The role of nitric oxide in the mechanism of lactic acid bacteria substituting for nitrite

ABSTRACT This study aims to reveal the role of nitric oxide (NO) in substituting nitrite with lactic acid bacteria. Three lactic acid bacterial strains (L. fermentum JCM1173, L. fermentum IFO3956, Weissella cibaria X31) with nitrite substitution ability were inoculated into fermented sausages, respectively. Results indicated that the ratio of nitrosomyoglobin (NO-Mb) to total pigments was positively correlated with the substitution ability of the strain. UV-visible and resonance Raman spectroscopy suggested the NO-Mb was five-coordinated. In addition, the intracellular NO detected by fluorescent DAF-FM probe showed that W. cibaria X31 had the strongest NO-producing ability. Inhibitory experiments showed that both nitrate reductase and nitric oxide synthase-like protein participated in the NO production. These results indicated the vital role NO played in the substitution through the formation of NO-Mb and demonstrated the sources of NO in lactic acid bacteria preliminarily. GRAPHICAL ABSTRACT


Introduction
Nitrite has been used as a major food additive in meat products for over a century. However, questions have been raised regarding its safety since the by-product N-nitrosamine has teratogenic and carcinogenic effects (Ahn et al., 2004;Wang, Ren, Liu, Zhu, & Wang, 2013). Until now, researches of nitrite substitution committed to produce safer and healthier meat products have focused on the lactic acid bacteria (Alahakoon et al., 2015;Slima et al., 2017). Lactic acid bacteria are essential for meat fermentation (Sriphochanart & Skolpap, 2010) and are generally recognized as safe (GRAS). To date, several lactic acid bacterial strains have been reported to be capable of substituting nitrite, including Lactobacillus fermentum JCM1173, Lactobacillus fermentum IFO3956, Lactobacillus plantarum 8P-A3, Lactobacillus plantarum CMRC6, Lactobacillus sakei CMRC15 and L. plantarum TN8 (Alahakoon et al. 2015;Chen et al., 2016;Slima et al., 2017). The substitution of these Lactobacillus strains is mainly due to their color-enhancing ability. It is well established that nitric oxide (NO) contributed greatly to the characteristic color of meat products in the usage of nitrite through the formation of nitrosomyoglobin (NO-Mb) (Chasco, Lizaso, & Beriain, 1996;Macdougall, Mottram, & Rhodes, 2010;Ras, Leroy, & Talon, 2018).
Nitrosomyoglobin (NO-Mb) is formed by NO binds to the heme iron atom of myoglobin (Pegg & Honikel, 2015). Until now, few experiments have been conducted to directly detect the formation of NO in lactic acid bacteria. Morita, Yoshikawa, Sakata, Nagata, and Tanaka (1997) demonstrated for the first time the production of NO synthesized from L-[guanidino-15 N 2 ]arginine by lactic acid bacteria and thus speculated that L. fermentum IFO 3956 possessed a bacterial NOS. Lamine et al. (2004) measured the NO production by electron paramagnetic resonance (EPR) with large bands masked the signal of [Fe(II)NO(DETC) 2 ]. Moreover, NO produced by Lactobacillus was determined based on the Griess reaction (Gündoğdu, Karahan, & Çakmakç, 2006). Ras et al. (2017) evaluated NO production of S. xylosus based on its ability to form nitrosomyoglobin and nitrosoheme.
Generally, there are two pathways generating NO in bacteria. The first is the nitrate reductase (NR) pathway with nitrate or nitrite as substrate (Wang et al., 2007). The second is the nitric oxide synthase (NOS) pathway with L-arginine as substrate (Ras et al., 2018). NR activity was measured mainly in staphylococci and it contributed to the color formation in cured raw ham (Bosse, Gibis, Schmidt, & Weiss, 2016) and fermented sausages (Gøtterup et al., 2008). Process-driven bacterial community dynamics were key to cured meat color formation by coagulase-negative staphylococci via NR or NOS activities (Mainar & Leroy, 2015). In addition, a nos deletion mutant (Δnos) in S. xylosus indicated NOSmediated NO production in S. xylosus (Ras et al., 2017). However, there are few studies on the contribution of NR and NOS in lactic acid bacteria to meat color at present. Thus, the hypothesis that lactic acid bacteria also contain the NOS enzyme and it plays an important role in NO production which lead to the formation of NO-Mb needs more proof (Li, Luo, Kong, Liu, & Chen, 2016).
Therefore, this study aimed to reveal the role of NO in the substitution of nitrite with lactic acid bacteria and the sources of NO in lactic acid bacteria, which are favor to produce safer meat products that are beneficial to human health.

Bacterial strains and growth conditions
Three lactic acid bacterial strains known to have the substitution ability for nitrite were used in this experiment. L. fermentum JCM1173 (AS1.1880) and S. aureus were obtained from China General Microbiological Culture Collection Center (CGMCC). L. fermentum IFO3956 was purchased from China Center of Industrial Culture Collection (CICC). W. cibaria X31 was isolated from the traditional Chinese dry-cured Xuanwei Ham in our laboratory and was primarily determined to have the ability to substitute for nitrite. All the strains were cultured in DeMan Rogosa and Sharp (MRS) broth at 37°C three times before use. MRS broth was prepared with 10.00 g·L −1 peptone, 10.00 g·L −1 beef extract, 5.00 g·L −1 yeast extract, 5.00 g·L −1 glucose, 5.00 g·L −1 sodium acetate, 2.00 g·L −1 ammonium citrate dibasic, 1.00 g·L −1 Tween 80, 2.00 g·L −1 K 2 HPO 4 , 0.20 g·L −1 MgSO 4 · 7H 2 O, and 0.05 g·L −1 MnSO 4 ·H 2 O, with a final pH of 6.8 ± 0.2. All reagents used were purchased from Sigma-Aldrich, unless otherwise specified.

Preparation of fermented sausages
Pigs were slaughtered at Jilin Huazheng Meat Processing Co. in compliance with animal ethical guidelines of China. The ham lean (without connective tissue) and back fat were chopped into 10 mm-thick squares. The nitrite and nitrate concentrations of raw meat were determined according to the method (Bahadoran et al., 2016), and they were 6.28 ± 0.37 ppm and 13.42 ± 1.05 ppm, respectively. Subsequently, the ground pork was mixed with 1.6% salt, 10.0% sugar, 2.0% rice wine, 0.02% sodium erythorbate, and 0.5% monosodium glutamate (Liu, Wu, & Tan, 2010). Then, cultures of L. fermentum JCM1173, L. fermentum IFO3956, and W. cibaria X31 were individually inoculated into raw meat batters at a concentration of 8 log CFU/g. Moreover, 100 mg/kg sodium nitrite and equal amount of culture medium were used as the control. Mixtures of all samples were stuffed into collagen casings and then the raw sausages were fermented at 25°C for 3 days and dried at 15°C for four weeks.

Determination of the substitution ability
Color is one of the most important measures in nitrite substitution (Slima et al., 2017), which is used to evaluate the substitution ability of different lactic acid bacterial strains. Color differences in the fermented sausages were determined using a Hunterlab colorimeter (ColorFlex®, USA) with an illuminant D65 10º observer. A white standard plate (L* = 94.52, a* = -0.86, b* = 0.68) was used for calibration. The L* value represents lightness. The a* value represents redness. The b* value represents yellowness. All values were analyzed in triplicates of each sample.

Determination of NO-Mb and total pigments
NO-Mb and total pigments were extracted and measured using the standard method (Hornsey, 1956) with slight modifications. The obtained meat samples (25 g) were added with 80% acetone (v/v) to a total volume of 125 mL and then homogenized for 1 min at 1,000 rpm. The pH of the homogenization sample JCM1173, IFO3956, X31, Nitrite, and Control was measured to be 4.60 ± 0.03, 4.51 ± 0.08, 4.83 ± 0.05, 5.22 ± 0.06, 5.79 ± 0.06, respectively. After standing on the ice for 4 min, the mixture was filtered using a quantitative filter paper. The absorption of the resulting supernatants at 540 nm (OD 540 ) was determined by a UV-visible spectrophotometer (TU-1901, Shanghai Tongyong Instruments Co., Ltd., China). The concentration of NO-Mb in meat samples was calculated by OD 540 multiplied by the factor 290. In addition, the supernatants were used for resonance Raman spectral analysis to identify the coordinate states of myoglobin derivatives.
For total pigments calculation, meat samples (25 g) were added with 80% acetone (containing 2% HCl (v/v)) to a total volume of 125 mL, homogenized for 1 min and allowed to stand on the ice for 30 min. The absorption of the supernatants at 640 nm (OD 640 ) was determined, and the concentration of total pigments was calculated by OD 640 multiplied by the factor 680. Finally, the ratio of NO-Mb to total pigments of each sample was evaluated.

Fluorescent detection of endogenous NO
To detect intracellular NO in lactic acid bacteria, a fluorescent assay using DAF-FM diacetate (Invitrogen) as the fluorescent probe was carried out according to Sapp et al. (2014) with slight modifications. The bacterial density of all the strains (three lactic acid bacteria grown in MRS broth, S. aureus grown in LB broth) was calculated based on the standard curve at 600 nm and diluted to 1.0 × 10 7 cfu/mL. Then, 50 mL of each culture was centrifuged at 4,000 × g for 15 min, the cell pellets were suspended in 1 × HBSS buffer containing 5 μM DAF-FM diacetate and immediately incubated at 37°C for 40 min. After washing for three times, the cells were collected by centrifugation and resuspended in HBSS buffer (50 mL). Then, 1 mL aliquots were absorbed into the 24-well plate and determined by fluorescence microscope (AF6000, Leica).

Determination of NOS-like activity
The NOS-like activity was measured using the NOS assay kit (Nanjing Jiancheng Bioengineering Co. Ltd., Nanjing, China). NOS can catalyze the reaction of L-Arginine with molecular oxygen to produce NO, and NO reacts with nucleophilic substances to form colored compounds. According to the absorbance of the colored compounds at 530 nm, the activity of NOS can be calculated. Briefly, cultures of L. fermentum JCM1173, L. fermentum IFO3956, and W. cibaria X31 were grown aerobically at 37°C for 18 h using a three-stage incubation (with an inoculum of 2%) and were centrifuged at 8,000 × g for 20 min at 4°C. The precipitate of cells was collected and resuspended in 20 mM PBS buffer, pH 6.0, containing 1 mM dithiothreitol (DTT), 1 mM EDTA, and 1 mM phenylmethylsulphonyl fluoride (PMSF) in a ratio of 1:4 (w/v). Then, the cells were lysed for 15 min using a Sonifier Cell Disrupter on ice. The soluble fraction was centrifuged at 20,000 × g for 10 min at 4°C. Then, pipette 200 μL supernatant in a test tube, immediately add 200 μL substrate buffer, 10 μL cofactors, and 100 μL chromogenic agents into it. Mixtures were incubated at 37°C for 15 min, and reaction was terminated by adding 2100 μL stop buffer. Optical density of sample was read at 530 nm using a spectrophotometer. The protein concentration of samples was determined by Coomassie (Bradford) protein assay kit (Thermo Scientific). NOS-like activity is expressed as nitric oxide produced in nanomoles per minute per milligram protein. The total NOS-like activity was calculated as follows: where k is nanomole extinction coefficient of color reagent (38.3 × 10 −6 ), t is the reaction time (15 min), l is the path length of quartz cuvette (1 cm), c is the protein concentration of samples (g prot/L), V 0 is the total volume of reaction mixture, and V 1 is the sample volume.
According to the literatures on NOS inhibitors (Guo, Okamoto, & Crawford, 2003;Holden et al., 2013;Kers et al., 2004), the inhibitor Nitro-L-Arginine methyl ester (L-NAME) had the smallest growth inhibition on the target strain. Therefore, L-NAME was chosen to perform the inhibitory assay of NOS. L-NAME was added to the bacterial supernatants with a final concentration of 50, 100, 400, 1000 μM, and then the NOS-like activity of each sample was measured in triplicates.

Preliminary determination of the sources of NO
To determine the sources of NO production in lactic acid bacteria, inhibition assay was performed according to Horchani et al. (2011) andTewari, Prommer, andWatanabe (2013) with minor modifications. One hundred milliliters of each strain culture mentioned above (1.0 × 10 7 cfu/mL) was centrifuged at 4,000 × g for 15 min, and then the cell pellets were resuspended in 10 mM Tris-HCl buffer (pH 7.5) with 10 mM KCl. The suspension was incubated with 1.2 mM sodium tungstate (NR inhibitor) or 100 μM L-NAME (NOS inhibitor) or both for 20 min. The same amount of buffer solution was used as control. Subsequently, 10 μM DAF-FM fluorescent probe was added and incubated for 40 min before determination by a fluorospectrophotometer (F4500, Hitachi) with excitation wavelength at 495 nm and emission wavelength at 515 nm.

Statistical analysis
All data are expressed as mean ± standard deviation. The SPSS Statistics software was used to analyze experimental data with one-way ANOVA, followed by Duncan's multiple range tests. P < .05 was considered significant and P < .01 was considered extremely significant.

Results
3.1. Comparison of the substitution ability of different lactic acid bacterial strains L*-value of samples inoculated with JCM1173, X31 and IFO3956 was significantly different from that of the control (p < 0.05); while the L*-value of meat samples inoculated with nitrite was not significantly different from that of the control (p > .05) ( Table 1). Results indicated that inoculating bacteria strains at a concentration of 8 log CFU/g could enhance the lightness of fermented sausages, which has a better effect than reported (Li, Kong, Chen, Zheng, & Liu, 2013). In addition, the X31 group had a higher a*/b* value than the reference strains JCM1173 and IFO3956, suggesting the X31 strain possessed better substitution ability. This was the first time that a lactic acid bacteria strain with nitrite substitution ability was found in Weissella, an emerging and promising genus (Zhu et al., 2018). Although the a*/b* value of X31 was lower than that of nitrite, X31 was one of the most promising lactic acid bacteria for nitrite substitution.

Identification of NO-Mb by UV-visible and resonance Raman spectroscopy
NO-Mb can be five-coordinated or six-coordinated under different conditions. The acetone extract of five-coordinated NO-Mb has absorption peaks at 395, 480, 542 and 565 nm , and the maximum peak at 396 nm is unique for NO-Mb. The six-coordinated NO-Mb in acetone has absorption peaks at 399 nm, 542 nm, and 576.5 nm (Morita, Sakata, & Nagata, 1998).
formation of five-coordinated NO-Mb can be influenced by many factors. Firstly, pH may have an effect on the coordination state (Morita et al., 1998). Secondly, iron atom in the central position of myoglobin is more likely to combine one nitric oxide molecule to form a five-coordinated complex rather than a six-coordinated one, since the globin moiety protected the heme ring against the second nitric oxide group binding (Soltanizadeh & Kadivar, 2012). Finally, the six-coordinated dinitrosomyoglobin had a weaker interaction of d π with π* than did the five-coordinated mononitrosomyoglobin (Wyllie, Schulz, & Scheidt, 2003), which means the six-coordinated dinitrosomyoglobin is easy to lose one NO molecule to form a more stable five-coordinated NO-Mb. The formation of NO-Mb was further confirmed by the characteristic frequencies of resonance Raman spectrum at about 1371 cm −1 and 1597 cm −1 (Figure 2) according to the literature . However, the second frequency was a little different from the reported. This may be due to the different excitation wavelengths used or the impact of NO in the system. In a word, these results revealed that meat samples either inoculated with lactic acid bacterial strains or nitrite all formed the five-coordinated NO-Mb, which suggested that NO did have an effect on the red color of meat through the five-coordinated NO-Mb.

Analysis of the contribution ratio of NO-Mb to total pigments
The content of NO-Mb increased gradually at the three production stages in all samples except the control, while the content of total pigments only increased at the fermentation stage and decreased at the drying stage ( Table 2). The contribution ratio of NO-Mb to total pigments increased greatly (p < 0.05) during the manufacturing process, and the values in final products of fermented sausages inoculated with L. fermentum JCM1173, W. cibaria X31, L. fermentum IFO3956 and nitrite (100 mg/kg) were 16.82 ± 1.72%, 15.82 ± 1.59%, 14.31 ± 1.17% and 22.37 ± 1.49%, respectively. NO-Mb accounted for 16% of the total red pigments in Staphylococcus xylosus (Ras et al., 2017), which was consistent with our results. Although the R-value of lactic acid bacterial group was not equivalent to that of the nitrite group, results indicated the substantial contribution of NO-Mb to the redness of fermented sausages.

Detection of intracellular NO by DAF-FM diacetate
The intracellular NO production in different strains was detected ( Figure 3). Compared with S. aureus, lactic acid bacteria had a weaker ability to produce NO. S. aureus was a known pathogenic strain with strong NO production capacity (Sorge et al., 2013). Among the three strains of lactic acid bacteria, W. cibaria X31 had the strongest ability to produce NO, while L. fermentum JCM1173 and IFO3956 were weaker. The strong capacity of S. aureus to produce NO could be attributed to NOS, because NO production was completely abrogated in the ΔNOS mutant compared with the wild-type strain (Sorge et al., 2013). However, in lactic acid bacteria, even the weaker production of NO, its source is not clear. NOS does not seem to exist, since no sequence similar to NOS has been found in the gene bank of many species of lactic acid bacteria and the western blot assay is failed in this study (data not shown).
With the increase in L-NAME concentration, the NOS-like activity was gradually inhibited (Figure 4). When the concentration of L-NAME was 100 μM and 400 μM, there was no significant difference between the reductions in NOS-like activity (p > .05). When the inhibitor concentration increased to 1000 μM, the activity of NOS-like decreased significantly (p < 0.05). This may be due to the high concentration of the inhibitor, which has a negative effect on the reaction system. As we all know, the working concentration of many enzymes (such as protease inhibitor PMSF) was only 1000 μM. Therefore, it was concluded that at the concentration of 100 μM L-NAME, the inhibition degree of NOS-like activity could represent the proportion of NO produced by the NOS pathway out of the total NO production. The NOS inhibition rate for W. cibaria X31, L. fermentum JCM1173, and L. fermentum IFO3956 was 53.39%, 39.07%, and 27.78% at 100 μM of inhibitor concentration, respectively. Coincidentally, a 35-40% decrease in NO production was observed in the nos deletion S. xylosus mutant when compared to the wild type (Ras et al., 2017), suggesting that 35-40% of NO was produced by NOS pathway. The resonance Raman spectra were obtained using an Ar + -laser confocal micro-Raman spectrometer (Renishaw, England). Experimental conditions were laser wavelength, 514 nm; Raman shift, 1000-3000 cm −1 ; accumulation time, 30 s; and output power, 30 mW. Control group: Sausages inoculated with an equal amount of MRS culture media; JCM1173, X31, IFO3956 groups: Sausages inoculated with cultures of L. fermentum JCM1173, W. cibaria X31, and L. fermentum IFO3956 at a concentration of 8 log CFU/g, respectively. Nitrite group: Sausages inoculated with nitrite at a concentration of 100 mg/kg. Figura 2. Espectros de resonancia Raman de la nitrosomiglobina extraída (NO-Mb). Estos espectros de resonancia se obtuvieron utilizando un espectrómetro micro-Raman confocal de láser Ar + (Renishaw, Inglaterra). Las condiciones experimentales fueron: longitud de onda del láser, 514 nm; cambio de Raman, 1000-3000 cm-1; tiempo de acumulación, 30 s; y potencia de salida, 30 mW. Grupo de control: salchichas inoculadas con una cantidad igual de medios de cultivo MRS; Grupos JCM1173, X31, IFO3956: salchichas inoculadas con cultivos de L. fermentum JCM1173, W. cibaria X31 y L. fermentum IFO3956 en una concentración de 8 log CFU/g, respectivamente. Grupo de nitritos: salchichas inoculadas con nitrito en una concentración de 100 mg/kg. N represents NO-Mb (mg·L −1 ); T represents total pigments (mg·L −1 ); R represents the ratio of N/T (%). All values are given as means ± standard deviation from triplicate determinations. Different lowercase letters represent the significant difference of data in the same column (P < 0.05). Different uppercase letters represent the significant difference (P < 0.05) of R-value in the same row. N representa NO-Mb (mg·L −1 ); T representa los pigmentos totales (mg·L −1 ); R representa la relación N/T (%). Todos los valores figuran como medias ± desviación estándar de las determinaciones por triplicado. Las diferentes letras minúsculas muestran la existencia de diferencias significativas entre los datos en la misma columna (P < 0.05). Las letras mayúsculas diferentes representan la diferencia significativa (P < 0.05) del valor R en la misma fila.

Preliminary determination of the sources of NO
The total NO production of W. cibaria X31 group was the highest without adding inhibitors, indicating that X31 strain had the strongest NO production capacity ( Figure 5). For L. fermentum JCM1173, IFO3956, and W. cibaria X31, NO catalyzed by NOS-like protein accounted for 21.70%, 25.00%, and 61.60% of total NO production, respectively; while NO catalyzed by NR was 61.32%,   Figura 4. Actividades NOS de las muestras de W. cibaria X31, L. fermentum JCM1173, y L. fermentum IFO3956, tratadas con diferentes concentraciones de L-NAME. Todos los valores se presentan como medias ± desviación estándar de las determinaciones por triplicado. Las diferentes letras minúsculas en el mismo grupo de concentración L-NAME dan cuenta de la existencia de una diferencia significativa (P < 0 .05).
59.78%, and 12.80% of total, respectively (Table S2). Results suggested that inhibitors Na 2 WO 4 and L-NAME were effective for NR and NOS-like protein, respectively. From the perspective of inhibition effect, the NO production of Lactobacillus was affected by not only NOS-like protein but also NR, and the influence of NR was greater than that of NOS-like protein.
However, the NO production of Weissella was mainly affected by NOS-like protein and was less affected by NR. AtNOS1/AtNOA1 was identified as the second enzymes in plants to convert Arginine to NO (Chandok, Ytterberg, van Wijk, & Klessig, 2003), but it turned out to be an ortholog of GTPase-YqeH (Sudhamsu, Lee, Klessig, & Crane, 2008;Zemojtel et al., 2006). YqeH and AtNOA1 probably act as G-proteins that regulate nucleic acid recognition and not as NOS (Sudhamsu et al., 2008). By blasting the currently named NOS sequences of Lactobacillus reuteri (GenBank accession number ABQ83494.1), L. kunkeei (KOY72921.1), L. frumenti (KRL27402.1), it was found that they were highly homologous to GTPase-YqeH, not the NOS. In addition, NOS was found mainly in phylum Firmicutes, and 81% of prokaryotic NOS were found in Gram-positive bacteria (Ras et al., 2018). If mammalian-like NOS does not exist in lactic acid bacteria, why does NOS inhibitor L-NAME inhibit NO production? Interestingly, this phenomenon was also observed in plants (Guo et al., 2003). No gene or protein with sequence similarity to known mammalian-like NOS was found in plants, but the NOS-like activity was inhibited. Although the existence of NOS seems almost impossible in lactic acid bacteria, NR is widely distributed including Lactobacillus ( Figure 6). Since there is no commercially available NR antibody for bacteria, Western blot test of NR has not been done.

Discussion
Weissella is a new genus of lactic acid bacteria, which belonged to Leuconostoc or Lactobacillus earlier. In recent years, Weissella is being extensively explored for its probiotic (Baruah, Maina, Katina, Juvonen, & Goyal, 2017), antioxidant (Zhu et al., 2018) and antimicrobial (Singh, Kim, Wang, Mathiyalagan, & Yang, 2016) activity. This study suggested the feasibility of inoculating lactic acid bacterial strains into meat to substitute for nitrite, especially the W. cibaria X31 strain. Moreover, W. cibaria X31 has been determined to conform to the selection criteria for starter cultures and its application in fermented sausages is under characterization in our laboratory. Although it is still in the experimental stage, the commercial use of W. cibaria X31 in the future is worth looking into.
Unlike the mammalian NOS, little is known about bacterial NOS (Crane, Sudhamsu, & Patel, 2010). The first bacterial NOS was identified in Nocardia species (Chen & Rosazza, 1994). Subsequently, only a few bacteria Bacillus anthracis (Midha et al., 2005), Geobacillus stearothermophilus (Kinloch, Sono, Sudhamsu, Crane, & Dawson, 2010), Staphylococcus aureus (Kinkel et al., 2016) and Staphylococcus xylosus (Ras et al., 2017) were found to possess a mammalian-type NOS. L-NAME is a competitive inhibitor of substrate L-arginine for mammalian NOS. Though lactic acid bacterial NOS seems to have much difference with the mammalian NOS-like protein, it is surprising to find the inhibitor L-NAME is effective for them both. Interestingly, the same phenomenon also occurs in plants (Guo et al., 2003). No protein with sequence similarity to known mammalian-type NOS has been found in plant, but the NOSlike activity can be inhibited by several mammalian NOS Figure 5. Total NO production of three experimental groups L. fermentum JCM1173, L. fermentum IFO3956, and W. cibaria X31. Each group was treated with four treatments: 1) None: no inhibitors were added, and the same amount of saline was used instead; 2) NOS: only nitric oxide synthase inhibitor (L-NAME) added; 3) NR: only nitrate reductase inhibitor (sodium tungstate) added; 4) NR+NOS: both nitrate reductase inhibitor (sodium tungstate) and nitric oxide synthase inhibitor (L-NAME) added. Different lowercase letters in the same group represent significant difference (P < 0.05).
inhibitors including L-NAME. Therefore, the protein with similar activity to mammalian NOS in lactic acid bacteria and plants had more secrets to be discovered.
NO has important physiological functions in animals, plants, and bacteria. Organisms use NO to mediate oxidative stress incurred during the innate immune response (Allain et al., 2011), and NO derived from macrophage is a crucial effector against invading pathogens (Hutfless, Chaudhari, & Thomas, 2018). Bacterial NO is important for toxin biosynthesis, signaling and biofilm formation (Allain et al., 2011). However, the role of NO in lactic acid bacteria and other probiotic bacteria has rarely been studied. In this paper, a new function for bacterial NO was found in the food industry.

Conclusion
Our findings reveal the vital role of NO in the substitution of nitrite with lactic acid bacteria and the sources of NO generation. The production ability of NO corresponds to the substitution ability, depending on the species of lactic acid bacteria. Lactobacillus mainly produces NO from nitrate reductase, while the attribution enzymes produced NO in Weissella are still a mystery. In addition, the hypothesis that a mammalian NOSlike protein involved in the mechanism was proved negative. Our results provide theoretical support for using lactic acid bacteria instead of nitrite to produce safer and healthier meat products in the coming future. Further study is required to standardize the application process of lactic acid bacteria used as starter cultures into various fermented meat products to substitute for nitrite.

Disclosure statement
We declared no conflict of interest.