Bacterial community dynamics during cold storage of minced meat packaged under modified atmosphere and supplemented with different preservatives

Food Microbiology 48 (2015) 192e199

Contents lists available at ScienceDirect

Food Microbiology journal homepage: www.elsevier.com/locate/fm

Bacterial community dynamics during cold storage of minced meat packaged under modified atmosphere and supplemented with different preservatives J. Stoops a, c, *, 1, S. Ruyters b, c, 1, P. Busschaert b, c, R. Spaepen a, c, C. Verreth b, c, J. Claes a, c, B. Lievens b, c, L. Van Campenhout a, c a

KU Leuven, Faculty of Engineering Technology, Department of Microbial and Molecular Systems (M2S), Lab4Food, Campus Geel, B-2440 Geel, Belgium KU Leuven, Faculty of Engineering Technology, Department of Microbial and Molecular Systems (M2S), Laboratory for Process Microbial Ecology and Bioinspirational Management (PME&BIM), Campus De Nayer, B-2860 Sint-Katelijne-Waver, Belgium c KU Leuven, Leuven Food Science and Nutrition Research Centre (LFoRCe), B-3001 Leuven, Belgium b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 8 September 2014 Received in revised form 21 November 2014 Accepted 27 December 2014 Available online 20 January 2015

Since minced meat is very susceptible for microbial growth, characterisation of the bacterial community dynamics during storage is important to optimise preservation strategies. The purpose of this study was to investigate the effect of different production batches and the use of different preservatives on the composition of the bacterial community in minced meat during 9 days of cold storage under modified atmosphere (66% O2, 25% CO2 and 9% N2). To this end, both culture-dependent (viable aerobic and anaerobic counts) and culture-independent (454 pyrosequencing) analyses were performed. Initially, microbial counts of fresh minced meat showed microbial loads between 3.5 and 5.0 log cfu/g. The observed microbial diversity was relatively high, and the most abundant bacteria differed among the samples. During storage an increase of microbial counts coincided with a dramatic decrease in bacterial diversity. At the end of the storage period, most samples showed microbial counts above the spoilage level of 7 log cfu/g. A relatively similar bacterial community was obtained regardless of the manufacturing batch and the preservative used, with Lactobacillus algidus and Leuconostoc sp. as the most dominant microorganisms. This suggests that both bacteria played an important role in the spoilage of minced meat packaged under modified atmosphere. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Minced meat Modified atmosphere packaging Preservatives Pyrosequencing

1. Introduction Minced meat is susceptible to microbial spoilage because of its high concentration of nutrients and high water activity (Ercolini et al., 2011). The deteriorative effects caused by bacterial growth are discolouration, off-odours, and slime production (Singh et al., 2011). The rate of deteriorative changes depends primarily on the meat composition, the hygienic practices during the grinding and packaging process, and the storage conditions (Limbo et al., 2010). Modified atmosphere packaging (MAP) is generally recognised as an effective method for food preservation. A typical modified atmosphere for storage of ground red meat is 80% O2 and 20% CO2

* Corresponding author. Lab4Food, Kleinhoefstraat 4, B-2440 Geel, Belgium. Tel.: þ32 14 56 23 10; fax: þ32 14 58 48 59. E-mail address: [email protected] (J. Stoops). 1 J.S. and S.R. contributed equally to this work. http://dx.doi.org/10.1016/j.fm.2014.12.012 0740-0020/© 2015 Elsevier Ltd. All rights reserved.

(McMillin, 2008). The positive effects of a high oxygen level (>50%) are related to colour retention of red meat, minimising drip losses and inhibition of anaerobic and microaerophilic microorganisms (Amanatidou, 2001). However, shelf life extension of meat also depends on the presence of CO2 (Limbo et al., 2010; Belcher, 2006; Dhananjayan et al., 2006; Amanatidou, 2001), with concentrations of 20e40% commonly used to suppress microbial growth €rkroth, 2007). Bacterial spoilage of meat is (Vihavainen and Bjo generally caused by species such as Brochothrix thermosphacta and species from the genus Pseudomonas, as well as by members of the Enterobacteriaceae and lactic acid bacteria (LAB) (Pennacchia et al., 2011). Their abundance and contribution to spoilage is largely influenced by the storage conditions (De Filippis et al., 2013; Doulgeraki et al., 2012; Ercolini et al., 2011; Pennacchia et al., 2011; Borch et al., 1996). Bacterial spoilage of meat is commonly determined using culture-dependent techniques such as viable plate counts, assessing the bacterial load and viability in the sample

J. Stoops et al. / Food Microbiology 48 (2015) 192e199

(Degirmencioglu et al., 2012; Esmer et al., 2011; Limbo et al., 2010; Koutsoumanis et al., 2008). It is generally accepted that microbial spoilage of meat occurs when counts reach levels of 7 log cfu/g (Degirmencioglu et al., 2012; Koutsoumanis et al., 2008). This level is commonly found to be correlated with sensory deterioration, like off-odours and the presence of slime in vacuum or gas packaged meat products (Limbo et al., 2010; Rao and Sachindra, 2002). Viable counts, however, are not suitable to characterise the microbial diversity of food products and to investigate thoroughly shifts in the bacterial communities during storage (Doulgeraki et al., 2012; Ercolini et al., 2006). Indeed, culturable microorganisms represent only a small fraction of the entire microbial diversity, and hence the microbial diversity in terms of species richness and abundance is grossly underestimated by culture-dependent techniques (Wintzingerode et al., 1997). In contrast, a thorough analysis of the microbial community can be achieved by culture-independent  et al., 2008), such as Polymerase Chain Reactionmethods (Juste Denaturing Gradient Gel Electrophoresis (PCR-DGGE) (Doulgeraki et al., 2012). This technique has, for example, been commonly applied to investigate shifts in the bacterial community during storage of beef in different conditions (Ercolini et al., 2011; Pennacchia et al., 2011; Brightwell et al., 2009; Ercolini et al., 2006; Fontana et al., 2006). However, as techniques such as PCRDGGE do not allow identification, sequence-based approaches such as 454 pyrosequencing (Margulies et al., 2005) are currently increasingly used for detailed characterisation of diverse microbial communities from different ecological niches, including, among many others, meat and meat products (De Filippis et al., 2013; Kiermeier et al., 2013; Xiao et al., 2013; Nieminen et al., 2012a, 2012b; Ercolini et al., 2011). Whereas the dynamics of the bacterial community of beef products have been studied before, little to nothing is known about differences in microbial load and dynamics during storage among different production batches. In addition, the effect of different preservatives on the microbial diversity and numbers remains to be investigated. Hence, the aim of this study was to investigate the dynamics of the bacterial community of minced meat during cold storage under modified atmosphere as well as to investigate the effect of different preservatives. This was assessed on three batches of minced meat sampled at three different manufacturing periods, and on one batch of minced meat using four different preservatives. Culture-dependent and culture-independent techniques were used. Microbial counting was performed to determine the number of culturable bacteria present. In addition, 454 pyrosequencing was used to gain insight into the bacterial community dynamics during storage. 2. Material and methods 2.1. Experiments This study included two separate experiments. The first experiment was performed to compare the dynamics of the bacterial community of minced meat during cold storage under modified atmosphere of different production batches produced over several months. Therefore, three batches of minced meat, originating from three different manufacturing periods (batch 1, 2 and 3 produced in May, June and August 2013, respectively) were used. In a second experiment, the effect of different preservatives on the dynamics of the bacterial community of minced meat during storage was investigated. 2.2. Meat samples Minced meat samples were obtained from a local meat processing company (Geel, Belgium). The samples composed of 98%

193

beef and inulin as a fat replacer. Additionally, all samples of the first experiment were supplemented with 2.0% of a default preservative applied by the meat processing company (NaL, Opti.Form® SA (56e59% (w/w) sodium lactate and 3.5e3.7% (w/w) sodium acetate), Brenntag N.V., Deerlijk, Belgium). In the second experiment, samples were supplemented with the default preservative (2.0% NaL) or one of three other additives. The other additives were 1.6% (w/w) potassium lactate (KL, Opti.Form® PPA plus (71.3e74.3% (w/ w) potassium lactate and 4.9e5.5% (w/w) potassium acetate), Brenntag N.V.), 2.0% (w/w) spice extract (SE, Misocarine LR, Barentz, Zaventem, Belgium) and a combination of 2.0% (w/w) NaL and 0.08% (w/w) ascorbic acid (AA, Brenntag N.V.). According to the manufacturer, the spice extract was a natural product based on the fermentation of spices (onion) and glucose syrup by lactic acid bacteria and acetic acid bacteria.

2.3. Packaging and storage Minced meat samples of 300 g were placed in a polypropylene tray (PS001117, 178  138  55 mm3, ES-Plastic GmbH, Hutthurm, Germany) and sealed with a foil (Rockguard 12-4-PET, adhesive, and Rockpeel-45-HT-AF, Rockwell Solutions Ltd, Dundee, UK) using a tray sealer (E 365 VG, G. Mondini S.p.A, Cologne, Italy). All samples were packaged using a gas mixture of 66% O2, 25% CO2 and 9% N2 by vacuum compensation (vacuum pressure of 200 mbar). The oxygen transmission rate of the foil was 8 cm3/ (m2.d.bar) at 23  C. The ratio between the volume of gas and the volume of product was 2:1 (Sivertsvik et al., 2004). To investigate the effect of microbial differences between production batches, 23 packages were prepared for each batch. Five packages were allocated for monitoring the gas composition during storage (five biological replicates). The other packages were used for microbial analyses. At each sampling point, three packages of minced meat were analysed (three biological replicates). For the second experiment, 15 packages were prepared for each preservative. Again, five packages were allocated to evaluate the gas composition (five biological replicates). Two packages were used for the microbial analyses at each sampling point (two biological replicates). For the study to be representative for storage of the meat products by consumers, packaged samples were stored in a home type refrigerator (Inspire Freestone, Electrolux, set point for temperature 5  C) for a period of 9 days. This storage period is more than the estimated shelf life of the products stored under these conditions (7 days). At the end of storage, discolouring of the meat occurred and a spoilage odour was perceived. The temperature was monitored during the entire storage period using a data logger (Escort iLog internal sensor, VWR International, Leuven, Belgium). The average storage temperature for the first experiment, concerning different batches, was 5.9 ± 0.8, 5.1 ± 1.1 and 5.8 ± 0.6  C for batch 1, 2 and 3, respectively. For the experiment including different preservatives, the average temperature during storage was 5.5 ± 1.5  C.

2.4. Gas analysis During storage the headspace gas composition was measured with a gas analyser (Checkpoint O2/CO2, PBI Dansensor, Ringsted, Denmark). To measure the gas composition, the needle (diameter 0.5 mm) of the gas analyser was pierced through a septum (reusable type, diameter 15 mm, DPI Dansensor) placed on the foil of the tray. Gas analyses were performed at the start of the experiment (day 0), as well as after 1, 6, 7, 8 and 9 days of packaging.

194

J. Stoops et al. / Food Microbiology 48 (2015) 192e199

2.5. Plate counts At each measuring point for gas analysis (except at day 0 in the second experiment including different preservatives), meat samples were subjected to microbial plating. Psychrotrophic aerobic and anaerobic counts were determined according to the ISO standards for microbial analysis of food as compiled by Dijk et al. (2007). At each sampling time, a sample of 20 g was transferred aseptically into a sterile stomacher bag and 180 ml of sterile peptone physiological salt solution (0.85% (w/v) NaCl, 0.1% (w/v) peptone, Biokar Diagnostics, Beauvais, France) was added. The mixture was homogenised for 60 s in a stomacher (Bagmixer® 400W, Interscience, St Nom, France). Next, a serial dilution series was prepared. Psychrotrophic aerobic and anaerobic counts were determined on Plate Count Agar (PCA, Biokar Diagnostic) using the pour plate technique. All plates were incubated at 6.5  C for 10 days. Anaerobic counts were determined by incubating plates in jars containing a GasPak™ EZ Anaerobe Container System Sachet (Becton Dickinson, Erembodegem, Belgium) and an Anaerobic Indicator (Oxoid, Erembodegem, Belgium). All microbial counts were expressed as log cfu/g. Afterwards, the mean log cfu/g value and the standard deviation were calculated. 2.6. Microbial assessment using 454 pyrosequencing In addition to classical plating, subsamples were subjected to molecular analysis at regular time points (i.e. 1, 7 and 9 days after packaging; for the first experiment concerning different production batches day 0 was also included). Samples were prepared as described by Nieminen et al. (2011). Briefly, 5 g of minced meat was homogenised in 45 ml peptone physiological salt solution for 60 s in a stomacher. Subsequently an aliquot of 40 ml of the solution was centrifuged at 800 g and 4  C for 5 min. Next, 35 ml of the supernatant was collected and centrifuged at 12,000 g and 4  C for 10 min. The pellet containing the DNA was suspended in 1.5 ml of 0.15 M NaCl and centrifuged at 12,000 g and 4  C for 5 min. DNA was extracted from the pellet obtained using the phenol chloroform method (Lievens et al., 2003). Subsequently, amplicon libraries were created using the barcode-tagged fusion primers 577F (forward primer 50 - AYTGGGYDTAAAGNG-30 ) and 926R (reverse primer 50 - CCGTCAATTCMTTTRAGT-30 ) targeting the bacterial 16S ribosomal RNA gene fragment containing the V4 and V5 variable region (Rosenzweig et al., 2012). More specifically, the forward fusion primer (55 bp) consisted of the “A” adaptor, amended with a unique Roche barcode and the original primer sequence (577F). The reverse fusion primer (48 bp) consisted of the “B” adaptor, amended with the original primer sequence (926R). PCR reactions were performed on a Biorad T100 thermal cycler in a 40 mL reaction volume, containing 0.15 mM of each dNTP (Invitrogen, Merelbeke, Belgium), 0.5 mM of each primer, one unit Titanium Taq DNA polymerase, 1  Titanium Taq PCR buffer (Clontech Laboratories, Inc., Palo Alto, CA, USA) and 10 ng genomic DNA (measured using a Nanodrop instrument (Thermo Scientific Nanodrop Products Inc., Wilmington, DE, USA)). Samples were subjected to an initial denaturation at 94  C for 2 min followed by 35 cycles of denaturation at 94  C for 45 s, primer annealing at 59  C for 45 s, extension at 72  C for 45 s, and a final extension step at 72  C for 10 min. After resolving the amplicons by agarose gel electrophoresis, amplicons within the expected size range were excised and extracted/purified from the gel using the Qiaquick gel extraction kit (Qiagen, Hilden, Germany). Purified dsDNA amplicons were then quantified with the Qubit fluorometer according to the manufacturer's instructions (Invitrogen). For each experiment, a separate library was constructed using equimolar concentrations (1.00  109 molecules/ml) of each amplicon. The quality of the resulting libraries was assessed

using an Agilent Bioanalyzer 2100 with high-sensitivity chip (Agilent Technologies, Waldbronn, Germany). Each amplicon library was separately sequenced using the 454/Roche GS FLX instrument with Titanium Lib L chemistry according to the manufacturer's instructions and resulted in a total of 93,600 and 114,408 reads for the first experiment and the second experiment, respectively. Using Cutadapt 1.0 (Martin, 2011), the sequences obtained were assigned to the corresponding sample based on both barcode and primer sequences, allowing zero discrepancies, and were subsequently trimmed from the barcodes and primers. Sequence processing and clustering were performed with Mothur version 1.32.1 (Schloss et al., 2009). Sequences were trimmed based on a minimum Phred score of 30 averaged over a 50 bp moving window. The minimum trimmed sequence length was set to 300 nt, the maximum trimmed sequence length was set to 350 nt. Sequences containing ambiguous base calls or homopolymers longer than 8 nucleotides were rejected, as were mitochondrial sequences and chimeras (detected using the Uchime chimera detection program (Edgar et al., 2011)). Filtered sequences were then aligned to the SILVA reference alignment, and sequences starting after position 15,647 or ending before position 27,659 of the reference alignment were eliminated from the dataset. Finally, a distance matrix was calculated based on the aligned sequences (using a cut off distance value of 0.40), and sequences were clustered with the average neighbour algorithm at a sequence identity level of 97%, a commonly used cut-off value to determine species level OTUs (Stackebrandt and Goebel, 1994). In order to minimise the risk of retaining sequences that resulted from sequencing errors, “global” singletons (OTUs represented by only a single sequence in the entire dataset) were removed. Removing these OTUs has previously been shown to improve accuracy of diversity estimates (Kunin et al., 2010; Waud et al., 2014). Remaining OTUs were assigned taxonomic identities, based on BLAST results of the OTU representative sequences (selected by Mothur) using the GenBank nucleotide (nt) database (Altschul et al., 1990; Benson et al., 2008), excluding uncultured/environmental entries. The ten most dominant OTUs were also verified using the EzTaxon database (Kim et al., 2012) and yielded similar results. For some OTUs no species could be assigned because different species were obtained with similar sequence similarity. 2.7. Statistical analysis SPSS (IBM© SPSS Statistics ver. 20, New York, USA) was used for statistical analyses. One-way ANOVA was performed to determine the effect of the different manufacturing batches and the use of different preservatives on the headspace gas composition, the microbial counts and the number of OTUs for each sampling time. Multiple comparison was performed by Tukey's post hoc test. For all statistical analyses, a significance level of 0.05 was considered. 3. Results 3.1. Microbial quality and community dynamics of different production batches of minced meat For this experiment, minced meat batches produced in three different months, including May, June and August, were assessed when packaged under modified atmosphere and stored at refrigerate temperature. Discolouration and off-odour were observed after 9 days of storage for all samples. The initial headspace gas composition (i.e. immediately after packaging) of the minced meat samples consisted of 71.9e74.0% O2 and 15.2e16.6% CO2. During storage, samples of all batches showed a decrease in O2 and increase in CO2 (Fig. S1A). Significant changes (p < 0.05) in the O2 and

J. Stoops et al. / Food Microbiology 48 (2015) 192e199

Number of OTU

A 220

Batch 1 Batch 2 Batch 3

200 180 160 140 120 100 80 60 40 20 0

0

1 7 Storage time (days)

B 220

Number of OTU

CO2 concentrations were observed between day 0 and day 9. On day 9, O2 and CO2 concentration ranged between 53.3e62.4% and 18.6e23.6%, respectively. The initial aerobic numbers were 4.6, 4.9, and 4.7 log cfu/g for batch 1, 2 and 3, respectively (Fig. 1A). Similar counts were obtained for the anaerobic microorganisms (ranging between 4.3 and 4.8 log cfu/g). Both initial microbial counts significantly (p  0.01) differed among the three batches. Within each batch no statistically significant difference was found between day 0 and day 1. After 6 days of storage, an overall increase in microbial counts was noticed for all samples. In batch 2 and 3, both microbial groups increased to above 7 log cfu/g. Aerobic and anaerobic loads increased even more after 9 days of storage and the significantly highest counts were observed in batch 2 (aerobic counts of 8.5 log cfu/g and anaerobic counts of 8.4 log cfu/g). Pyrosequencing of the 16S rRNA gene resulted in 58,671 filtered sequences across all samples (average of 1630 sequences per sample). Fig. 2A shows the number of OTUs (OTU richness) for all batches at day 0, 1, 7 and 9. These results demonstrated a diverse microflora of fresh meat (day 0 and day 1) with 85, 116 and 61 different OTUs for batch 1, 2 and 3, respectively. OTU richness was not significantly different between the three manufacturing batches at day 0 (p ¼ 0.440) and day 1 (p ¼ 0.101). However, the most dominant bacterial species differed between the three manufacturing batches (Fig. 3A). As a result, OTU 4 (Pseudomonas

195

9

NaL KL SE NaL+AA

200 180 160 140 120 100 80 60 40 20 0 1

7 Storage time (days)

9

Fig. 2. Number of Operational Taxonomic Units (OTUs, 3% sequence dissimilarity cutoff) in minced meat packaged under modified atmosphere during cold storage for different production batches (A, data are expressed as the mean of 3 replicates) and preservatives 2 (B, data are expressed as the mean of 2 replicates). Error bars represent standard errors.

sp.) and OTU 5 (Propionibacterium acnes) were the most abundant bacteria occurring at day 0 in batch 1, with an incidence of 30% and 21%, respectively. Batch 2 was initially dominated by OTU 1 (Lactobacillus algidus) (12%), OTU 2 (Leuconostoc sp.) (17%), and OTU 5 (P. acnes) (12%). The most dominant OTUs in batch 3 were OTU 1 (L. algidus) (50%) and OTU 3 (Photobacterium sp.) (27%). During storage, the number of OTUs decreased (Fig. 2A). At the end of the storage period, the microbial communities of batch 1, 2 and 3 consisted of 4, 15 and 9 OTUs. Additionally, the microbial communities in the three manufacturing batches converged to a quite similar community as shown in Fig. 3A. Three bacterial OTUs (OTU 1 (L. algidus), OTU 2 (Leuconostoc sp.) and OTU 3 (Photobacterium sp.) comprised more than 97% of the total microflora in all samples (Supplementary Information Table S1). In all batches, a high relative abundance of OTU 1 was observed (ranging from 34 to 45%). In the minced meat samples of batch 1 and 2, OTU 2 (Leuconostoc sp.) was the most predominant OTU, with a relative abundance of 65% and 59%, respectively. The microbial community of batch 3 was also dominated by OTU 3 (Photobacterium sp.), with a relative abundance of 45%.

Fig. 1. Effect of different production batches (A, data are expressed as the mean of 3 replicates) (batch 1: square, batch 2: circle and batch 3: triangle) and preservatives (B, data are expressed as the mean of 2 replicates) (NaL: diamond, KL: square, SE: triangle or a combination of NaL þ AA: circle) on aerobic and anaerobic plate counts (log cfu/g) of minced meat packaged under modified atmosphere during cold storage. Closed symbols represent aerobic counts, and open symbols represent anaerobic counts. Error bars represent standard errors.

3.2. Microbial quality and community dynamics of minced meat supplemented with various preservatives In this experiment, a batch of minced meat supplemented with one of the four different preservatives was packaged under modified atmosphere and stored at refrigerate temperature. Colour

196

J. Stoops et al. / Food Microbiology 48 (2015) 192e199

Relative abundance (%)

A

120 100

Other OTUs

OTU 10 (Psychrobacter urativorans) OTU 9 (Carnobacterium divergens)

80 60 40

OTU 8 (Pseudomonas grimontii) OUT 7 (Pseudoxanthomonas sp.) OTU 6 (Lactococcus sp.) OTU 5 (Propionibacterium acnes) OTU 4 (Pseudomonas sp.)

20 0

OTU 3 (Photobacterium sp.) OTU 2 (Leuconostoc sp.) OTU 1 (Lactobacillus algidus)

Relative abundance (%)

B

120

Other OTUs

100

OTU 10 (Psychrobacter urativorans) OTU 9 (Carnobacterium divergens)

80 60

40 20 0

OTU 8 (Pseudomonas grimontii) OUT 7 (Pseudoxanthomonas sp.)

OTU 6 (Lactococcus sp.) OTU 5 (Propionibacterium acnes) OTU 4 (Pseudomonas sp.) OTU 3 (Photobacterium sp.) OTU 2 (Leuconostoc sp.) OTU 1 (Lactobacillus algidus)

Fig. 3. Relative abundance (%) and dynamics of the bacterial community during cold storage of minced meat packaged under modified atmosphere for different production batches (A, data are expressed as the mean of 3 replicates) and preservatives (B, data are expressed as the mean of 2 replicates). Only the 10 most abundant OTUs obtained in this study are specifically indicated. Other OTUs are grouped together in “Other OTUs”. Error bars represent standard errors. These data including the number of sequences and OTUs are also summarized in Supplementary Information Table S1 for different batches of minced meat and for minced meat supplemented with different preservatives.

deterioration and off-odour were observed for all samples after 9 days of storage. Immediately after packaging, O2 and CO2 concentration ranged between 72.2e72.7% and between 15.9e16.9%, respectively (Fig. S1B). During storage, changes in the O2 and CO2 concentrations were similar to the changes observed in the first experiment. On day 9, the headspace composition consisted of 65.7e70.7% O2 and 17.4e21.0% CO2. At day 1, aerobic and anaerobic numbers ranged from 4.4 to 4.8 log cfu/g and 3.6 to 4.8 log cfu/g, respectively (Fig. 1B). Both aerobic and anaerobic microbial counts increased in a similar way during the entire storage period. Microbial counts significantly differed (p < 0.05) among the different preservatives after 9 days of storage. Samples supplemented with NaL and KL showed aerobic and anaerobic numbers above 7 log cfu/g. On the contrary, lower microbial loads were obtained for samples supplemented with SE (6.3 log cfu/g for both microbial groups) and a combination of NaL þ AA (6.6 log cfu/g for both microbial groups). A total of 34,066 filtered sequences were obtained from all minced meat samples (1419 sequences per sample on average). The

number of OTUs during cold storage of minced meat supplemented with different additives are given in Fig. 2B and the dynamics of the bacterial community are given in Fig. 3B. In this case, the bacterial community was not assessed at day 0, as the first experiment revealed no difference between the bacterial communities at day 0 and day 1 (Fig. 3A). At the beginning of the storage period, the highest diversity was observed in minced meat treated with SE (73 OTUs). The composition of the bacterial community was similar between samples supplemented with KL and SE. The most abundant OTU was OTU 4 (Pseudomonas sp.), with a relative abundance of 63% for samples with KL and 60% for samples with SE. In both treatments, OTU 4 was followed by OTU 7 identified as Pseudoxanthomonas sp. (>6%). In contrast, sequencing of samples supplemented with NaL and a combination of NaL þ AA indicated that OTU 1 (L. algidus) was the main OTU at day 1. OTU 2 (Leuconostoc sp.) and OTU 3 (Photobacterium sp.) were also found, with a relative abundance above 12%. It was also noted that OTU 4 (Pseudomonas sp.) occurred with a higher relative abundance in samples supplemented with NaL þ AA (17%) compared to supplementation

J. Stoops et al. / Food Microbiology 48 (2015) 192e199

with NaL alone (5%). After storage (day 9), the bacterial community showed a significant decrease (p < 0.05) in microbial diversity (Fig. 2B), irrespective of the preservative tested. The bacterial community of all samples was dominated by OTU 1 (L. algidus) and OTU 2 (Leuconostoc sp.) (>95%). The most abundant OTU in samples supplemented with NaL was OTU 2 (Leuconostoc sp.) (80%), while OTU 1 (L. algidus) was the main OTU in samples with KL (59%), SE (65%) and a combination of NaL þ AA (80%). 4. Discussion In this study we investigated the microbial quality and community dynamics during cold storage of minced meat packaged under modified atmosphere. Gas analyses, culture-based classical plating and 16S rRNA gene targeted 454 pyrosequencing were performed on minced meat samples of three different manufacturing batches produced over several months (May, June, August) and on meat samples supplemented with four different preservatives. For all samples, O2 concentrations decreased and CO2 concentrations increased during storage. These changes were frequently found in previous studies concerning meat packaged under modified atmosphere (Friedrich et al., 2008; Esmer et al., 2011; Degirmencioglu et al., 2012) and can be due to microbial respiration and/or respiration of the meat muscle (Jakobsen and Bertelsen, 2002). Both aerobic and anaerobic counts increased during storage. At the end of the storage period, most samples reached counts above 7 log cfu/g, which is commonly reported as the microbial spoilage threshold causing sensorial deterioration, like off-odours and slime (Degirmencioglu et al., 2012; Koutsoumanis et al., 2008; Limbo et al., 2010; Rao and Sachindra, 2002). Microbial counts did not exceed the level of 7 log cfu/g during the entire storage period when samples were supplemented with SE (6.3 log cfu/g for both microbial counts) or a combination of NaL þ AA (6.6 log cfu/g for both microbial counts). This finding suggested that these preservatives increased shelf life. Nevertheless, discolouration and off-odours were also perceived for these samples at the end of the storage period (9 days), probably because they were close to the spoilage threshold. 454 pyrosequencing was used to gain more insight into the bacterial community composition and diversity during storage. We were especially interested in the variability of the community composition among different production batches of minced meat and among different preservatives applied in minced meat. It is clear that regardless the manufacturing batch or the preservative used, the microbial diversity decreased dramatically during storage. Lactic Acid Bacteria (LAB), such as L. algidus and Leuconostoc sp., became the most dominant bacteria (>50%) in all samples during storage. LAB are facultative anaerobic bacteria and can grow under aerobic and anaerobic conditions. The finding that aerobic and anaerobic counts were similar and that LAB were the dominant microorganisms, indicated that LAB constituted the majority of the culturable fraction. LAB species are considered to be important for preventing the growth of other bacteria (Ercolini et al., 2006). Several LAB strains produce antibacterial substances such as hydrogen peroxide, lactic acid and bacteriocins (Borch et al., 1996). This property likely contributed to the observed decrease in diversity upon the dominance of these bacteria, outcompeting for example Pseudomonas spp. that were initially present. Competition for the substrate and adaptation to the meat environment might also explain the decreased diversity, as only a fraction of the microbial species initially present will grow and outcompete other species in the fresh meat ecosystem (Huis in't Veld, 1996; Ercolini, 2013; De Filippis et al., 2013). Both aforementioned bacteria are generally recognised as part of the microbiota of spoiled meat packaged under vacuum or modified atmosphere (Doulgeraki et al.,

197

2012; Nieminen et al., 2011; Pennacchia et al., 2011; Vihavainen and € rkroth, 2007). Also in this study, these LAB were probably Bjo responsible for spoilage as their dominance coincided with an increased microbial count close to or above the spoilage level and with sensory deterioration. The initial bacterial community composition was relatively different between the production batches of May (1), June (2) and August (3). An OTU identified as Propionibacterium acnes was found in all samples, but its relative abundance differed among the batches. This OTU is a human skin-associated bacterium and it might have originated from meat handling during manufacturing. Pseudomonas sp. and Psychrobacter urativorans were only found in batch 1 and 2. However, all these OTUs disappeared after cold storage under modified atmosphere and the dominant bacteria appeared to be LAB at day 9. Interestingly, these bacteria were taxonomically identical in all three batches manufactured at different periods, i.e. L. algidus and Leuconostoc sp. This shows that the community dynamics during storage were probably determined by the gas mixture and the composition of the meat matrix. However, also factory procedures and the factory environment might play deterministic roles in the bacterial dynamics, but this aspect was not included in this study. Photobacterium sp. was found as dominant species, in addition to L. algidus, at the end of the storage period for one batch (August), at the expense of Leuconostoc sp. This genus has been identified as the main spoilage bacterium in  et al., 2013), but it is not well several chilled marine fishes (Mace recognised as a predominant microorganism in meat products. Photobacterium spp. were previously found in beef stored at 4  C in air or under vacuum (Pennacchia et al., 2011). The high abundance of this bacterium after storage might be explained by its initial larger abundance in this batch (27%) compared to the other two batches (between 3 and 5%). This finding is in agreement with De Filippis et al. (2013), who reported that the initial contamination of meat can influence the spoilage dynamics during storage. Batch 2 and 3 reached the spoilage threshold more rapidly compared to batch 1, i.e. after 6 days of storage. This is likely due to the larger initial relative abundance of the LAB species compared to batch 1 (only 3% of the sequences). These LAB also contributed to the higher initial microbial counts recorded in these samples. It shows the importance of limiting the initial microbial counts to a minimum. Nevertheless, similar relative abundances of LAB were established after 9 days of storage in this batch compared with the other two batches. Importantly, the microbial spoilage level was achieved more rapidly for batch 2 and 3 than the shelf life set by the company (7 days). In addition, the use of different preservatives resulted in the occurrence of the same bacteria at the end of the storage period among different preservatives. These bacteria were identical to the ones observed in the previous experiment. After one day of storage, a high relative abundance of a Pseudomonas sp. was found in the samples supplemented with KL and SE. LAB species, that are considered to be responsible for spoilage, appeared to be more abundant in samples to which NaL was added (with or without ascorbic acid). This suggests that the other preservatives, KL and SE (with only about 10% of LAB) might extend the shelf life of minced meat. Based on the microbial counts, this could be true for SE, because both microbial counts were below the spoilage level at the end of the storage period. However, samples supplemented with KL showed counts above this level, while microbial count of samples with NaL þ AA were below this level. An obvious relation between the initial relative abundance of LAB and the spoilage level at the end of storage was therefore not observed. After 9 days of storage, L. algidus and Leuconostoc sp. were the most dominant species in all samples, irrespective of the preservative used. Nevertheless, the relative abundance of both OTUs differed among the preservatives.

198

J. Stoops et al. / Food Microbiology 48 (2015) 192e199

L. algidus was predominant for SE and NaL þ AA, and to a lesser extent for KL, whereas Leuconostoc sp. was largely predominant when NaL without ascorbic acid was used. The absence of ascorbic acid changed the predominance of both OTUs towards Leuconostoc sp., as L. algidus was initially predominant. Also in batch 1 and 2 of the previous experiment (using NaL without ascorbic acid) Leuconostoc sp. tend to became more dominant than L. algidus. Previous studies showed that Enterobacteriaceae and Brochothrix thermosphacta can also play an important role in the spoilage of raw meat (Rao and Sachindra, 2002; Doulgeraki et al., 2011; Esmer et al., 2011). In our study, however, no dominant Enterobacteriaceae or B. thermosphacta were observed. Similarly, Ercolini et al. (2011) never detected Enterobacteriaceae in the microbial community of meat packaged under a high oxygen (60%) and carbon dioxide (40%) concentration by culture-independent analyses. In contrast, they reported the presence of B. thermosphacta in beef during the entire storage period of 45 days. In conclusion, the combination of both culture-dependent and culture-independent analyses enabled us to unravel the microbial quality of minced meat during cold storage under modified atmosphere. In this study, microbial changes during storage of fresh minced meat were monitored, which is of utmost importance for the meat processing industry. In accordance with previous studies we found that LAB dominate the community during storage resulting in an increase of the microbial counts and a decrease of the diversity. These species are often associated with meat spoilage and are identical for all samples irrespective of the manufacturing batch and the preservative used. However, the relative abundance of the dominant bacteria can vary. The gas mixture and the meat matrix used in this study are probably the most important factors in the bacterial community dynamics of minced meat. Further research on the importance of the meat processing company, such as processing environment and procedures, on the bacterial community dynamics is desirable as well as on the production of spoilage related-compounds by the meat microbiota. This will provide a more complete evaluation of the microbial quality of minced meat. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.fm.2014.12.012. References Amanatidou, A., 2001. High Oxygen as an Additional Factor in Food Preservation (PhD thesis). Wageningen University, The Netherlands. Altschul, S.F., Gish, G.W., Miller, W., Myers, E.;W., Lipman, D.J., 1990. Basic local alignment search tool. J. Mol. Biol. 215, 403e410. Belcher, J.N., 2006. Industrial packaging developments for the global meat market. Meat Sci. 74, 143e148. Benson, D.A., Karsch-Mizrachi, I., Lipman, D.J., Ostell, J., Wheeler, D.L., 2008. GenBank. Nucleic Acids Res. 36, 25e30. Brightwell, G., Clemens, R., Adam, K., Urlich, S., Boerema, J., 2009. Comparison of culture-dependent and independent techniques for characterization of the microflora of peroxyacetic acid treated, vacuum-packaged beef. Food Microbiol. 26, 283e288. Borch, E., Kant-Muermans, M.-L., Blixt, Y., 1996. Bacterial spoilage of meat and cured meat products. Int. J. Food Microbiol. 33, 103e120. De Filippis, F., La storia, A., Villani, F., Ercolini, D., 2013. Exploring the source of bacterial spoilers in beefsteaks by culture-independent high-throughput sequencing. PLoS ONE 8, e70222. http://dx.doi.org/10.1371/ journal.pone.0070222. Degirmencioglu, N., Esmer, O.K., Irkin, R., Degirmencioglu, A., 2012. Effects of vacuum and modified atmosphere packaging on shelf life extension of minced meat chemical and microbiological changes. J. Anim. Vet. Adv. 11, 898e911. Dhananjayan, R., Han, I.Y., Acton, J.C., Dawson, P.L., 2006. Growth depth effects of bacteria in ground turkey meat patties subjected to high carbon dioxide or high oxygen atmospheres. Poult. Sci. 85, 1821e1828. Dijk, R., van den Berg, D., Beumer, R.R., de Boer, E., Dijkstra, A.F., Kalkman, P., Stegeman, H., Uyttendaele, M., 2007. Microbiologie van voedingsmiddelen:

methoden, principes en criteria. In: Manual of Systematic Bacteriology. Noortvliet Media BV, Houten. Doulgeraki, A.I., Ercolini, D., Villani, F., Nychas, G.E., 2012. Spoilage microbiota associated to the storage of raw meat in different conditions. Int. J. Food Microbiol. 157, 130e141. Doulgeraki, A.I., Paramithiotis, S., Nychas, G.E., 2011. Characterization of the Enterobacteriaceae community that developed during storage of minced beef under aerobic or modified atmosphere packaging conditions. Int. J. Food Microbiol. 145, 77e83. Edgar, R.C., Haas, B.J., Clemente, J.C., Quince, C., Knight, R., 2011. UCHIME improves sensitivity and speed of chimera detection. Bioinformatics 27, 2194e2200. Ercolini, D., 2013. High-throughput sequencing and metagenomics: moving forward in the culture-independent analysis of food microbial ecology. Appl. Environ. Microbiol. 79, 3148e3155. http://dx.doi.org/10.1128/AEM.00256-13. Ercolini, D., Ferrocino, I., Nasi, A., Ndagijimana, M., Vernocchi, P., La storia, A., Laghi, L., Mauriello, G., guerzoni, M.E., Vallani, F., 2011. Monitoring of microbial metabolites and bacterial diversity in beef stored under different packaging conditions. Appl. Environ. Microbiol. 77, 7372e7381. Ercolini, D., Russo, F., Torrieri, E., Masi, P., Villani, F., 2006. Changes in the spoilagerelated microbiota of beef during refrigerated storage under different packaging conditions. Appl. Environ. Microbiol. 72, 4663e4671. Esmer, O.K., Irkin, R., Degirmencioglu, N., Degirmencioglu, A., 2011. The effect of modified atmosphere gas composition on microbial criteria, color and oxidation values of minced beef meat. Meat Sci. 88, 221e226. Fontana, C., Cocconcelli, P.S., Vignolo, G., 2006. Direct molecular approach to monitor bacterial colonization on vacuum-packaged beef. Appl. Environ. Microbiol. 77, 5618e5622.  , I., Dalmadi, I., Horva th, K., Agoston, Friedrich, L., Siro R., Balla, Cs, 2008. Influence of various preservatives on the quality of minced beef under modified atmosphere at chilled storage. Meat Sci. 79, 332e343. Huis in't Veld, J.H.J., 1996. Microbial and biochemical spoilage of foods: an effect on chemical quality changes in the meat: a review. J. Muscle Foods 13, 143e168. Jakobsen, M., Bertelsen, G., 2002. The use of CO2 in packaging of fresh red meats and its effects on chemical quality of changes in the meat: a review. J. Muscle Foods 13, 143e168. , A., Thomma, B., Lievens, B., 2008. Recent advances in molecular techniques to Juste study microbial communities in food-associated matrices and processes. Food Microbiol. 25, 745e761. Kiermeier, A., Tamplin, M., May, D., Holds, G., Williams, M., Dann, A., 2013. Microbial growth, communities and sensory characteristics of vacuum and modified atmosphere packaged lamb shoulders. Food Microbiol. 36, 305e315. Kim, O.S., Cho, Y.J., Lee, K., Yoon, S.H., Kim, M., Na, H., Park, S.C., Jeon, Y.S., Lee, J.H., Yi, H., Won, S., Chun, J., 2012. Introducing EzTaxon: a prokaryotic 16S rRNA gene sequence database with phylotypes that represent uncultured species. Int. J. Syst. Evol. Microbiol. 62, 716e721. Koutsoumanis, K.P., Stamatiou, A.P., Drosinos, E.H., Nychas, G.E., 2008. Control of spoilage microorganisms in minced pork by a self-developed modified atmosphere induced by the respiratory activity of meat microflora. Food Microbiol. 25, 915e921. Kunin, V., Engelbrektson, A., Ochman, H., Hugenholtz, P., 2010. Wrinkles in the rare biosphere: pyrosequencing errors can lead to artificial inflation of diversity estimates. Environ. Microbiol. 12, 118e123. vesque, C.A., Cammue, B.P.A., Lievens, B., Brouwer, M., Vanachter, A.C.R.C., Le Thomma, B.P.H.J., 2003. Design and development of a DNA array for rapid detection and identification of multiple tomato vascular wilt pathogens. FEMS Microbiol. Lett. 223, 113e122. Limbo, S., Torri, L., Sinelli, N., Franzetti, L., Casiraghi, E., 2010. Evaluation and predictive modeling of shelf life of minced beef stored in high-oxygen modified atmosphere packaging at different temperatures. Meat Sci. 84, 129e136. , S., Mamlouk, K., Chipchakova, S., Pre vost, H., Joffraud, J.J., Dalgaard, P., Mace Pilet, M.F., Dousset, X., 2013. Development of a rapid real-time PCR method as a tool to quantify viable Photobacterium phosphoreum bacteria in salmon (salmo salar) steaks. Appl. Environ. Microbiol. 79, 2612e2619. Margulies, M., Egholm, M., Altman, W., Attiya, S., Bader, J., Bemben, L., Berka, J., Braverman, M., Chen, Y.-J., Chen, Z., 2005. Genome sequencing in microfabricated high-density picolitre reactors. Nature 437, 376e380. Martin, M., 2011. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. 17, 10e12. McMillin, K.W., 2008. Where is MAP going? A review and future potential of modified atmosphere packaging for meat. Meat Sci. 80, 43e65. Nieminen, T.T., Koskinen, K., Laine, P., Hultman, J., S€ ade, E., Paulin, L., Paloranta, A., € rkroth, J., Auvinen, P., 2012a. Comparison of microbial comJohansson, P., Bjo munities in marinated and unmarinated broiler meat by metagenomics. Int. J. Food Microbiol. 157, 142e149. €litalo, H., Sa €de, E., Paloranta, A., Koskinen, K., Bjo €rkroth, J., 2012b. Nieminen, T.T., Va The effect of marination on lactic acid bacteria communities in raw broiler fillet strips. Front. Microbiol. 3, 1e8. http://dx.doi.org/10.3389/fmicb.2012.00376. Nieminen, T.T., Vihavainen, E., paloranta, J., Lehto, J., Pulin, L., Auvinen, P., € rkroth, K.J., 2011. Characterization of psychrotrophic bacterial Solismaa, M., Bjo communities in modified atmosphere-packed meat with terminal restriction fragment length polymorphism. Int. J. Food Microbiol. 144, 360e366. Pennacchia, C., Ercolini, D., Villani, F., 2011. Spoilage-related microbiota associated with chilled beef stored in air or vacuum pack. Food Microbiol. 28, 84e93. Rao, D.N., Sachindra, N.M., 2002. Modified atmosphere and vacuum packaging of meat and poultry products. Food Rev. Int. 18, 263e293.

J. Stoops et al. / Food Microbiology 48 (2015) 192e199 Rosenzweig, N., Tiedje, J.M., Quensen, J.F., Mang, Q., Hao, J.J., 2012. Microbial communities associated with potato common scab-suppressive soil determined by pyrosequencing analyses. Plant Dis. 96, 718e725. Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, I.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A., Oakley, B.B., Parks, D.H., Robinson, C.J., Sahl, J.W., Stress, B., Thallinger, G.G., Van Horn, D.J., Weber, C.F., 2009. Introducing mother: opensource, platform-independent, community-supported software for describing and comparing microbial communities. Appl. Environ. Microbiol. 75, 7537e7541. Singh, P., Wani, A.B., Saengerlaub, S., Langowski, H., 2011. Understanding critical factors for the quality and shelf-life of MAP fresh meat: a review. Crit. Rev. Food Sci. Nutr. 51, 146e177. Sivertsvik, M., Jeksrud, W.K., Vagane, A., Rosnes, J.T., 2004. Solubility and adsorption rate of carbon dioxide into non-respiring foods. Part 1: development and validation of experimental apparatus using a manometric method. J. Food Eng. 61, 449e458.

199

Stackebrandt, E., Goebel, B.M., 1994. Taxonomic note: a place for DNA-DNA reassociation and 16S rRNA sequence analysis in the present species definition in bacteriology. Int. J. Syst. Bacteriol. 44, 846e849. €rkroth, K.J., 2007. Spoilage of value-added, high-oxygen Vihavainen, A.J., Bjo modified-atmosphere package raw beef steaks by Leuconostoc gasicomitatum and Leuconostoc gelidum. Int. J. Food Microbiol. 119, 340e345. Waud, M., Busschaert, P., Ruyters, S., Jacquemyn, H., Lievens, B., 2014. Impact of primer choice on characterization of orchid mycorrhizal communities using 454 pyrosequencing. Mol. Ecol. Resour. http://dx.doi.org/10.1111/1755-0998.12229. € bel, U.B., Stackebrandt, E., 1997. Determination of microbial Wintzingerode, F.V., Go diversity in environmental samples: pitfalls of PCR-based rRNA analysis. FEMS Microbiol. Rev. 21, 213e229. Xiao, X., Dong, Y., Zhu, Y., Cui, H., 2013. Bacterial diversity of analysis of zhenjiang yoa meat during refrigerated and vacuum-packed storage by 454 pyrosequencing. Curr. Microbiol. 66, 398e405.