Farm level survey of spore‐forming bacteria on four dairy farms in the Waikato region of New Zealand

Abstract The aim of our study was to determine the occurrence and diversity of economically important spore‐forming bacteria in New Zealand dairy farm systems. Farm dairy effluent (FDE) collected from Waikato dairy farms were tested for the presence of spore‐forming bacteria, using a new culture‐based methodology followed by genomic analysis. An enrichment step in which samples were inoculated in cooked meat glucose starch broth under anaerobic conditions, aided in the differential isolation of Bacillus and Clostridium species. Furthermore, the use of molecular methods such as ERIC genotyping, 16S rRNA gene sequence analysis identified different spore‐forming bacteria present in FDE. C. sporogenes signature PCR gave further information on the phylogenetic relationship of the different Clostridium spp. isolated in this study. In total 19 Bacillus spp., 5 Paenibacillus spp. and 17 Clostridium spp. were isolated from farm dairy effluent. Sequence types similar to economically important food spoilage bacteria viz: C. butyricum, C. sporogenes and members of the Paenibacillus Genus were isolated from all four farms, whereas, sequence types similar to potential toxigenic, B. cereus, C. perfringens, C. butyricum, and C. botulinum were found on at least three of the farms. Sampling of farm dairy effluent provides a good indicator of farm level prevalence of bacterial load as it is used to irrigate dairy pasture in New Zealand. This study highlights the presence of various spore‐forming bacteria in dairy waste water and indicates the implementation of good hygienic farm practices and dairy waste effluent management.

from the farm environment as well as from heat-treated milk highlighting the importance of the environment on spore resistance, survival and germination properties (Gerhardt & Marquis, 1989;Guillaume-Gentil et al., 2002;Scheldeman, Pil, Herman, De Vos, & Heyndrickx, 2005). As new products and technologies have been developed, concerns have also been raised around the impact of spore contamination on quality and safety in minimally processed and shelf stable dairy foods (Guinebretiere, Girardin, Dargaignaratz, Carlin, & Nguyen-The, 2003;Peck, 2006;Ranieri, Huck, Sonnen, Barbano, & Boor, 2009).
Moreover, if extensive proteolysis occurs during cheese ripening, the release of amino acids and increase in pH will favor the growth of many Clostridium species, especially C. tyrobutyricum (Klijn, Nieuwendorf, Hoolwerf, van derWaals, & Weerkamp, 1995). Silage has been implicated as the principal source of spores of ruminant feed (Vissers, Driehuis, Te Giffel, De Jong, & Lankveld, 2007a,b) and according to a study by Klijn et al. in 1995, C. tyrobutyricum was found to be one of the commonly isolated Clostridium spp. in silage samples. Other Clostridium species commonly found in silage include C. beijerinckii and C. acetobutylicum. Importantly, the reduction or elimination of these bacteria can result in an extended shelf life for pasteurized milk providing an overall higher quality product.
Human food poisoning caused by spore-forming bacteria are often associated with heat-treated foods subjected to mishandling and temperature abuse during storage, which results in spore germination, bacterial multiplication, and food consumption with hazardous levels of cells or toxins. For Bacillus spp., B. cereus is considered the most important human pathogen because of the ability of some strains to produce toxins (Ghelardi et al., 2002;Turnbull et al., 2002).
Common Clostridium pathogenic species include C. perfringens, C. difficile, C. botulinum, with C. perfringens being the most commonly associated with raw milk products (McAuley, McMillan, Moore, Fegan, & Fox, 2014). The majority of the Clostridium spp. is of relevance to the dairy industry possess the metabolic ability to reduce sulphite to sulphide under anaerobic conditions and are termed as sulphite reducing Clostridia (SRCs) (Prevot, 1948;Weenk, Van den Brink, Struijk, & Mossel, 1995). SRCs are often used by the dairy industry as a hygienic indicator (Dodds, 1993;Aureli & Franciosa, 2002;New Zealand Ministry of Primary Industries (MPI), 2014) but the reduction of sulphite to sulphide is not a discriminatory factor between species.
Very little is known about the impact of farm management practices on the relative abundance and diversity of spore-forming bacteria in dairy farm systems. The dairy industry relies on pasteurization to control the number of pathogenic and spoilage microorganisms.
As pasteurization is relatively ineffective against spores (Griffiths & Phillips, 1990) it is important to understand how on farm practices may increase or reduce risk for raw bulk milk contamination. Silage type and quality have been implicated as the major primary source of spore-forming bacteria in bulk tanker milk (Garde, Gaya, Arias, & Nuñez, 2012;Julien, Dion, Lafreniere, Antoun, & Drouin, 2008;Vissers, Te Giffel, Driehuis, De Jong, & Lankveld, 2007;Vissers et al., 2007a,b). Other factors that may influence silage quality include the starting material, pH, dry matter content, fermentation conditions and microbial content (Rammer, 1996;Vissers et al., 2007a,b); Vissers, Te Giffel, et al 2007. Current thinking is that raw milk becomes contaminated through consumption of lower grade silage by herd, followed by the survival of spores in the bovine gastrointestinal tract resulting in contaminated faeces. Subsequent fecal contamination of teats and udder surfaces then result in contamination of raw milk particularly if good hygienic practice is not followed (Aureli & Franciosa, 2002).
Whether the contamination of bulk milk with spore-forming bacteria can be eliminated is unknown. Annually, New Zealand dairy companies process around a billion litres of milk, most of which will go overseas as whole milk powder to be used globally as ingredients or for the production of infant formula. If studies to investigate the diversity of spore-forming bacteria and their prevalence on farm are to generate mitigation strategies to control their entry to the food chain, simple tools for their rapid detection and differentiation are required.
This study aimed at investigating diversity of spore-forming bacteria in farm dairy effluent (FDE) as cultures from dairy effluent provide an indication of microbial strains that may be cycling on the dairy farm. We developed a new culture based methodology to detect and separate Clostridium and Bacillus species anaerobically. To date, we believe that no study has been undertaken to investigate the occurrence of sporeforming bacteria at a farm level in New Zealand.

| Study sites
A cross-sectional pilot study was carried out on four Waikato dairy farms. All the 4 study farms had some component of pasture grazing; three were pasture-only and in one, the cows had access to HerdHomes ® -like facility (http://herdhomes.co.nz). Analysis was carried out on a "whole herd" basis by collecting samples of FDE from the exit point of the milking parlor at final wash down.

| Sample processing
Farm dairy effluents were collected in July and September 2014 (late winter early spring) and in January 2015 (summer). A one litre grab sample of FDE was collected in a sterile Schott Duran glass bottle from the shed wash collection sump immediately after morning milking. Samples were transported to the laboratory in an insulated box and kept frozen until processed. For ease of recording and to retain confidentiality of individuals, farms were allocated a number.
Twenty ml of farm dairy effluent was dissolved in 250 ml of prewarmed Butterfield's diluents (Fort Richard, New Zealand) and centrifuged at 3466g for 1 hr. The supernatant was removed and pellet suspended in 5 ml of prewarmed Butterfield's diluent. To isolate spores from vegetative cells, the respective suspensions were heated at 80°C for 15 min in a water bath.

| Bacterial isolates
Clostridium sporogenes NCTC 532 spores were used as a control and to develop the new methodology.

| New culture-based methodology for
isolating mesophilic spore-forming bacteria from environmental samples

| Aerobic spore formers
A 1 ml aliquot from each of the heated sample suspensions described previously was serially diluted in Butterfield's diluent and plated directly in duplicate on Sheep Blood Agar (SBA) (Fort Richard, New Zealand). Plates were incubated under aerobic conditions at 35°C for 24 hr and then colonies enumerated. This preparation was termed, heated and direct (HD).

| Anaerobic spore formers
A 1 ml aliquot of each of the heated sample suspensions was added to 9 ml of prereduced cooked meat glucose broth (Fort Richard) supplemented with casein (0.03%), L-cysteine (0.0005%), Haemin (0.1%), Vitamin K1 (1%) and yeast extract (0.0005%) for sample enrichment and incubated under anaerobic conditions at 35°C for 3 day. This treatment was termed, heated and enriched (HE). Enriched cultures were removed and transferred to fresh sterile centrifugation tubes. The samples were centrifuged at 6000g for 15 min and the pellets were re-suspended in 2 ml of Butterfield's diluent. A 1 ml preparation from each of the enriched suspensions was serially diluted in Butterfield's diluent and plated in duplicate on Shahidi-Ferguson Perfringens agar (Shahidi & Ferguson, 1991) with 50% egg yolk, polymyxin B (3 mg l −1 ) and kanamycin (12 mg l −1 ) (EYA). Plates were incubated under anaerobic conditions at 35°C for 24-48 hr. Enumeration of anaerobic spore formers was not carried out as the samples were enriched in growth medium for 3 days.
All the colonies from both HD, and HE plates (black as well as white colonies on EYA plates) were further sub cultured on Sheep blood agar for obtaining pure cultures. Each of the cultures was further inoculated in thioglycollate broth (Fort Richard) for genomic analysis.

| DNA extraction
A boiled lysate of each of the cultures was prepared by boiling the culture at 100°C for 10 min and collecting the supernatant after centrifugation at 9838g for 5 min. The supernatant was used as a source of template DNA for Enterobacterial Repetitive Intergenic Consensus (ERIC) PCR. Genomic DNA was isolated, using the Roche High Pure PCR template preparation kit (Roche diagnostics, Manheim, Germany), according to the manufacturer's instructions and used for amplification of the 16S rRNA gene from spore-forming bacteria.

| 16S rRNA gene amplification and Sequencing
Amplification of the 16S rRNA gene was carried out using forward primer sequence pA 5′-AGAGTTTGATCCTGGCTCAG-3′ (Invitrogen) and reverse primer sequence pH* 5′-AAGGAGGTGATCCAGCCGCA-3′ (Invitrogen) as described by Boddinghaus, Wolters, Heikens, and Bottger (1990). Pure DNA isolated from the Roche High Pure PCR template preparation kit was used for the reaction. Each 50 μl PCR mixture contained 0.2 mmol/L of each dNTPs (Invitrogen), 1 μmol/L of forward and reverse primers, 1X reaction buffer (Invitrogen) with 2 mmol/L of MgCl2 (Invitrogen), 1.25 U of Taq polymerase (Invitrogen) and 5 μl of DNA. All the reagents were procured from Life Technologies and nuclease free water was used in all the reactions.
PCR was carried out in a PTC-100 ™ Programmable Thermal Cycler (MJ Research Inc), using the following conditions: 93°C for 3 min; 92°C for 1 min, 55°C for 1 min and 72°C for 2 min for 30 cycles followed by a final extension at 72°C for 3 min. PCR products were visualized on a 0.8% ultrapure agarose (GibcoBRL) gel stained with ethidium bromide (Bio-Rad, 10 mg l-1). The PCR products were purified, using Qiagen DNA extraction kit as per manufacturer's instructions (Qiagen, Bio-strategy Ltd, New Zealand) and the products sequenced, using an ABI3730 DNA Analyzer (Massey Genome Service, Palmerston North, New Zealand) and the primers described above. 16S rRNA gene consensus sequences were used to investigate the phylogenetic relationship of the spore-forming bacteria obtained from the samples, using

| Phylogenetic analysis
Phylogenetic analysis was carried out, using the software Geneious version 8.1 by Biomatters. A Polar formatted, unrooted phylogenetic tree was created from the converted 16s rRNA gene sequence alignment of all the representative isolates identified during this study and botulinum Group I, II and III to investigate genetic distance between the sequences. 16S rRNA gene sequences of botulinum Groups I, II and III were obtained from Genbank (http://www.ncbi.nlm.nih.gov/ genbank/) and PATRIC (Pathosystems Resource Integration Centre, https://www.patricbrc.org). Alignment of all 16S rRNA gene sequences was prepared, using Muscle alignment tool with UPGMA clustering. The most likely tree found using a General Time Reversible model with optimized nucleotide equilibrium frequencies, optimized invariable sites, optimized across site variation, and NNI tree-searching operations was bootstrapped with 100 replicates. Branches with less than 50% support were collapsed, using TreeCollapserCL version 4 (Hodcroft, 2013) (Weigand et al., 2015). Each 25 μl PCR mixture contained 0.2 mmol/L of each dNTPs (Invitrogen), 1 μmol/L of forward and reverse primers, 1X reaction buffer (Invitrogen) with 1 mmol/L of MgCl 2 (Invitrogen), 1.25 U of Taq polymerase (Invitrogen) and 2 μl of DNA. All reagents were procured from Life Technologies and nuclease free water was used in all the reactions. PCR amplification was carried out in a T100 ™ Bio-Rad Thermal Cycler, using the following conditions: 95°C for 3 min; 95°C for 30 s, 53°C for 30 s and 72°C for 90 s for 35 cycles followed by a final extension at 72°C for 10 min. PCR products were visualized on a 2% ultrapure agarose (GibcoBRL) gel stained with ethidium bromide (Bio-Rad, 10 mg l −1 ). Farm 1, did not show any difference in number of colony-forming units ml −1 in winter (20 × 10 3 CFU ml − 1, July 14) and summer (23 × 10 3 CFU ml − 1, Jan 15). Farm 2 had 25 × 10 5 in winter (July 14) and 38 × 10 4 CFU ml −1 in summer (Jan 15), Farm 3 had 21 × 10 5 in winter (July 14) and 28 × 10 3 CFU ml −1 in summer (Jan 15) whereas, Farm 4 did not show huge difference between two seasons; 8 × 10 3 July 14 and 28 × 10 3 Jan 15 CFU ml −1 , respectively (Table 3).

| ERIC PCR genotypic fingerprinting and 16S rRNA sequencing
ERIC PCR was able to differentiate between isolates cultured from all FDE samples which aided in selecting unique representatives for 16S rRNA gene sequencing (Tables 1 and 2). The maximum identity for the 16S rRNA gene sequences obtained in this study in comparison to type strains ranged from 92 to 100% (Tables 1 and 2) therefore, species were identified on the basis of closest-related taxonomically described species.

| Occurrence and diversity of aerobic sporeforming bacteria in FDE
In total 19 different mesophilic Bacillus spp., five different Paenibacillus spp. and six non Bacillus spp. (identified by 16S rRNA gene closest taxonomically described species) were isolated from the four farms ( from Farm 2 in winter were different to Farm 4 and none of those species were common to both the farms (Table 1).
Bacillus strains with a 16S rRNA sequence type most closely matching B. licheniformis were the most prevalent and also gave rise to the most ERIC types per farm, followed by B. altitudinis (Table 3).

| Phylogenetic analysis of Clostridium botulinumlike species
A phylogenetic tree was created, using PhyML tree builder in Geneious version 8.1 from the converted 16S rRNA gene sequence alignments.
The tree shows four distinct clustered sets consisting of isolates closely related to botulinum Group I and II, along with few isolates clustering in a completely different set. Four of the isolates clustered in botulinum Group I were found to be closely related to C. sporogenes type strain ATCC 15579 where as one isolate was closely related to C. sporogenes subsp. tusciae which clustered in a different set. The remaining nine isolates clustered under botulinum Group I were found to be closely related to C. botulinum B1 strain Okra. Twelve isolates were grouped in botulinum Group II in which five were closely related to C. butyricum E4, two were closely related to C. botulinum E3 strain Alaska and five isolates were closely related to C. botulinum B strain Eklund as well as C. botulinum E1 strain BONT E Beluga (Figure 1).
Phylogenetic analysis was also performed, using PCR primer sets to four-specific orthologs from the C. sporogenes lineage (Weigand et al., 2015). All farm isolates with 16S rRNA gene sequences closely related to either botulinum Group I and II representatives were tested, using the Sporogenes Signature PCRs (Figure 1 and 2) and the phylogenetic relationship between the Sporogenes Signature Positive isolates compared with Sporogenes Signature Negative, botulinum Group II and Group III isolates are shown in Figure 1. representatives did not generate any amplicons, using the Sporogenes Signature PCR assays (Figures 1 and 2 . 6,8,10,15,16,17,19, 22,27,29,38,44,45 identified from one farm during the summer sampling event and other isolates were found in winter. All the dapL positive isolates were obtained from a single farm (Farm 1) during the winter sampling event (Figure 1). Total count (cfu/ml) 20 × 10 3 25 × 10 5 21 × 10 5 8 × 10 3 23 × 10 3 38 × 10 4 28 × 10 3 28 × 10 3 Prevalence was based on the percentage of each bacterial species (identified by 16srRNA sequencing) from the total number of colonies isolated from farm dairy effluent samples collected from each farm.

| DISCUSSION
In New Zealand, FDE is used to irrigate dairy pasture as it is a resource full of nutrients and when managed properly increases pasture production. However, dairy effluent also contains bacteria excreted from the dairy cow and hence provides a good indicator of farm level prevalence of different bacteria. Isolation of environmental spore-forming bacteria is notoriously challenging. A major challenge has been to separately isolate Clostridium spp. and Bacillus spp. in anaerobic conditions. This is  (Lukasova, Vyhnalkova, & Pacova, 2001;Sutherland & Murdoch, 1994;Tatzel, Ludwig, Schleifer, & Wallnofer, 1994) and B. cereus being the most common psychrotolerant species (Sutherland & Murdoch, 1994 Bacillus sequence types were found on all farms on at least one of the sampling events. Our results were found to be concordant to a study which showed a remarkable diversity of aerobic spore-forming bacteria in dairy farms and B. licheniformis amongst the most common ones (Scheldeman et al., 2005). In our study, we isolated a range of aerobic spore-forming bacteria from farm dairy effluent used for irrigation Bacillus species have been found in liquid milk and other milk products (Griffiths, 1992). Toxin producing B. cereus strains can cause emetic or diarrhoeal food poisoning, while diarrhoeal toxin is produced as a result of spore germination and outgrowth in the small intestine, the emetic toxin is produced by vegetative cells of B. cereus growing in the preheat-treated milk (Kramer & Gilbert, 1989). However, the present cross-sectional study was carried out investigating the diversity and occurrence of dairy associated spore-forming bacteria at the farm level and determination of toxicity was not in the scope of this study.
B. cereus, B. licheniformis and Paenibacillus spp. are all psychrotrophic thermophilic sporeformers and a particular problem for the dairy industry. They can grow at refrigeration temperatures and often multiply and sporulate in the bulk milk (Murphy, Lynch, & Kelly, 1999).
Their spores can then survive various heat treatments and processing and may go on to cause food poisoning, or reduce the shelf life of pasteurized milk and dairy products (Te Giffel, Beumer, Granum, & Rombouts, 1997). B. licheniformis is often the most frequently isolated bacterial contaminant in raw milk (Phillips & Griffiths, 1986;and Crielly, Logan, & Anderton, 1994;Miller et al., 2015). Moreover, some strains of this species have been found to show enhanced growth in skim milk under anaerobic environment (Ronimus et al., 2003).
There is very little data on the occurrence, diversity and seasonal distribution of Clostridium species on dairy farms. Clostridium species are ubiquitous in soil and presence in feed is common source of contamination of raw milk. Commonly found species are C. sporogenes, C. butryicum, C. tyrobutyricum, C. disporicum, and C. saccharolyticum. Feligini, Panelli, Sacchi, Ghitti, and Capelli (2014) investigated the occurrence of Clostridium spp. in raw milk from Northern Italy. They found that C. butyricum and C. sporogenes were more abundant in winter, C. tyrobutyricum in spring and C. beijerinckii in summer. Similarly, in this study, C. butyricum and C. sporogenes were also found to be abundant in winter than in summer and were representatives from botulinum Groups I and II. From a food safety perspective, C. perfringens is considered significant because of the ability of some strains to induce illness (Grass, Gould, & Mahon, 2013) and have also been recurrently identified as the causative agents of mastitis in bovine and other ruminants (Osman, El-Enbaawy, Ezzeldeen, & Hessein, 2009 Ahiko,1990). In this study, C. butyricum sequence types (maximum identity 96.7%-100%) were identified on all four farms in winter and on Farm 2 during summer. C. sporogenes sequence types (maximum identity 93%-100%) were identified on three farms (2, 3, and 4) during winter but only on Farms 2 and 4 during summer.
Currently there is not enough data to determine if seasonal distribution of Clostridium and Bacillus species on dairy farms impacts on the risk of contamination of raw milk and downstream processed products.
In 2014  However, the problem arises when the product is rehydrated and kept in anaerobic conditions at nonrefrigeration temperatures, which allow the C. botulinum spores to germinate and produce toxins.
The use of plastic-wrapped and non-acidified silage as cattle feed has increased the number of botulism outbreaks over the last two decades due to botulinum Groups I-III in dairy cattle (Böhnel & Gessler, 2013;Lindström, Myllykoski, Sivelä, & Korkeala, 2010). In this study, representatives of botulinum Group I and II were identified on all farms, highlighting the presence of these spores in the dairy farm environment.
Consumption of silage contaminated with spores by cattle was found to be the main reason of raw milk getting contaminated with spore-forming bacteria (Vissers, Driehuis, Te Giffel, De Jong, & Lankveld, 2006). These spores, survive in the gastrointestinal tract of animals and contaminate the manure by shedding spores in feces which subsequently results in contaminating teats and udder surfaces causing contamination of raw milk during the process of milking (Bergère, Gouet, Hermier, & Mocquot, 1968). Some of the isolates identified in this study may have potential to spoil food or produce toxin. There is always a possibility that if appropriate farming practices are not conducted, these spores may enter raw milk from the dairy environment and can cause problems as described by Driehuis, 2013. This study did not undertake any risk assessment to investigate the possibility of the contamination of raw milk by these isolates. However, this could be a future scope for another valuable study to understand the contamination level of raw milk in dairy farms.
Good farming practices are perhaps the most significant element in controlling spore numbers in raw milk. The use of good quality silage, cleaning, and maintenance of parlor/milking equipment, as well as a stringent udder cleaning and teat preparation before milking are all considered to be good farming practice. Consequently, the management of dairy waste effluent is also a critical component to consider for the reduction in number and spread of spore-forming bacteria on dairy farms. Furthermore, careful application of dairy waste effluent is particularly important if the farm or neighboring areas are used for wild harvest of foods such as water cress or horticultural operations.

| FUTURE DIRECTIONS
To investigate seasonal variation in the number of spore-forming bacteria and their diversity, more samples from different farms will be analyzed during different seasons.