Effects of different starter culture combinations on microbial counts and physico‐chemical properties in dry fermented mutton sausages

Abstract This study was conducted to evaluate the effects of inoculation with different mixed starter culture combinations on microbial counts, physico‐chemical properties, and proteolytic and lipolytic properties of dry fermented mutton sausages during fermentation and ripening. Four different batches of mutton sausages were manufactured: CO batch, no starter cultures, used as control; LB batch with Lactobacillus sakei; LS batch with L. sakei + Staphylococcus xylosus; and LSS batch with L. sakei + S. xylosus + Staphylococcus carnosus. The results showed that adding starter culture caused Lactobacillus and Staphylococcus to become dominant bacteria and reduced the Enterobacteriaceae count in the inoculated sausages. The mixed starter cultures (LS & LSS) accelerated acidification and reduced water activity and lipid oxidation. Statistical analysis revealed that the use of mixed starter cultures, especially the combination L. sakei + S. xylosus + S. carnosus, contributed to proteolysis and lipolysis, increasing total free amino acids and polyunsaturated fatty acids. The above results demonstrate that the use of mixed starter cultures contributes to improving the composition of free amino acids and free fatty acids as well as the hygienic quality of dry fermented mutton sausages.

Commonly used microbial starter cultures include lactic acid bacteria (LAB) and some staphylococcal species, such as Lactobacillus sakei, Staphylococcus xylosus, and Staphylococcus carnosus, as commercial starters for the manufacture of several types of dry fermented sausages (Corbière Morot-Bizot, Leroy, & Talon, 2006). Staphylococci and lactic acid bacteria can help methyl branched amino acids synthesize into 2-and 3-methyl butanal thus improving the flavor of dried-fermented sausage (Schmidt & Berger, 1998;Stahnke & Marie, 1999). In addition, hexanal, octanal, and nonanal caused by β-oxidation of lipids are related not only to substrate concentration but also to microbial starter activity (Olivares, Navarro, & Flores, 2011).
Mutton is characterized by high protein, low fat, and low cholesterol content, and using mutton as raw meat for dry fermented sausages can increase the nutritional value of fermented sausages. To our knowledge, studies on different mixed starter cultures inoculated in dry fermented mutton sausages during fermentation and ripening are rare. The aim of this work was to evaluate the effect of different mixed starter culture combinations on the microbiological and physico-chemical properties and free amino acid and free fatty acid composition of dry fermented mutton sausages.

| Dry fermented sausage preparation and sampling procedures
The sausages were low temperature fermentation mutton sausages manufactured in the meat production center of Inner Mongolia Agricultural University of China, while the Food Microbiology Lab of the center provided the starter cultures of L. sakei, S. xylosus, and S. carnosus. Dry fermentation mutton sausage formulation includes lean mutton meat (80% w/w), tallow (20% w/w), sucrose (0.5% w/w), glucose (0.5% w/w), ascorbic acid (0.05% w/w), sodium nitrate (0.015% w/w), and nitrite (0.01% w/w). Four different batches of mutton sausages were included as follows: (a) CO batch, without commercial starter cultures; (b) LB batch, with L. sakei; (c) LS batch, with L. sakei + S. xylosus; and (d) LSS batch, with mixed starter cultures of L. sakei + S. xylosus + S. carnosus. Mutton meat and fat were made into 4-6 mm particle sizes, which were mixed with starter cultures and other ingredients. The mixture was pickled for 24 hr at 0-4°C and then stuffed into collagen casings with a diameter of 30 mm and length of 15 cm. First, the sausages were fermented for 2 days at 25°C and 90%-95% relative humidity (RH). After fermentation, the sausages entered the dry-ripening process for 10 days at 14-15°C and 75%-85% (RH). Samples were obtained at 0 (end of pickling), 2, 5, and 12 days for subsequent experimental analyses.

| Microbial analysis
Microbiological analyses were performed at the end of the pickling, fermentation, drying, and ripening processes using the methods of Wang et al. (2015) with a slight modification. After aseptically removing and discarding the outer casing, 10 g of each sample was weighed in a sterile plastic bag. Then, samples were homogenized with 90 ml of sterile physiological saline at 20-25°C for 2 min in a stomacher (Seward Medical, London, UK) and were 10-fold diluted in sterile 0.1% peptone water and 0.85% NaCl. Total viable counts were analyzed using plate count agar (PCA) after incubation at 30°C for 48 hr. Lactic acid bacteria and staphylococci were enhanced using MRS agar and MSA agar after incubation at 30°C for 48 hr.
Enterobacteriaceae were determined on VRBG Agar after incubation at 30°C for 48 hr.

| pH, water activity, and thiobarbituric acidreactive substances (TBARS) analysis
The sausage pH and water activity were measured using a digital pH meter (Mettler Toledo, Shanghai, China) and LabMaster-aw (Novasina AG, Switzerland). Ten × 10 mm samples (diameter × height) were compressed at a probe (P/36R) speed of 1 mm/s using a texture analyzer (TA.XT2i, SMS, Germany). Texture analysis was performed by compressing 50% using a P/36R probe, and each sample was compressed twice and the interval time was 5 s.

| Proteolysis index (PI) analysis
Proteolysis index of sausages was determined by the methods of Hughes et al. (2002). Two grams of test samples was diluted in 18 ml distilled water. The solution was homogenized for 2 min and centrifuged at 1,000 g at 4°C for 15 min (AllegraTM, 64R, Beckman, American). The above supernatant was filtered using Whatman #1 filters (Mosutech, Shanghai, China). The extraction process was repeated once, and the merged filtrate was subjected to the tests.
Fifteen milliliters of test fluid described above was mixed with 15 ml 10% trichloroacetic acid and then filtered by Whatman #1 filters.
Five milliliters of the resulting filtrate was measured for nonprotein nitrogen content using Kjeldahl nitrogen determination instruments (Jinghe analytical instruments, Shanghai, China). Protein content of the sausage was determined by the same instrument. The formula for the protein decomposition index is as follows: PI = 0.2 V × N 1 /N; where V: filtrate; N 1 : nonprotein nitrogen content; and N: protein content in the sausage.

| Free amino acid analysis
Free amino acid content was analyzed by high-performance liquid chromatography (HPLC) according to the per-column derivatization method. The samples were dried at temperatures < 63°C to a constant weight, and the fat of the dried samples was removed with ether (Sinopharm Chemical Reagent Co., Shanghai, China). Thirty milligrams of sample was crushed and added to a long neck tube and extracted using 0.1M HCl. Then, vacuum samples were hydrolyzed at 100°C for 24 hr and filtered and fixed to a 50-ml volumetric flask. In addition, 150 μl of the mixture described above was mixed with 50 μl 60 mM adjacent nitrobenzene sulfonic acid chloride and 1 ml 0.05 mM borax buffer solution (adjusted to pH 9.0 with acetic acid), which were derived at 25°C for 10 min. The above were filtered using a 0.45-μm filter for high-performance liquid chromatography analysis. The free amino acid results were expressed in mg/kg of dry matter.

| Free fatty acid analysis
Total lipids were extracted according to the method of Folch, Lees, and Sloane Stanley (1957) with a slight modification using chloroform: methanol (2:1) as the extracted solvent. The extracts were concentrated in a rotating vacuum evaporator. Free fatty acids were determined as described by Olivares et al. (2011) and expressed in mg/100 g of fat.

| Sensory analysis
The sausages were submitted to sensory evaluation to determine the effect of processing method and starter inoculation on the quality of final product. Appearance, color, flavor, texture, and overall quality attributes were evaluated using a hedonic scale with nine points (1 = very bad, 9 = very good). The sensory panel consisted of nine trained panelists. Tests were conducted at 20-22°C in a wellventilated room. Samples were sliced to 4 mm thickness and held in a 5-mm-diameter plastic container with cover. Water and bread were provided for panelists to rinse and clean their mouths between TA B L E 1 Effect of different mixed starter cultures on microbial counts (log CFU/g) of total mesophilic aerobic bacteria (PCA), staphylococci (MSA), lactic acid bacteria (MRS), and Enterobacteriaceae (VRBGA) of dry fermented sausages at various processing stage (means ± SD of six replicates) Mann-Whitney test to determine the effect of processing method and starter inoculation (Gibbons, 1976).

| Statistical analysis
All statistical analyses were performed using IBM SPSS Statistics 19 software (IBM, Chicago, IL, USA). The statistical significance (p < 0.05) was determined using one-way ANOVA. Duncan's test and the LSD method were performed to compare the mean values during processing time.

| Effect of starter culture combinations on microbial counts during fermentation and ripening
The effect of different starter cultures on microbial counts of dry fermented mutton sausages is shown in Table 1. Statistical analysis showed a significant difference (p < 0.01) between the control and the inoculated microbial counts. This result was similar to Essid and Hassouna (2013), who reported that using the selective starter significantly influenced the total viable counts of PCA, staphylococci, LAB, and Enterobacteriaceae. LAB counts in all batches were very close to the total viable number. Initial LAB counts in LB, LS, and LSS were three log units higher than the control (6.74 vs. 6.77 vs. 6.74 vs 3.69 log CFU/g, respectively). The LAB counts in the four groups increased gradually until reaching maximum levels on the 5th day of ripening, and a slight decrease was observed at the end of ripening, which could be due to the decrease in carbohydrate content and decline of water activity in fermented sausages (Lorenzo & Franco, 2012). In addition, LAB were the most competitive microorganisms, which may be due to LAB being well adapted to the meat environment and to a positive interaction among species (Essid & Hassouna, 2013;Lorenzo & Franco, 2012;Zhao et al., 2011). LAB contribute to the development of the physico-chemical qualities of fermented sausages, such as texture, flavor, hygiene, and safety-related properties (Essid & Hassouna, 2013). During the whole process of fermentation and ripening, LAB and staphylococci counts were significantly higher in LB, LS, LSS than in the control (p < 0.001). As reported by Lu  (2010), staphylococci were the second predominant bacteria in mixed starter batches, and the growth rate of the inoculated batches was faster than that of the control. At day 5, staphylococci counts reached maximum levels (8.14, 9.08, and 9.10 log CFU/g for LB, LS, and LSS batches, respectively). However, the staphylococci count in the inoculated batches decreased by approximately 7.42%-8.88% at the end of ripening. This could be due to the poor competitiveness of staphylococci and the decrease in pH and water activity in the sausage, as reported in other works (Essid & Hassouna, 2013;Zhao et al., 2011). During fermentation and ripening, Enterobacteriaceae counts decreased gradually in the four treatment groups, and the inoculated batches had significantly lower Enterobacteriaceae counts than those of the control (4.90, 3.93, 3.73, and 3.94 log CFU/g for CO, LB, LS, and LSS batches) on day 5 (p < 0.01), which may be the result of the lower a w and pH values in the sausage and growth of LAB (Lorenzo, Gomez, & Fonseca, 2014;Lorenzo, Sarriés et al., 2014). At day 12, Enterobacteriaceae counts of LB, LS, and LSS batches (2.36, 2.80, and 2.38 log CFU/g, respectively) presented significantly (p < 0.001) lower than those of the control (3.41 log CFU/g), which agreed with the report of Tabanelli et al. (2012). A lower amount of Enterobacteriaceae improves the product quality and safety of dry fermented sausages by producing less harmful substances, such as biogenic amines (Lu et al., 2010;Ma et al., 2015).

| Effect of starter cultures on pH, water activity, TBARS, and protein decomposition index of dry fermented mutton sausages
The changes in pH, water activity, TBARS, and the protein decomposition index of dry fermented mutton sausages during fermentation and ripening are presented in  (2014) and Zhao et al. (2011). The rapid decline of pH value during fermentation is critical because it helps to inhibit the growth of undesirable microorganisms and improve fermented sausages with redder color (Lorenzo, Gomez, et al., 2014;Lorenzo, Sarriés, et al., 2014). The increase in pH values in the inoculated sausages at 12 days may be caused by the increase in proteolytic peptides and amines, which was induced by bacterial proteases (Ruiz-Moyano et al., 2011). It is essential to ensure the drying process of mutton sausages by stronger acidification to reduce the water-binding capacity of proteins and promote water evaporation (Lorenzo, Gomez, et al., 2014;Lorenzo, Sarriés, et al., 2014).
At the beginning of the production of the fermented sausages, the water activity (a w ) of all four batches was above 0.95. After 2 days, the a w of the inoculated batches decreased significantly (p < 0.001) compared with the control, and all batches decreased to the lowest a w at the end of ripening. This result disagreed with the study of Lorenzo, Gomez, et al. (2014) and Lorenzo, Sarriés, et al. (2014) and was similar to Kaban and Kaya (2009). Lower a w at the end of ripening improves fermented sausage quality and extends shelf life.
According to the report of Ulu (2004), the TBARS value was used to evaluate the lipid oxidation level of fermented sausages. TBARS values ranging from 0.6 to 2.8 mg MDA/kg are considered to be normal in fermented sausages (Marco, Navarro, & Flores, 2006).

| Effect of starter cultures on free amino acids at the end of ripening
Free amino acids, as precursors of many odorants, participate indirectly in flavor development and contribute directly to the taste of sausage products. The analysis of free amino acids (FAA) (expressed as mg/kg of dry sausage samples) of dry fermented mutton sausage is reported in Table 3. The total FAA in LB, LS, and LSS batches was higher than in the control (p < 0.001), which implied that the total FAA could be increased by inoculated starter cultures. However, the total FAA content did not differ significantly among the inoculated batches, which agree with the research of Casaburi et al. (2007) and Candogan, Wardlaw, and Acton (2009). It is well known that the release of free amino acids is attributable to the proteolytic action of microbial enzymes and endogenous enzymes in fermented sausages (Candogan et al., 2009). The types of amino acids play an important role in flavor and taste development (Bermúdez, Franco, Carballo, Sentandreu, & Lorenzo, 2014). As Mau & Tseng (1998) reported, glutamic acid and aspartic acid are the main amine acids that 'impart fresh taste to food, glycine and alanine impart a sweet taste to food, while arginine, leucine, lysine, valine, and phenylalanine cause food to have a bitter taste. Other FAAs show sour or salty characteristics (Mau & Tseng, 1998

| Effect of starter cultures on free fatty acids during fermentation and ripening
The fatty acid compositions in dry fermented mutton sausages are shown in Table 4. The composition and content of fatty acids contribute to the nutritional value of fermented sausage. Appropriate ratio of omega-6/omega-3 fatty acid intake can improve dietary energy homeostasis, thus preventing humans from cardiovascular, stroke, and other diseases (Regulaka-llow et al., 2013). In addition, lipolysis is directly involved in flavor formation during ripening of cured products (Casaburi et al., 2007). In this study, monounsaturated fatty acids (MUFA) were found at the highest levels followed by saturated fatty acids (SFA) and polyunsaturated fatty acids (PUFA) in the production process, which was similar to that reported by Alicia, José, and Mónica (2015). The total FFA in all batches showed a significant TA B L E 3 Effect of different mixed starter cultures on free amino acids (expressed as mg/kg of dry matter) of dry fermented mutton sausage at the end of ripening (means ± SD of six replicates) Significance: n.s.: not significant; *(p < 0.05); **(p < 0.01); ***(p < 0.001).

Days of processing
TA B L E 4 (Continued) (p < 0.05) increasing tendency from day 0 to day 5, and the total content changed from 599.30 to 899.17 mg/100 g of fat and had a small decrease from day 5 to day 12. However, the total FFA amounts in this study were lower than the research of Lorenzo, Gomez, et al. (2014), Lorenzo, Sarriés, et al. (2014) and Lorenzo, Fonseca, Gómez, and Domínguez (2015) in foal sausages and Trani et al. (2010) in pork dry-cured sausages. The different contents of FFA among these studies were mainly due to the various raw materials, formulas, conditions of the processes, and distinct activity and specificity of the lipase of endogenous enzymes and starter cultures (Lorenzo & Franco, 2012). The FFA contents in the four groups showed the same tendency, increasing first and then decreasing during the process.
Alpha-linolenic acid (C18:3 n3) and docosahexaenoic acid (C22:6 n3) can only be acquired through diet and reduce the incidence of diseases, such as diabetes, cancer, and cardiovascular disease (Akpinar, Görgün, & Akpinar, 2009;Jakobsen et al., 2009;Ruiz-Núñez, Janneke Dijck-Brouwer, & Muskiet Frits, 2016). For the free fatty acid composition, our results agreed with those reported by Gómez and Lorenzo (2013) that the MUFA were liberated in higher proportions than SFA and PUFA, indicating that the liberation also originates from the triglycerides rich in MUFA. In addition, Marco et al. (2006) observed that higher specific fatty acids were released from the polar fraction, while the majority of FFA derived from the triglycerides, which could be due to the neutral lipids, was the most abundant lipid fraction in intramuscular and subcutaneous fat. However, some of the fatty acids were significantly (p < 0.05) influenced by the use of the starter cultures. During ripening, the MUFA content of the inoculated batches decreased sharply and was significant (p < 0.001) lower than that of the CO, which may be due to β-oxidation or decomposition into aldehydes and other flavor substances. At the end of ripening, PUFA and total FFA of LSS were significantly higher than those of LB, LS, and CO. In addition, the total FFA of the CO (820.76 mg/100 g of fat) was significantly higher than those of the LB and LS batches (between 764.17 and 777.19 mg/100 g of fat). This suggested that endogenous enzymes played a main role in lipid decomposition, while microbial enzymes could play a promoting role in the composition of free fatty acids.

| Sensory analysis
The sensory attributes evaluated by the trained panel are shown in Table 5. The inoculated sausages had significantly higher scores than the CO in all attributes analyzed. And, mixed starter-inoculated samples were perceived as having better flavor, texture, and overall quality than the LB and control sausages. Table 5 shows the sensory profile of the sausages as affected by the mixed starters of L. sakei + S. xylosus and L. sakei + S. xylosus + S. carnosus.

| CON CLUS IONS
The LAB and staphylococci counts were higher in the inoculated batches and were the most dominant bacteria. promoted proteolysis and improved the content of total free amino acids, total free fatty acids, and polyunsaturated fatty acids. These results suggest that the mixed starter cultures (L. sakei + S. xylosus and L. sakei + S. xylosus + S. carnosus) were conducive to improving the quality and safety of dry fermented mutton sausages.

ACK N OWLED G M ENTS
The authors are grateful to "Double first-class" discipline innovation team construction talent cultivation project ( for their financial supports.

CO N FLI C T S O F I NTE R E S T
The authors declare no conflict of interest or relationship, financial or otherwise.

E TH I C A L A PPROVA L
It is not applicable.