representing 31 isolates (AT1, AT6, AT7, AT15, AT20, AT21, AT27, AT29, AT62, AT65, AT68, AT69, and AT97) isolated from the NYS dairy farm had been previously isolated from raw and pasteurized milk samples (Durak et al., 2006; Huck et al., 2007a,b; Figure 4). These 13 rpoB AT, which represent 35% of all 624 isolates collected from this and previous studies, were identified as B. licheniformis (AT1 and AT6), B. pumilus (AT20, AT62, AT68, and AT69), Bacillus subtilis (AT65), Bacillus weihenstephanensis (AT97), Paenibacillus amylolyticus (AT29), and Paenibacillus spp. (AT7, AT15, AT21, and AT27),
View larger version (17K):
[in this window]
[in a new window]
|
Figure 4. Nonproportional Venn diagram showing the distribution of rpoB allelic types isolated from dairy farm (F) samples (n = 93), raw milk tank truck (T) samples (n = 57), raw milk storage silo (S) samples (n = 138), and pasteurized milk (P) samples (n = 336). Letters in each square indicate sample types in which these rpoB allelic types were commonly identified.
|
|
View larger version (19K):
[in this window]
[in a new window]
|
Figure 5. Distribution of rpoB allelic types isolated from 3 or 4 of the sample locations, including the (1) dairy farm, (2) raw milk tank trucks, (3) raw milk storage silos, and (4) pasteurized milk samples. Shaded bars indicate the number of isolates of each rpoB allelic type that were isolated from (A) the dairy farm (n = 93), (B) raw milk tank trucks (n = 57), (C) dairy plant raw milk storage silos (n = 138), and (D) pasteurized milk (n = 336) collected from 4 New York State dairy plants (on d 12 of shelf life or beyond).
|
|
|
DISCUSSION
|
|---|
In this study we applied to dairy farm environmental samples a previously developed culture-dependent method for isolating psychrotolerant endospore-forming bacteria from milk. The subtype diversity of endospore-forming bacteria isolated from the dairy farm environment was analyzed within the context of a larger data set of bacterial isolates that had been collected throughout the milk-processing continuum. Overall, our data indicate that considerable subtype diversity of endospore-forming bacteria is present in multiple locations on the farm, representing likely sources of raw milk contamination.
Isolation of Psychrotolerant Endospore-Forming Bacteria from Both Raw Milk and the Dairy Farm Environment
Development and application of a universal, economical method for isolating psychrotolerant endospore-forming bacteria from raw milk is essential for gaining a better understanding of the ecology and transmission of these fluid milk spoilage microorganisms. The most widely used method, an 80°C heat treatment, originally described by Mikolajcik (1978) for isolating heat-resistant psychrotrophs from raw milk, has been applied by many, including Crielly et al. (1994), Huck et al. (2007a), Shehata and Collins (1971), and te Giffel et al. (2002), to isolate spore-forming bacteria from dairy farm environmental samples and raw milk. As further demonstrated in this study, the application of a heat treatment to diluted environmental samples (i.e., soil, manure, and animal feeds), followed by a period of cold incubation (6°C), eliminates background flora (i.e., vegetative cells) and promotes the growth of psychrotolerant endospore-forming microorganisms. Data from this study validate the use of this isolation method, in conjunction with rpoB DNA sequence-based subtyping, as a valuable set of tools for tracking psychrotolerant endospore-forming bacterial contaminants through milk production and processing systems, including from the dairy farm environment.
rpoB Sequencing Allows for Rapid and Economical Characterization and Subtyping of Endospore-Forming Dairy Spoilage Organisms
Application of molecular subtyping methods, including pulsed-field gel electrophoresis, ribotyping, and other banding pattern-based methods, has been critical for gaining a better understanding of the transmission of foodborne pathogens and for reducing transmission of these pathogens (Swaminathan et al., 2001). Recent advances in the development of DNA sequencing-based subtyping methods, including the development of an rpoB sequencing-based subtyping method for Bacillus and Paenibacillus spp., now offer improved opportunities to explore the transmission and ecology of these organisms with an economical subtyping method. A comparison of results from automated EcoRI ribotyping and rpoB DNA sequence-based subtyping suggests that whereas ribotyping can provide reproducible and interpretable subtype discrimination, rpoB sequence-based subtyping provides a much more economical and efficient tool for molecular subtype analysis of large collections of Bacillus and Paenibacillus spp. isolates. In addition, the continued development and availability of an rpoB sequence database that includes a diverse collection of >600 Bacillus and Paenibacillus spp. isolates (this study; Durak et al., 2006; Huck et al., 2007a,b), comprising >150 rpoB AT that have also been characterized by 16S rDNA sequencing, allows for characterization of many Bacillus and Paenibacillus isolates to the genus and species level. Even if species-level identification is not possible, which is currently typical for many Paenibacillus isolates, rpoB sequence data still allow for identification of distinct Paenibacillus clusters, which are likely to represent different species. Furthermore, these sequencing-based subtyping strategies allow simple Web-based data curation as well as reliable and robust comparison of subtype data generated by different research groups (Aanensen and Spratt, 2005).
Dairy Farm Environment as Host to a Diversity of Endospore-Forming Bacteria Contributing to Fluid Milk Spoilage
Overall, our data indicate that a considerable diversity of endospore-forming bacteria, including those subtypes previously identified in raw and pasteurized milk (Durak et al., 2006; Huck et al., 2007a,b) exist in the dairy farm environment. In particular, B. licheniformis, B. pumilus, B. subtilis, B. weihenstephanensis, P. amylolyticus, and Paenibacillus spp. represent a large proportion of the endospore-forming bacteria isolated in the production chain from dairy farm to HTST-pasteurized fluid milk products (Huck et al., 2007a). Bacillus spp., as well as Paenibacillus spp., are known for their ability to form heat-resistant endospores that can survive heat-treatment processing conditions (Collins, 1981; Huck et al., 2007b). Although a number of studies have identified Bacillus spp. in milk from the farm (Crielly et al., 1994; Sutherland and Murdoch, 1994; Lukasova et al., 2001) and the processing plant, including in processing plant raw milk supplies (Lin et al., 1998; Huck et al., 2007b) and in pasteurized packaged products (Griffiths and Phillips, 1990; Lin et al., 1998; Douglas et al., 2000; Fromm and Boor, 2004; Durak et al., 2006; Huck et al., 2007b), the present study used DNA sequence-based subtyping to characterize these spoilage bacteria and illustrate the potential contamination route from the dairy farm environment to packaged HTST-pasteurized fluid milk products. This on-farm survey provides direct subtype-based evidence that endospore-forming bacteria, including those able to grow under refrigeration temperatures, exist in the dairy farm environment (i.e., cow feed, cow bedding materials, manure, soil, and wash water) and represent a potential source of raw milk contamination, ultimately contributing to HTST-pasteurized fluid milk spoilage. This route of transmission from raw milk to pasteurized products is consistent with previous studies tracking spore-forming contaminants from raw milk tank trucks to finished fluid milk products (Huck et al., 2007a).
The data reported here indicate that Bacillus spp., as well as Paenibacillus spp., are commonly present in the dairy farm environment. These microorganisms may contaminate the udder or milking equipment and ultimately contaminate the raw milk supply on the farm. Aerobic spore-formers of the genus Bacillus are ubiquitous and have been isolated from dairy farm environments, including silage (te Giffel et al., 2002), pasture (Slaghuis et al., 1997), soil (Christiansson et al., 1999), and fecal material (Labots et al., 1965). Environmental contamination of udders (Lukasova et al., 2001) and milking equipment (McKinnon and Pettipher, 1983) may lead to the presence of Bacillus spp., including B. licheniformis and B. cereus, in bulk tanks on dairy farms (Griffiths and Phillips, 1990; Crielly et al., 1994). The present study determined that a portion of these endospore-forming bacteria found in the dairy farm environment exhibit the ability to grow under refrigeration temperatures. Thus, contamination of fluid milk with endospore-forming spoilage microorganisms may occur at the dairy farm, with the potential for recontamination to occur at various points in the production continuum from farm to finished product.
|
CONCLUSIONS
|
|---|
As the dairy industry progresses toward extending the shelf life of conventionally pasteurized HTST fluid milk products, the presence of endospore-forming bacteria capable of surviving pasteurization and growing at refrigeration temperatures is the specific biological hurdle limiting shelf-life extension past 14 d (Fromm and Boor, 2004; Durak et al., 2006; Huck et al., 2007a,b). Our data clearly show that although a considerable diversity of psychrotolerant endospore-forming bacteria are present throughout the fluid milk production continuum, from farm to package, some subtypes appear to be particularly more frequently encountered. Molecular typing methods, such as rpoB subtyping, provide the reproducibility and sensitive discrimination needed to identify and track psychrotolerant endospore-forming bacteria through dairy-processing systems. Ultimately, further extension of fluid milk product shelf lives will require the elimination of these spoilage bacteria from the raw milk supply. Because our data suggest multiple potential entry points for psychrotolerant endospore-forming bacteria into the raw milk at the farm, as well as various points of recontamination throughout the production chain, implementation of control strategies for these spoilage organisms represents a considerable challenge that will require a comprehensive focus on the entire farm-to-table continuum.
|
ACKNOWLEDGEMENTS
|
|---|
This project was supported by the New York State Milk Promotion Advisory Board through the New York State Department of Agriculture and Markets (Albany, NY)—NYS dairy farmers committed to the production of high-quality dairy products. The authors thank the management and employees of the NYS dairy farm involved in the study as well as Martin Wiedmann for helpful conversations and critical review of the manuscript, and staff members of the Cornell University Milk Quality Improvement Program for technical support.
Received for publication September 17, 2007.
Accepted for publication November 20, 2007.
|
REFERENCES
|
|---|
Aanensen, D. M. and B. G. Spratt. 2005. The multilocus sequence typing network: mlst.net. Nucleic Acids Res. 33(Web Server issue):W728–W733.[Abstract/Free Full Text]
Bruce, J., R. J. Hubner, E. M. Cole, C. I. McDowell, and J. A. Webster. 1995. Sets of EcoRI fragments containing ribosomal RNA sequences are conserved among different strains of Listeria monocytogenes. Proc. Natl. Acad. Sci. USA 92:5229–5233.[Abstract/Free Full Text]
Christiansson, A., J. Bertilsson, and B. Svensson. 1999. Bacillus cereus spores in raw milk: Factors affecting the contamination of milk during the grazing period. J. Dairy Sci. 82:305–314.[Abstract]
Collins, E. B. 1981. Heat resistant psychrotrophic organisms. J. Dairy Sci. 64:157–160.[Abstract/Free Full Text]
Crielly, E. M., N. A. Logan, and A. Anderton. 1994. Studies on the Bacillus flora of milk and milk products. J. Appl. Bacteriol. 77:256–263.[Medline]
Douglas, S. A., M. J. Gray, A. D. Crandall, and K. J. Boor. 2000. Characterization of chocolate milk spoilage patterns. J. Food Prot. 63:516–521.[Medline]
Drancourt, M., V. Roux, P. E. Fournier, and D. Raoult. 2004. rpoB gene sequence-based identification of aerobic Gram positive cocci of the genera Streptococcus, Enterococcus, Gemella, Abiotrophia and Granulicatella. J. Clin. Microbiol. 42:497–504.[Abstract/Free Full Text]
Durak, M. Z., H. I. Fromm, J. R. Huck, R. N. Zadoks, and K. J. Boor. 2006. Development of molecular typing methods for Bacillus spp. and Paenibacillus spp. isolated from fluid milk products. J. Food Sci. 71:50–56.
Eneroth, A., A. Christiansson, J. Brendehaug, and G. Molin. 1998. Critical contamination sites in the production line of pasteurised milk, with reference to the psychrotrophic spoilage flora. Int. Dairy J. 8:829–834.[CrossRef]
Eneroth, A., B. Svensson, G. Molin, and A. Christiansson. 2001. Contamination of pasteurized milk by Bacillus cereus in the filling machine. J. Dairy Res. 68:189–196.[CrossRef][Medline]
Fromm, H. I., and K. J. Boor. 2004. Characterization of pasteurized fluid milk shelf-life attributes. J. Food Sci. 69:207–214.
Fugett, E. B. 2006. Development of molecular subtyping databases to improve control of Listeria monocytogenes. MS Thesis. Cornell University, Ithaca, NY.
Griffiths, M. W., and J. D. Phillips. 1990. Incidence, source and some properties of psychrotrophic Bacillus spp. found in raw and pasteurized milk. J. Soc. Dairy Technol. 43:62–66.[CrossRef]
Huck, J. R., B. H. Hammond, S. C. Murphy, N. H. Woodcock, and K. J. Boor. 2007a. Tracking spore-forming bacterial contaminants in fluid milk processing systems. J. Dairy Sci. 90:4872–4883.[Abstract/Free Full Text]
Huck, J. R., N. H. Woodcock, R. D. Ralyea, and K. J. Boor. 2007b. Molecular subtyping and characterization of psychrotolerant endospore-forming bacteria in two New York State fluid milk processing systems. J. Food Prot. 70:2354–2364.[Medline]
Kabuki, D. Y., A. Y. Kuaye, M. Wiedmann, and K. J. Boor. 2004. Molecular subtyping and tracking of Listeria monocytogenes in Latin-style fresh-cheese processing plants. J. Dairy Sci. 87:2803–2812.[Abstract/Free Full Text]
Labots, H., G. Hup, and T. E. Galesfoot. 1965. Bacillus cereus in raw and pasteurized milk. III. The contamination of raw milk with Bacillus cereus spores during its production. Neth. Milk Dairy J. 19:191–221.
Lin, S., H. Schraft, J. A. Odumeru, and M. W. Griffiths. 1998. Identification of contamination sources of Bacillus cereus in pasteurized milk. Int. J. Food Microbiol. 43:159–171.[CrossRef][Medline]
Lukasova, J., J. Vyhnalkova, and Z. Pacova. 2001. Bacillus species in raw milk and in the farm environment. Milchwissenschaft 56:609–611.
McGinnis, S. and T. L. Madden. 2004. BLAST: At the core of a powerful and diverse set of sequence analysis tools. Nucleic Acids Res. 32(Web Server issue):W20–W25.[Abstract/Free Full Text]
McKinnon, C. H., and G. L. Pettipher. 1983. A survey of sources of heat-resistance bacteria in milk with particular reference to psychrotrophic spore-forming bacteria during transport, storage and processing. J. Dairy Res. 50:163–170.[Medline]
Meer, R. R., J. Baker, F. W. Bodyfelt, and M. W. Griffiths. 1991. Psychrotrophic Bacillus spp. in fluid milk products: A review. J. Food Prot. 54:969–979.
Mikolajcik, E. M. 1978. Psychrotrophic sporeformers: A possible keeping-quality problem in market milk. Am. Dairy Rev. 40:34A–34D.
Rozas, J., and R. Rozas. 1999. DnaSP version 3: An integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics 15:174–175.[Abstract/Free Full Text]
Shehata, T. E., and E. B. Collins. 1971. Isolation and identification of psychrotrophic strains of Bacillus from milk. Appl. Environ. Microbiol. 21:466–473.[Abstract/Free Full Text]
Slaghuis, B. A., M. C. T. Giffel, R. R. Beumer, and G. Andre. 1997. Effect of pasturing on the incidence of Bacillus cereus spores in raw milk. Int. Dairy J. 7:201–205.[CrossRef]
Sutherland, A. D., and R. Murdoch. 1994. Seasonal occurrence of psychrotrophic Bacillus species in raw milk, and studies on the interactions with mesophilic Bacillus spp. Int. J. Food Microbiol. 21:279–292.[CrossRef][Medline]
Svensson, B., A. Eneroth, J. Brendehaug, G. Molin, and A. Christiansson. 2000. Involvement of a pasteurizer in the contamination of milk by Bacillus cereus in a commercial dairy plant. J. Dairy Res. 67:455–460.[CrossRef][Medline]
Swaminathan, B., T. J. Barrett, S. B. Hunter, and R. V. Tauxe. 2001. PulseNet: The molecular subtyping network for foodborne bacterial disease surveillance, United States. Emerg. Infect. Dis. 7:382–389.[Medline]
te Giffel, M. C., R. R. Beumer, L. P. M. Langeveld, and F. M. Rombouts. 1997. The role of heat exchangers in the contamination of milk with Bacillus cereus in dairy processing plants. Int. J. Dairy Technol. 50:43–47.[CrossRef]
te Giffel, M. C., A. Wagendorp, A. Herrewegh, and F. Driehuis. 2002. Bacterial spores in silage and raw milk. Antonie Leeuwenhoek 81:625–630.[CrossRef][Medline]
Washam, C. J., H. C. Olson, and E. R. Vedamuthu. 1977. Heat-resistant psychrotrophic bacteria isolated from pasteurized milk. J. Food Prot. 40:101–108.
TOP
ABSTRACT
INTRODUCTION
