Effect of Starter Culture and Low Concentrations of Sodium Nitrite on Fatty Acids, Color, and Escherichia coli Behavior during Salami Processing

The reduction of NaNO2 and safety in meat products have been a concern to the meat industry for the last years. This research evaluated the changes in total fatty acids (TFAs) and myoglobin forms by adding starter culture (Lactobacillus sakei/Staphylococcus carnosus) and 50 ppm of NaNO2 during salami processing. In the postripening stage, the starter culture influenced the concentration of the palmitic, oleic, vaccenic, and γ-linolenic TFAs, whereas the metmyoglobin concentration was lower (which could be related to the antioxidant effect of the starter culture). In this stage, an increase in enthalpy, specific heat, and onset temperature was found when adding starter culture and NaNO2, which is directly related to polyunsaturated TFA. However, when adding just the starter culture without 50 ppm NaNO2, the E. coli population was reduced in 4 log CFU/g. This study proposes the analysis of changes in meat product processing like salami in a holistic form, where the application of starter culture with low nitrite concentrations could be in the meat industry an upward trend for reducing this additive.


Introduction
During the processing of salami, physical, physicochemical, biochemical, and microbiological transformations occur. These changes are influenced by factors such as raw material, sodium nitrite, and starter culture. This process has an impact on the total fatty acid profile and it increases the unsaturation level and the susceptibility of oxidation [1,2]. Furthermore, a concentration of 50 ppm sodium nitrite contributes in delaying lipid oxidation in meat products as a result of its reaction with iron [3]. The addition of NaNO 2 influences the formation of the bright red color that corresponds to the nitrosylmyoglobin complex. Moreover, the addition of starter cultures influences changes in the concentration and types of fatty acids and oxidative phenomena of the lipid fraction [4].
Lipid oxidation, including autoxidation and enzymecatalysed oxidation of fatty acids in fermented and drycured meat products, is involved in several aspects of meat products quality (functional, sensory, and nutritional) [5,6]. Primary products of lipid oxidation generate changes in myoglobin such as the oxidation of oxymyoglobin (OMB) to metmyoglobin (MMB) inducing color changes [1,7]; these changes result in a decrease in heme redox stability, rather than the oxidation of specific amino acid residues. This reaction generally proceeds in parallel to the lipidic oxidation, where the products of both reactions can mutually accelerate pigmentation and lipid oxidation [2,8].
In addition, during the processing of salami a w and pH are reduced; as a result, in these products foodborne pathogens are inhibited. However, deviations in the process parameters 2 International Journal of Food Science (temperature and/or humidity) affect the assurance of the elimination of food-borne pathogens in the final products [9,10]. As a consequence, it is of great importance for the safety of the product the addition of starter cultures and the validation of the reduction of E. coli during salami processing.
Therefore, the design of the starter culture is critical for salami quality, where the addition of lactic acid bacteria (LAB) is aimed to ensure product safety and coagulasenegative staphylococci (CNS) is related with formation and stabilization of color and prevention of rancidity due to nitrate reductase and catalase activity [11]. In fact, the LAB Lactobacillus sakei and the CNS Staphylococcus carnosus have been selected as starter culture for their technological properties, where L. sakei decrease the pH through the fermentation of carbohydrates [12]. Furthermore, L. sakei produces NO and N 2 O due to its nitrite reductase and catalase activity heme independent [13]. In addition, S. carnosus can inhibit the oxidation of fatty acids due to its high nitrate reductase activity, which generates nitrites (NO 2 − ) with antioxidant properties, taking part in nitrosomyoglobin formation, being involved in the development of color in fermented meat products [11].
Differential scanning calorimetry (DSC) is a thermo analytical technique that has been applied for lipid characterization. Thermal properties reported good relation with major triacylglycerols (TAG) and total fatty acids (TFA) and minor components free fatty acids (FFA), and oxidation products [14]. Multiple melting behaviors are explained by the melting of TAG and TFA with different melting peaks and crystal reorganization effects; also, enthalpy transitions are coupled with each other and strongly overlap. Consequently, the objective of this study was to evaluate changes of total fatty acids profile by CG-EM and MDSC, and myoglobin forms, during salami processing led by starter culture and the addition of 50 ppm of NaNO 2 . Furthermore, the reduction of E. coli was validated in salami with starter culture addition.

Salami Preparation.
Four treatments were designed for the preparation of salamis (30 samples for each treatment): control (C) without the addition of the starter culture or NaNO 2 ; control (C + ) without the starter culture and 50 ppm of NaNO 2 ; and treatments with the starter culture without NaNO 2 (Ls-Sc) and with the starter culture and 50 ppm NaNO 2 (Ls-Sc + ).
Each treatment was evaluated in 3 different batches. Several salami samples units of 25±1 g each were prepared, the experimental samples were designed from mixtures based on pork meat (42% w/w) and pork back fat (23% w/w) under the formulation and procedures of a local industry (Bogotá, Colombia).
A commercial starter culture composed of Lactobacillus sakei and Staphylococcus carnosus (T-SC-150 Bactoferm6, CHR-Hansen, Hoersholm, Denmark) was added to the Ls-Sc and Ls-Sc + . The starter culture was previously activated in a 2% w/v dextrose (Scharlau Chemie SA, Barcelona, Spain) solution in sterile water and incubated at 25±0.1 ∘ C for 20 h. For every 10 kg of the mixture, 2.5g was added. The meat and fat were ground through a 10 mm diameter and mixed with 2.5% m/m NaCl, 0.1% m/m sodium erythorbate and NaNO 2 according to the treatment and then stuffed into natural casings of 5 cm diameter.
Salami processing was performed at 80% of relative humidity in 4 stages: (i) the prefermentation stage, the first 12 h at 4±0.1 ∘ C; (ii) the postfermentation stage, the following 48 h at 21±0.5 ∘ C; (iii) ripening up to 14 days at 17±0.5 ∘ C. (iv) The last day of processing was considered post-ripening.
Measurements of pH and a w were done according to protocols described by Blanco-Lizarazo et al. [15]. The initial pH for all treatments was 5.89±0.14 and in postripening stage was for Ls-Sc 4.67±0.01, Ls-Sc + 4.85±0.02, and C 4.43±0.01 and for C + 4.52±0.04. The initial a w was 0.910±0.12 and in postripening stage was 0.810±0.03 for all treatments.

Determination of Total Fatty
Acids. Fatty acids were determined on the lipid extract from the salami samples at different processing times. Samples of 10g of salami were placed within an extraction cartridge and inserted into a reflux flask using petroleum ether under the Soxhlet methodology during 5 h. The obtained extracts were concentrated using a vacuum evaporator at 45 ∘ C.
Extracted lipids were methylated and transesterified using sodium methoxide at pH 3, according to the method of Park and Goins [16]. Subsequently, twice the volume of a saturated sodium chloride solution and 3 times the volume of ethyl acetate was added in a separatory funnel. Finally, 5g of anhydrous sodium sulfide was added and decanted. The samples were distilled at 40 ∘ C in vacuum.
The fatty acids methyl esters obtained from the TFAs and triglycerides of fatty acids were separated and quantified with a gas chromatographer (Agilent model 62630A, China) coupled to a mass spectrometer model 5975C with the Agilent triple axis detector (USA). The GC-MS system was equipped with a DB-5 column (Agilent, USA, 30 m x 0.25 mm x 0.25 m). Helium was used as the carrier gas at a flow rate of 1.0 ml/min, and the injection ratio was 1:10. The oven temperature was adjusted to 150 ∘ C with a holding time of 2 min, an initial ramp of 5 ∘ C/min up to 200 ∘ C, followed by a ramp of 3 ∘ C/min up to 215 ∘ C, and finally a ramp of 10 ∘ C/min up to 260 ∘ C, which was held for 2 min. The injector and detector temperatures were adjusted to 250 ∘ C and 290 ∘ C, respectively, and the injection volume was 1 l. The fatty acids methyl esters were identified by comparison of the respective retention times and analysis of their corresponding mass spectra with a mass spectral database of fatty acid methyl esters, ISBN: 978-1-1181-4394-0, MSD ChemStation (E.02.00.493, Agilent, USA) Software, (November 2011). Quantification was performed through areas under the curve of each peak based on total methyl esters. All analyses were performed by triplicate.

Modulated Differential Scanning Calorimetry (MDSC)
Analysis of Salami Lipids. For the analysis of nonisothermal transitions, 5±0.01 g samples of salami of the treatment Ls-Sc + were collected at 0 and 14 days. To ensure the homogeneity and reproducibility of the sample, the 10 samples for each treatment were subjected to cryomilling with liquid nitrogen for 1 min and were milled by impact for five cycles of 10 s each with a grinder (IKA A11 Basic S1, USA).
A Modulated DSC 2920 (TA Instruments, New Castle, DE, USA) was used. A total of 12±1 mg of sample was placed in a hermetically sealed aluminum cell. The experimental conditions for the samples are based on the back fat pork analysis proposed by Sasaki et al. [17], with an initial temperature of 10 ∘ C and a final temperature of 100 ∘ C, a rate (ß) of 5 ∘ C.min-1, and modulation of ±0.796 ∘ C min-1. An empty cell was used as the reference. Indium was used for the temperature calibration, and Sapphire was used for the heat capacity.

Measurement of the Reflectance for Determinate
Concentrations of Myoglobin Forms. Nine 17 mm thick samples were evaluated with 5 replicates for each salami processing time (0, 1, 2, 3, 7, and 14 days). The reflectance was measured between 400 nm and 700 nm in 10 nm intervals in accordance with protocols proposed by the American Meat Science Association [18] with the ColorquestXE spectro-colorimeter (HunterLab, Reston, USA) (light source A, standard geometric observation at 10 ∘ to 1 in of distance, and white standard). The concentrations of the chemical forms of myoglobin were calculated according to the protocols proposed by Tang et al. [19]. The reflectance values (503, 557, and 582 nm) were calculated using linear interpolation.

E. coli Reduction.
For the analysis of E. coli reduction, salami samples for the treatments Ls-Sc and Ls-Sc + were inoculated with 2.5±0.1 ml (4.4±0.19 Log CFU/g) of E. coli ATCC 259226 previously activated with glass beads in 100 mL of BHI medium (Scharlau Chemie SA, Barcelona, Spain) and incubated (Friocell 22 -Comfort, MMM group, Munich, Germany) for 12 h at 37±1 ∘ C. The strain was maintained in glass beads with glycerol/Nutrient Broth (Scharlau Chemie SA, Barcelona, Spain) (20/80%) (v/v) and stored at -80 ∘ C in liquid nitrogen. The salami samples inoculated with E. coli were cultured in Eosin methylene blue agar (EMB) under identical conditions described above for analysis of bacterial growth.

Modelling of Bacterial Behavior.
To analyse the bacterial growth during the processing of salami the model developed by Baranyi and Roberts was applied [20]. This model was selected to determine the maximum growth rate ( max ) and lag phase ( ). The software Matlab R2010a (The MathWorks, USA) was used. For the analysis of E. coli reduction during the processing of salami was applied the model developed by Coroller et al. [21].

Statistical Analysis.
This experiment was based on a completely randomized design with 3 replicates corresponding to 3 batches of salami for fatty acid analysis by GC-MS. For each TFA from each treatment at the evaluated and time points was compared through independent analysis by Duncan's test.
For the analysis of the concentration of myoglobin forms, 3 samples of 3 different batches of salami were analyzed for each treatment and processing time. Percentages of chemical forms of myoglobin were compared for each treatment using the F-test (P<0.05) and Duncan's test. The parameters obtained from the mass spectra and the analyses of the MDSC thermograms were compared for each treatment through the analysis of variance (P<0.05) and Duncan's test.
Microbial growth and E. coli reduction were evaluated for 2 independent samples from 3 different batches of salami for each treatment and processing time. The growth parameters were compared for each treatment through the analysis of variance (P<0.05) and Tukey's test. The statistical software SAS 9.2 (32) was used (SAS Institute Inc., Cary, NC, USA).

Changes in the Total Fatty Acid (TFA)
Profile. The profiles of total fatty acids (TFA) during the processing of salami are expressed in mg/g of fat and are shown in Table 1. The concentration of the saturated fatty acids (SFA) myristic acid (C14:0) and palmitic acid (C16:0) during processing with the addition of the starter culture did not change significantly (P>0.05) as a function of the addition of NaNO 2 . The SFA myristic and stearic were higher abundance in the starter culture treatments; this finding was in concordance with results reported by Aksu and Kaya [22] on Pastirma (Turkish Dry Meat Product) where the SFA concentration was lower in control group than in treatments with starter culture composed by Staphylococcus xylosus + Lactobacillus sakei and Staphylococcus carnosus + Lactobacillus pentosus.
Regarding the concentration of monounsaturated fatty acids (MUFA), palmitoleic acid (C16:1) exhibited higher concentration in the postfermentation and postripening phases for treatment C + . The MUFA oleic acid (C18:1) had the highest concentrations in the control treatments and no differences (P>0.05) were associated with the addition of NaNO 2 . For the Ls-Sc and Ls-Sc + , there were no significant differences (P>0.05) in the concentration of the MUFA vaccenic acid (C18:1 trans-11) after the 7th day of ripening.
The polyunsaturated fatty acids (PUFA) linoleic acid and -linolenic acid were not identified in stages prior to prefermentation. During the ripening stage, Ls-Sc + showed the highest concentration of this PUFA compared to treatment C, which could be correlated with the greater population density of the starter culture. According to Talon et al. [23] S. carnosus inhibited the oxidation of linoleic and linolenic acids in a pork fat but LAB such as L. sakei, L. curvatus, and P. pentosaceus did not present an antioxidant effect on linoleic acid in vitro. Zanardi et al. [24] did not find significant differences on Table 1: Changes in total fatty acids (TFAs) during salami processing expressed in mg/g of fat. Treatments with (Ls-Sc) and without (C) starter addition; the superscript + indicates the addition of 50 ppm NaNO Ls-Sc Palmitoleic acid (C16:1)  International Journal of Food Science 5 Table 2: Onset temperature of the endothermic peak (T o ), maximum temperature (T max ), and enthalpy (ΔH) of the lipid fractions of salami treatment Ls-Sc + in the pre-fermentation and postripening stages.

Stage
Non-reversible heat flow Total heat flow  total fatty acids profile for different fermented sausages with different concentrations for NaNO 2 . However, these authors reported differences on polyunsaturated fatty acids (PUFAs) on the meat product associated with starter culture strain composition, where a starter culture composed by Lactobacillus spp. and S. carnosus presents lower PUFAs concentration than meat products with starter cultures composed by L. curvatus and Micrococcus varians; L. sakei, P. pentosaceus, S. xylosus, and M. varians; and L. sakei, P. pentosaceus, S. xylosus, S. carnosus, and M. varians.
In contrast, the PUFA eicosadienoic acid was not found during the postripening stage; this result is concordant with the results reported by Qiu et al. [25] where decrease the proportion of the long chain PUFAs such as C20:4, C20:2, and C20:1 during cantonese sausage processing. This behavior is due to the oxidation of polyunsaturated fatty acids releasing aldehydes and ketones that affect the flavor of the final product during ripening [26].
During the post-fermentation stage, treatments with and without starter culture (addition of 50 ppm NaNO 2 ) generated the highest concentration of SFA myristic acid and MUFA palmitoleic acid. These results could be related to the high max values of CNS probably dominated by S. carnosus in treatments with the addition of NaNO 2 . The SFA/ (MUFA + PUFA) ratio was higher during the entire process for treatments with the starter culture compared to the controls; however, this behavior was not associated with the addition of NaNO 2 . The TFA ratio increased during the post-fermentation stage and decreased during ripening for the Ls-Sc and Ls-Sc + . The results previously reported by Marco et al. [27] are indicative that the total polyunsaturated fatty acid concentration change as a function of processing time on different lipid fraction (free fatty acids, phospholipids and triglycerides), where the free fatty acids within the total fatty acids present high increase on the concentration and proportion of PUFAs. Martín-Sánchez et al. [28] report PUFAs increase in dry-cured fermented Spanish sausage inoculated with L. sakei and S. carnosus as the starter culture during ripening, but during fermentation, there is a high oxidation rate of unsaturated triglycerides.

Modulated Differential Scanning Calorimetry (MDSC) of Salami Lipids between the Pre-Fermentation and Post-
Ripening Stages for the Ls-Sc + . The lipid thermograms of salami for the Ls-Sc + are shown in Figure 1. Transitions in the total heat flow (Figure 1(a)) show combination on reversing and nonreversing curves (Figure 1(b)), which indicate that the transitions show a combination of melting and crystallization of fatty acids between the prefermentation and postripening stages; and their contribution can be determined by enthalpy [29]. For this reason, a higher enthalpy for transitions in the nonreversible heat flow in the prefermentation stage could be related to crystallization and fatty acids fusion events as the main energetic contribution in this processing stage.
For the nonreversible and total heat flow, an increase was observed (P<0.05) on T o and ΔH in the post-ripening stage compared with prefermentation stage ( Table 2). The Table 3: Transitions of nonreversible and total heat flow, onset temperature of the endothermic peak (T o ), midpoint temperature (T m ), and specific heat (ΔCp) of lipid fractions of salami of treatment Ls-Sc + in the prefermentation and postripening stages.

Stage
Non-reversible heat flow Total heat flow Pre-fermentation increase on enthalpy (ΔH) can be related to the increase in the concentration of primary and secondary products of the lipid oxidation and free fatty acids [14] during the salami processing. Furthermore, in the nonreversible period three transitions were observed for prefermentation stage and two transitions for the postripening stage (Table 2). This behavior indicates a kinetic process on salami lipids, and it can be associated with changes on the enthalpy relaxation reflecting structural reorganization and modifications [29].
Transitions with similar T o had an increment on Cp in the postripening stage respect to the pre-fermentation stage; this was more evident for nonreversible heat flow ( Table 3). The phenomena observed is comparable to the results reported by Samyn et al. [29]; Hidalgo and Zamora [30]; Coupland and McClements [31], where the increment on Cp is in function of the increase of the degree of unsaturation and the lipid autoxidation reactions developed during salami ripening. This is also seen in the results for the treatment Ls-Sc + which are in concordance with the decreasing levels of SFA stearic acid, with significant differences (P<0.05) observed between the prefermentation and post-ripening periods. The concentration of PUFAs (linoleic and -linolenic acid) increased significantly (P<0.05) during the postripening period and the increment on lipid oxidation related to the relative abundance of MMb.

Changes in the Concentrations of Myoglobin Forms.
The concentrations of DMB, OMB and MMB are shown in Table 4. According to Pérez-Alvarez et al. [32], the decrease in the concentration of OMB in the treatments with the addition of NaNO 2 was inverse to the formation of nitrosomyoglobin complexes. For treatments C + , Ls-Sc, and C the decrease was more evident in the OMB related with the processing time. However, in the treatment with NaNO 2 the OMB decrease was perceived from day 7 of processing (ripening), whereas for Ls-Sc and C this decrease was gradual from first day (fermentation). The oxidation of OMB increases proportionally at a lower pH, lower partial oxygen pressure, and lipid oxidation [8].
For all treatments, the concentration of MMB increased as a function of the processing time ( Table 4). The highest concentration of MMB with significant differences (p<0.05) was for C + and C. Consistent with Tang et al. [33], these changes in myoglobin oxidation were directly related to lipid oxidation and its products (i.e., aldehydes) that initiate conformational changes in myoglobin and cause an increase in the oxidation of the haem group and, therefore, it is darkening [18]. Moreover, the lipid oxidation during salami processing could be related with an increase on NaCl concentration; according to Ying et al. [34], who reported in dry-cured goose sausages that this reaction during the ripening stage could be due to the pro-oxidative effect of salt. On the other hand, the increase in NaCl concentration increases the activity of endogenous enzyme lipoxygenase (LOX) involved in lipid oxidation.
The OMB oxidation to MMB generates reactive intermediates capable of enhancing further oxidation of unsaturated fatty acids; in this sense, a superoxide anion is achieved which one dismutase to hydrogen peroxide. The latter can react with MMB concurrently generated in this oxidation sequence to form an activated MMB complex capable of enhancing lipid oxidation that is attributed to ferryl myoglobin [8]. This phenomenon indicates that the starter culture addition could generate an antioxidant effect, associated with the S. carnosus capacity to consume the free oxygen and nitrate reductase, catalase, and SOD activities and for L. sakei the nitrite reductase and catalase activities. However, the 50 ppm of NaNO 2 addition did not accomplish a clear effect on the decrease of MMB formation. Furthermore, during Table 4: Myoglobin species (%), deoxymyoglobin (DMB), oxymyoglobin, (OMB) and metmyoglobin (MMB). Treatments with (Ls-Sc) and without starter culture addition (C); the superscript + indicates the addition of 50 ppm NaNO 2 . the salami processing the treatments with the addition of a starter culture acquired higher concentration of PUFAs than the control treatments; this behavior is also related to an antioxidant capacity from the starter culture, because primary and secondary products derived from unsaturated fatty acids oxidation can enhance myoglobin oxidation [8].

Microbial Dynamics during Salami
Processing. The dynamics of microbial growth during the salami processes are shown in Figure 2. The population density of CNS was maintained for the treatment with the starter culture (Figures 2(a) and 2(b)), whereas for the control treatment a decrease in the microorganism population was observed during the first 6 hours of fermentation of the salami (Figures 2(c) and 2(d)).
In accordance with previous reports by Hospital et al. [35], the addition of low levels of nitrite (<60ppm) can promote microbial growth under low levels of oxygen pressure (inner part of the meat product) also acting as a terminal acceptor of electrons, stimulating reductase activity. Microorganisms from the Staphylococcal genus can rapidly grow under aerobic conditions (surface), and their reductase activity decreases with high oxygen pressure; in this way, excess nitrite may cause the inhibition of growth associated with the interference of the electron transport chain. Furthermore, in the treatments Ls-Sc and Ls-Sc + , oxygen consumption (reduction of redox potential) and nitrate reductase activity is expected, according to Tjener et al. [36]. This phenomenon generates a positive influence on color formation and a negative influence on lipid oxidation.
In all treatments of this study, there were no differences in the bacterial concentrations of BAL and CNS with the addition of 50 ppm NaNO 2 during the process of salami. The influence on changes in lipid fraction could be related to the fatty acid biosynthesis pathway by S. carnosus [37], where branched-chain fatty acids represent the majority of fatty acids produced by S. carnosus. However, for L. sakei were not found enzymes implicated with fatty acid biosynthesis or lipid transformation such as elongases or desaturases.
For the total LAB, there were no differences (P>0.05) in the max between the Ls-Sc, Ls-Sc + , and C + treatments; however, max decreased to 0.003 h −1 following treatment C. In Ls-Sc + , increased to 14.7 h ( Table 5). The population density of LAB in all treatments increased until 72 h of salami processing to approximately of 7 log CFU.g −1 and the control treatments was maintained until the post-ripening stage, while for Ls-Sc and Ls-Sc + from 7th day (ripening) began to slight decrease, this behavior is concordant with results reported by Lorenzo et al. [38] where LAB population declines during ripening associated to the reduction of fermentable carbohydrates and a w . The growth rate ( max ) of aerobic mesophiles was determined as the total bacterial count. The C and C + treatments showed a higher max ; for these treatments, the dominant microbiota was native to the meat and adapted to the environment and therefore could present a higher growth rate. In contrast, no differences were associated with the addition of 50 ppm NaNO 2 for the evaluated treatments or the cell concentrations in the postripening stage (P>0.05).   Figure 3. E. coli was reduced in the treatment Ls-Sc in 4.15±0.75 log CFU/g (90.5%) and in Ls-Sc + in 4.22±0.81 log CFU/g (78.7%). The inactivation parameters showed that ( ) was 15.99 for Ls-Sc and 19.47 for Ls-Sc + implying greater resistance in the first phase of the salami processing for treatment with NaNO 2 . The parameter (p) was 6 for Ls-Sc and 1.79 for Ls-Sc + , thus E. coli treatment resistance related with adaptation speed was reduced in the NaNO 2 addition. Inactivation rate ( ), for Ls-Sc ( 1 ) was 269.55 h −1 and ( 2 ) was 544.31 h −1 , and for Ls-Sc + ( 1 ) was 169.47 h −1 and ( 2 ) was 321.88 h −1 . As a consequence, the NaNO 2 addition reduces decimal inactivation time for both subpopulation of E. coli. Therefore, this behavior indicates the variability of the stress response of E. coli attributed to multiple factors combined as direct and indirect inhibition caused by the starter culture addition, NaNO 2 , pH, and a w . Holck et al. [39] mentioned that E. coli is generally inhibited in fermented sausages due to a combination of several hurdles: high salt, the presence of nitrite curing salt, low a w , decrease of redox potential, the growth of competitive starter culture, and a decrease in pH. Wang et al. [40] found the growth inhibition of E. coli during fermentation and ripening of Chinese fermented sausages by L. sakei and 100 ppm of NaNO 2 addition. Results by Casey and Condon [41] reported that E. coli O157: H45 inhibition rate was faster with 50 ppm NaNO 2 addition than without NaNO 2 during processing of sausage fermented with P. acidilactici.

Conclusions
The addition of starter culture in salami processing increase the concentration of PUFAs (linoleic, -linolenic, and eicosadienoic), in contrast to the MMb concentration, which was lower. This latest concentration could be associated with an antioxidant effect produced by S. carnosus in the starter culture. On the other hand, the addition of 50 ppm NaNO 2 did not have influence in the concentrations of the TFAs and MMb formation. Moreover, the starter culture without and with 50 ppm NaNO 2 addition allowed the reduction of 4 log CFU/g of E. coli. However, the stress response of E. coli showed a broad variability because of bacteria inhibition (attributed to multiple factors combined). Therefore, these results allow analyzing changes during salami processing in a holistic form, where the application of starter culture with low nitrite concentrations could be a strategy for the meat industry to reduce this additive.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no conflicts of interest.