Dry-fermented sausages are a type of fermented meat traditionally produced in many countries all over the world. These products have increasingly recognized as health food and become an important sector of the meat products in the market (Karwowska et al., 2022). In each region or culinary heritage, fermented sausage products are often characterized by a typical taste, aroma, and nutritional value (Carballo, 2021). The typical characteristics of fermented sausages depend much on the formulation of ingredients and raw materials, the technological conditions during fermentation and ripening processes, and the autochthonous microflora derived from raw materials or the environment (Karwowska et al., 2022). Recent studies using culture-dependent and culture-independent methods have shown that lactic acid bacteria (LAB) and coagulase-negative staphylococci (CNS) are the dominant bacteria species in traditional fermented sausages (Ammor and Mayo, 2007; Pisacane et al., 2015; Wang et al., 2018), and the most dominant LAB species are Latilactobacillus sakei, Latilactobacillus curvatus, and Lactiplantibacillus plantarum (Fontana et al., 2005; Aymerich et al., 2006; Urso et al., 2006). LABs have been known to play an essential role during the fermentation of sausages by inhibiting the growth of pathogenic and spoilage bacteria (Hugas and Monfort, 1997), modifying raw materials, and contributing to the development of flavor, color, and texture of products (Babić et al., 2011; Hu et al., 2022).

LABs are generally recognized as safe (GRAS), with a long history of safe use for human beings in traditional fermented foods (Parlindungan et al., 2021). As recognized by many health benefits, several LAB species are being used as probiotic bacteria and commercialized as food supplements on the market (Jensen et al., 2012). The probiotic LAB strains are commonly isolated from traditional fermented foods, accurately identified at the species level, and subsequently screened for probiotic and safety characteristics prior to large-scale production for commercial products (Babić et al., 2011; Landeta et al., 2013). Like other fermented products, fermented sausages are served as a rich reservoir for isolating and selecting probiotic LABs (Rouhi et al., 2013). Therefore, some probiotic candidates are able to apply as starter cultures in fermented foods instead of commercialized food supplements for providing benefits in human diet (Ruiz-Moyano et al., 2008; Agüero et al., 2020).

As produced by traditional techniques, the fermentation of sausages may be affected by the unstable autochthonous microflora present in raw materials and the environment, resulting in undesirable or under-control sensory quality of the sausages (Carballo, 2021). To improve this, starter cultures for sausages have been recently developed and evaluated for the stable quality and safety of the final products. Development of the starter cultures can be approached based on the selection of either commercially available probiotic LAB strains (Agüero et al., 2020) or dominant LAB species (Babić et al., 2011; Landeta et al., 2013), which possess technological properties for the production of sausages. Undoubtedly, each fermented product in a given region has its own composition of fermentative microorganisms that makes its typical sensory characteristic, and the biological properties of LAB species are strain-specific. Therefore, the selection of bacteria in a given type of sausages and the development of their starter cultures that possess both probiotic and technological properties are still great of interest.

Vietnam is a tropical country with a long history of many traditional fermented foods (La Anh, 2015). In the northwestern region, lap xuong is a popular smoked fermented sausage traditionally produced by ethnic minorities by mixing minced pork meat with salt, spices, sugar, alcohol, and plant seeds such as “mac khen” (Zanthoxylum rhetsa) or “doi” (Michelia tokinesis). The homogenized mixture is stuffed into a clean pig's small intestine before hanging out for fermentation and ripening under the sunlight (during the day) and surrounding the wood stove (at night). Up to now, no research has been conducted on the composition and functional characteristics of LABs from the sausage. In this study, we aimed to isolate and identify the autochthonous LAB species from lap xuong. The LAB strains were subsequently investigated on their probiotic, technological and safe properties as a potential stater culture for making a lap xuong fermented sausage.

Testing bacteria and cell line

A reference strain of Lacticaseibacillus casei Shirota (LcS), isolated from a commercial fermented milk product (Yakult) in June 2021, was used in all of the comparative probiotic experiments. Eight enteric or foodborne pathogens were used as indicator bacteria to test the antimicrobial activity of the LAB strains. They included Staphylococcus aureus VTCC 12275, Listeria monocytogenes VTCC 70147, Enterococcus faecalis VTCC 70177, Escherichia coli VTCC 12272, Salmonella enterica VTCC 70080, Aeromonas dhakensis VTCC 70106, Vibrio vulnificus VTCC 70092, and Campylobacter jejuni VTCC 70176, which were obtained from the Vietnam Type Culture Collection (VTCC), Institute of Microbiology and Biotechnology, Vietnam National University, Hanoi, Vietnam. A human colon adenocarcinoma cell line (HT-29) was kindly provided by Dr. Pham Thi Thu Huong at the Key Laboratory of Enzyme and Protein Technology, Hanoi University of Science, Vietnam National University, Hanoi, Vietnam.

Sample collection and isolation of lactic acid bacteria

From January to February 2015, nine traditional fermented sausages of lap xuong were purchased at different local convenience stores in three provinces of Lang Son, Cao Bang, and Son La in northwest Vietnam. A ten-gram of each sausage was homogenized in 90 ml of sterile saline (0.85% NaCl; w/v), and 10-fold serial dilutions were prepared prior to plate out on de Man Rogosa Sharpe (MRS) agar (Becton, Dickinson and Company). After three-day incubation at 30°C under anaerobic conditions using gas generating sachets (GasPak^TM EZ, Becton, Dickinson and Company), presumptive LAB counts were calculated at appropriate dilutions, and the number of bacteria was expressed as log colony-forming unit (CFU) per gram of sausage. Then distinct bacterial colonies with different characteristics of color, size, shape, margin, elevation, and texture were picked up from each sausage sample. These strains were subjected to Gram staining and catalase test. The catalase test was done by adding a few drops of 3% H₂O₂ to the colonies grown in MRS plates and bubbles formation was considered a positive reaction (Agüero et al., 2020). Sixty-three of the Gram-positive and catalase-negative strains were stored at -80^oC in MRS broth supplemented with 20% (v/v) glycerol.

Bacterial identification

From the frozen vials, the LAB strains were streaked onto MRS agar plates. After incubation at 37^oC under anaerobic conditions for 24 h, the bacterial colonies were collected using a sterile cotton swab, suspended in sterile phosphate-buffered saline (PBS, pH 7.2), and washed twice by a centrifugation step of 6,797 × g for 10 min (Centrifuge 5430R, Eppendorf). The obtained fresh bacterial cells were used for genomic DNA extraction following the method described by Gabor et al. (2003). Briefly, the fresh bacterial cells were treated in 200 µl of lysis buffer (100 mM Tris HCl, 100 mM Na₂EDTA, 1.5 M NaCl, 1% [w/v] cetyltrimethylammonium bromide [CTAB], and pH 8.0), 10 µl lysozyme (50 mg/ml), and 10 µl proteinase K (10 mg/ml). The mixture was incubated at 37°C for 30 min. Subsequently, the mixture was added with 50 µl of 10% (v/v) SDS solution, vigorously mixed, and further incubated at 56°C for 2 h. An equal volume of chloroform:isoamyl alcohol solution (49:1, v/v) was added to precipitate protein. After gently mixing, the mixture was centrifuged at 17,949 × g for 10 min (Centrifuge 5430R, Eppendorf). The upper phase was transferred into a new tube. Total bacterial DNA was precipitated with cooled isopropanol at -20^oC for an hour and washed twice with cooled 70% (v/v) ethanol. After centrifugation, the DNA pellet was left for air-drying, resuspended in TE buffer, and stored at -20°C for further use (Gabor et al., 2003).

The 16S rRNA gene amplification was conducted in a 25 µl reaction mixture consisting of 1 × Dream Taq PCR Master Mix (Thermo Scientific), 400 nM of each universal primers 8F and 1492R, and approximately 50 ng of bacterial DNA template (Marroki et al., 2011). The PCR conditions were 95°C for 5 min, followed by 35 cycles of 95°C for 30 sec, 55°C for 30 sec, and 72°C for 1 min 45 sec, and a final extension at 72°C for 5 min (ProFlex^TM Base, Thermo Fisher Scientific). The PCR products were visualized in 1% (w/v) agarose gel using the nucleic acid staining solution of Redsafe (Intron Biotechnology). Successful PCR products were sent to the First Base Laboratories for DNA sequencing analysis. The highest similarity of the 16S rRNA gene sequences was searched on the EzBioCloud database (https://www.ezbiocloud.net) accessed on October 2022, and the nearest related species of the bacterium were identified. Species in Lactiplantibacillus plantarum group were distinguished by the multiplex PCR assay described in previous study (Guantario et al., 2018). The partial 16S rRNA gene sequences of these strains were deposited in the GenBank database under the accession numbers OQ519696 to OQ519705, OQ569371 to OQ581742, OK271413, OK271420, and OQ733250.

Tolerance to low pH

All of the LAB strains were preliminarily screened for low pH tolerance. Briefly, the fresh bacterial cells were suspended to obtain an optical density at 600 nm (OD₆₀₀) of 0.1 in sterile PBS pH 7.2. After adjusting to pH 2.5 using HCl 1N, the cell suspensions were incubated at 37°C for 2 h. Then 10 µl of the cell suspension was dropped on MRS agar plates. After incubating anaerobically at 37°C for 48 h, the strains, which showed stable growth, were selected for testing pH tolerance. In the next experiment, the bacterial cells of the selected LAB strains were counted before and after low pH treatment using MRS agar plates. The number of bacterial cells was expressed as log CFU/ml and survival rates of the LAB strains were calculated as follows:

Tolerance to simulated upper gastrointestinal tract (GIT)

The tolerance to simulated upper human GIT was performed as previously described (Charteris et al., 1998b; Huang and Adams, 2004). Briefly, 200 µl of the fresh LAB cell suspension was mixed with either 1000 µl of simulated gastric juice (3 g/L pepsin [Sigma-Aldrich] in 0.5% [v/v] sterile saline, pH 2.5) and 300 µl of 0.5% sterile saline or 1000 µl of simulated small intestinal juice (1 g/L pancreatin [Sigma-Aldrich] and 3 g/L bile salt [Biobasic] in 0.5% sterile saline, pH 8.0) and 300 µl of 0.5% sterile saline. The treatment steps were performed at 37°C for 2 h and 6 h in the simulated gastric juice and the simulated small intestinal, respectively. Bacterial cells were counted before and after the treatment steps, and survival rates of the LAB strains were calculated as described above.

Adhesion ability to HT-29 cell line

The bacterial adhesion ability to HT-29 cell line was investigated using a method described by Argyri et al. (2013). Briefly, the cell line was sub-cultured at 2 × 10⁵ cells/ml in 24-well tissue culture plates containing Dulbecco’s modified Eagle’s medium (DMEM, 1 g/L glucose) (PAN-Biotech) supplemented with 10% (v/v) fetal bovine serum (Gibco) and 1% (v/v) penicillin/streptomycin (Corning). After incubation at 37°C in 5% CO₂ (SCA-165D, Astec) for 24 h, the culture medium was removed, and the cell layer was washed twice with PBS (pH 7.2) before adding 900 µl of the DMEM without antibiotics into each well. For the adhesion test, a mixture of 100 µl fresh LAB cell suspension and 900 µl of the DMEM was added into each well (approximately 10⁶ bacterial cells/well). After incubation at 37°C in 5% CO₂ for 2 h, the cell layer was washed twice with PBS to remove non-adherent bacterial cells and lysed with 200 µl of 1% (v/v) Triton X-100 solution in PBS for 5 min. Then, the cell lysates were serially diluted, plated onto MRS agar medium, and incubated for 48 h under anaerobic conditions before counting bacterial cells. The ability of bacterial adhesion was calculated as a percentage of the number of adhered bacteria to the number of initially added bacteria (Argyri et al., 2013).

Antimicrobial activity

From the fresh colonies, the LAB strains were cultured into MRS both at 37°C for 24 h under anaerobic conditions. The culture broths were centrifuged at 6,797 × g at 4°C for 10 min and sterilized through 0.2 µm pore size filters (Minisart NY, Sartorius). The cell-free supernatants (CFSs) were neutralized to pH 7.0 before testing the antimicrobial activity using the agar well diffusion method. Briefly, the pathogenic indicator bacteria were streaked on Columbia agar supplemented with 5% sheep blood (MELAB). After overnight incubation at 37°C, individual colonies of each pathogen were suspended in sterile saline to the optical density (OD) of a 0.5 McFarland turbidity standard. The suspensions were spread out on Müller-Hinton (MH) agar plates (Becton, Dickson and Company) using sterile cotton swabs. Then the agar plates were punched with sterile plastic pipes (5 mm in diameter). After removing the agar plugs, the neutralized and non-neutralized CFSs (50 µl) obtained from the LAB cultures were dropped into the wells. The agar plates were incubated at 37°C for 24 h, and the antimicrobial activity of the LAB strains was determined by an inhibition zone size (mm in diameter) around the wells (Choudhary et al., 2019).

Antioxidant activity

The scavenging capacity of free radical DPPH (1,1-diphenyl -2-picrylhydrazyl) was determined as previously described by Cao et al. (2019). Briefly, 1.5 ml of 0.2 mM DPPH (TCI chemical) dissolved in methanol was mixed with 1.5 ml of the CFSs prepared from the LAB cultures. The mixture was shaken vigorously and incubated at room temperature for 30 min in the dark. In a blank sample, the CFS was substituted with the methanolic DPPH. The absorbance was then measured at 517 nm using a UV-Vis spectrophotometer (DU^® 730, Beckman Coulter). The percentage of the radical scavenging activity (RSA) was calculated using the following formula (Cao et al., 2019):

Gas production

The fresh LAB colonies were inoculated into a test tube containing 10 ml of MRS broth supplemented with 1% (w/v) glucose and an inverted Durham bell. After incubation at 37°C for 24 h, the appearance of bubbles inside the Durham bells was considered a positive test (Agüero et al., 2020).

Effects of temperature and simulated fermentation and ripening conditions on LAB growth

The growth of selected LAB strains was performed at 22°C, 30°C, 37°C, and 43°C using MRS broth. After incubation for 48 h, OD at 600 nm of the LAB cultures was determined using spectrophotometer (Ultrospec 10, Biochrom) (Babić et al., 2011).

The combined effects of pH, sodium nitrite, and sodium chloride to mimic fermentation and ripening conditions on the growth of the LAB strains were also evaluated following the method described by Agüero et al. (2020). Briefly, fermentation and ripening media prepared from MRS broth supplemented with 100 mg/L NaNO₂ and 3% (w/v) NaCl were adjusted to pH 4.5 and 5.5, respectively. Then 100 µl of overnight cultures of the LAB strains in MRS broth was added into 10 ml of these media. After incubation at 30°C for 24 h, bacterial cells were counted to determine the bacterial growth under these conditions (Agüero et al., 2020).

Determination of lactic acid concentration

The lactic acid content was determined by a method adapted from Borshchevskaya et al. (2016). Briefly, the fresh LAB colonies were inoculated in MRS broth at 30°C for 48 h. After centrifugation, 50 µl of 20-fold diluted CFSs in deionized water was added into a test tube containing 2 ml of a 0.2% (w/v) iron (III) chloride (Merck) solution prepared in the deionized water. The mixture was vortexed and left at 28°C for 30 min. The absorbance of the mixture was measured at 390 nm using a UV-Vis spectrophotometer. The lactic acid content in the CFSs was extrapolated from a standard curve generated from lactic acid concentrations ranging from 0 to 10 g/L (Borshchevskaya et al., 2016).

Biogenic amine production

Biogenic amine production was determined by a qualitative method described by Agüero et al. (2020). Briefly, 50 µl of overnight cultures of the LAB strains in MRS broth was added into a test tube containing 5 ml medium (5 g/L Difco^TM tryptone [Gibco], 3 g/L yeast extract [Biobasic], 1 g/L D-glucose [Wako], 0.016% (w/v) bromocresol purple [Merck], and pH 5.3) supplemented individually with each amino acid of L-arginine, L-histidine, L-lysine, L-tyrosine, and L-tryptophan (Biobasic) at a final concentration of 0.5% (w/v). Tubes were aerobically incubated at 37°C for 72 h. The amino acid decarboxylase activity of the LAB strains was positive when the color of the medium converted from yellow to purple (Agüero et al., 2020).

Antibiotic susceptibility

The disc diffusion assay was used to determine the antibiotic susceptibility of the selected LAB strains. Briefly, 0.5 McFarland turbidity standard prepared from the fresh LAB colonies were spread on MRS agar plates. Antibiotic discs (Oxoid) of ampicillin (10 µg), ciprofloxacin (5 µg), chloramphenicol (30 µg), clindamycin (2 µg), erythromycin (15 µg), gentamicin (10 µg), streptomycin (10 µg), tetracycline (30 µg), and vancomycin (30 µg) were placed on the prepared MRS agar plates. After incubation at 37°C for 24 h under anaerobic conditions, inhibition zone diameters were measured. The susceptibility of the LAB strains was categorized as resistant (R), moderately susceptible (MS), or susceptible (S) using the cut-off values described in previous study (Charteris et al., 1998a).

Statistical analysis

Data are expressed as means ± standard deviation (SD) obtained from three independent experiments, with each performed in duplicates. Differences between the two numerical data sets were compared using the student’s t-test, and p values less than 0.01 were considered statistically significant.

Isolation and identification of LAB strains

On MRS agar plates, the mean value of the bacterial counts in nine lap xuong samples was 8.62 (range from 4.38 to 10.44) log CFU per gram of the sausages. From each sample, five to ten presumptive LAB strains with distinct colony morphology were picked up. A total of 63 strains that showed Gram-positive bacilli or cocci and catalase-negative were isolated. Based on the 16S rRNA gene sequence analysis, the strains were identified as Enterococcus faecium (n = 3), E. hermanniensis (n = 1), Lactiplantibacillus plantarum group (n = 6), Lactococcus garvieae (n = 2), Lactococcus lactis (n = 6), Lactococcus petauri (n = 1), Latilactobacillus curvatus (n = 2), Latilactobacillus sakei (n = 25), Levilactobacillus brevis (n = 6), Leuconostoc citreum (n = 1), Pediococcus pentosaceus (n = 3), Weissella hellenica (n = 3), and W. viridescens (n = 4).

Of six strains in Lactiplantibacillus plantarum group, the multiplex PCR assay using species-specific primers targeting the recA gene showed a typical DNA band size (318 bp) of Lactiplantibacillus plantarum in five strains and a typical DNA band size (107 bp) of Lactiplantibacillus paraplantarum in the remainder strain. Totally, 14 LAB species belonging to 8 genera were identified in the lap xuong sausages (Table 1). Recent investigations on LAB diversity in traditional fermented products can be approached by DNA-based methods such as denaturing gradient gel electrophoresis technique and high-throughput sequencing technology (Pisacane et al., 2015; Wang et al., 2018). However, our aim focused on selecting viable bacteria for the starter culture of sausage production. Thus, this routine culture technique enabled us to isolate LAB strains for further downstream applications.

Table 1 LAB strains identified based on the 16S rRNA gene sequences

Species	Number of strains (n = 63)	Frequency of strains (%)	Frequency in samples (%)	Identity with reference species in EzBioCloud database (%)	Reference accession number
Enterococcus faecium	3	4.76	11.11	99.79–100	AJ301830
Enterococcus hermanniensis	1	1.59	11.11	99.72	JXKQ01000033
Lactiplantibacillus paraplantarum*	1	1.59	11.11	99.86	AJ306297
Lactiplantibacillus plantarum*	5	7.94	44.44	99.86–100	ACGZ01000098
Lactococcus garvieae	2	3.17	22.22	99.79–100	CP065637
Lactococcus lactis	6	9.52	33.33	99.72–100	BALX01000047
Lactococcus petauri	1	1.59	11.11	100	MUIZ01000023
Latilactobacillus curvatus	2	3.17	22.22	100	BBBQ01000060
Latilactobacillus sakei	25	39.68	77.77	99.93–100	BALW01000030
Levilactobacillus brevis	6	9.52	22.22	98.02–100	KI271266
Leuconostoc citreum	1	1.59	11.11	99.86	AF111948
Pediococcus pentosaceus	3	4.76	33.33	99.86–100	JQBF01000022
Weissella hellenica	3	4.76	22.22	99.93–100	BJOF01000040
Weissella viridescens	4	6.35	22.22	100	AB023236

*Strains were identified by 16S rRNA gene sequences and multiplex PCR

From the nine collected sausages, Latilactobacillus sakei was found in seven (77.7%) samples, followed by Lactiplantibacillus plantarum in four (44.4%) samples, and Lactococcus lactis and P. pentosaceus in three (33.3%) samples. The other ten species were presented in only one or two sausages, indicating the heterogeneous composition of LAB species in the lap xuong samples collected in northwest Vietnam. This heterogeneity was similar to that reported in other countries such as Hungary, Italy, and Greece, where 14 different LAB species were detected in traditional fermented sausages, and only three common species of Latilactobacillus sakei, Lactiplantibacillus plantarum, and Latilactobacillus curvatus were found, with a different picture of LAB population in each sausage (Rantsiou et al., 2005). A review from other Mediterranean countries such as France, Spain, and Portugal also showed a diverse LAB species in traditional fermented meat products (Aquilanti et al., 2016). The heterogeneous LAB species might be explained by various formulations of raw meat materials and spices ingredients mixed for specific sausages and the different conditions at fermentation and ripening stages (Rantsiou et al., 2005). Our study was in good agreement with the previous observations, which showed Latilactobacillus sakei was the most common LAB species in traditional fermented meat products manufactured from different geographical locations (Rantsiou et al., 2005; Bonomo et al., 2008; Landeta et al., 2013; Aquilanti et al., 2016; Wang et al., 2021). Similarly, we also found other common species of Lactiplantibacillus plantarum, Lactococcus lactis, and P. pentosaceus, except for Latilactobacillus curvatus, which was isolated from two out of nine samples (Table 1). W. viridescens grew in two samples in which Latilactobacillussakei was absent. This bacterial species could produce ethanol, acetoin, and diacetyl which contributed to the spoilage of morcilla de Burgos (blood sausage) during cold storage (Diez et al., 2009).

Potential probiotic characteristics of LAB strains

To reach the small intestine and colonize the human host, probiotic candidates have to survive in the acidic condition of the stomach. Thus, 63 LAB strains were initially screened for acid tolerance. Of those, nine (14.3%) strains of E. faecium (n = 1), Lactococcus lactis (n = 3), Latilactobacillus sakei (n = 1), and Lactiplantibacillus plantarum (n = 4) were able to survive after exposure to pH 2.5 for 2 h (Table 2). Bacterial counts were performed, and five strains designated as E. faecium LM0204, Lactococcus lactis LM0302, Latilactobacillus sakei LM0405, Latilactobacillus plantarum LM0705, and Latilactobacillus plantarum LM0901 showed survival rates above 44.67% of the reference LcS strain. These strains were selected for further tests on exposure to stimulated conditions of the stomach and duodenum. Under stimulated gastric conditions consisting of pepsin in low pH, Latilactobacillus sakei LM0405 lost its cell viability, while other strains showed survival rates above 43.01% of the reference LcS strain. Enterococcus faecium LM0204 showed an excellent survival rate at 95.04% in the simulated gastric juice, followed by Lactococcus lactis LM0302 (62.73%), Lactiplantibacillus plantarum LM0901 (49.59%), and Lactiplantibacillus plantarum LM0705 (48.82%). When treated with simulated small intestinal juice consisting of pancreatin and bile salt, only Latilactobacillus sakei LM0405 showed a cell reduction, while other tested LAB strains were unaffected and had equal survival rates to the reference LcS strain (Table 3). Our findings are in agreement with previous studies, which showed that Latilactobacillus sakei strains originating from various fermented meat products were sensitive to gastric conditions and the viability of LAB species in simulated small intestinal juice were greater than in gastric juice (Jensen et al., 2012; Parlindungan et al., 2021).

Table 2 Effect of pH on survival of selected LAB strains

Strains	Acid tolerance (viable counts, log CFU/ml)
	0 h	2 h	Survival rate (%)
Enterococcus faecium LM0204	7.65 ± 0.03	4.46 ± 0.03	58.30 ± 0.23
Lactococcus lactis LM0108	7.54 ± 0.05	3.15 ± 0.15	41.77 ± 2.42
Lactococcus lactis LM0302	8.25 ± 0.01	4.70 ± 0.04	56.96 ± 0.59
Lactococcus lactis LM0404	7.77 ± 0.01	3.27 ± 0.12	42.13 ± 1.60
Latilactobacillus sakei LM0405	7.08 ± 0.01	3.69 ± 0.12	52.11 ± 2.11
Lactiplantibacillus plantarum LM0501	8.13 ± 0.02	2.99 ± 0.11	36.79 ± 1.50
Lactiplantibacillus plantarum LM0701	7.83 ± 0.04	2.85 ± 0.15	36.36 ± 2.16
Lactiplantibacillus plantarum LM0705	7.35 ± 0.04	5.65 ± 0.03	76.94 ± 0.82
Lactiplantibacillus plantarum LM0901	8.04 ± 0.03	4.60 ± 0.04	57.26 ± 0.76
Lacticaseibacillus casei Shirota	8.07 ± 0.01	3.61 ± 0.10	44.67 ± 1.13

Table 3 Tolerance to simulated gastric and small intestinal juices and adherence to intestinal cells of selected LAB strains

Strains	Initial count (log CFU/ml)	Survival to gastric juice			Survival to intestinal juice		Adhesion to HT-29 cells (%)	DPPH radical scavenging activity (%)
		2 h (log CFU/ml)	Survival rate (%)		6 h (log CFU/ml)	Survival rate (%)
Enterococcus faecium LM0204	7.05 ± 0.02	6.70 ± 0.02	95.04 ± 0.02*		8.11 ± 0.01	115.04 ± 0.26	3.10 ± 0.21*	62.23 ± 2.76
Lactococcus lactis LM0302	7.46 ± 0.04	4.68 ± 0.04	62.73 ± 0.28*		8.00 ± 0.02	107.24 ± 0.43*	0.87 ± 0.01*	60.00 ± 0.88
Latilactobacillus sakei LM0405	6.22 ± 0.01	n.d.	n.d.		5.26 ± 0.02	84.57 ± 0.26*	0.48 ± 0.02*	49.47 ± 3.78
Lactiplantibacillus plantarum LM0705	6.76 ± 0.08	3.25 ± 0.08	48.82 ± 0.19		8.08 ± 0.02	119.54 ± 1.58	2.41 ± 0.03*	55.64 ± 4.10
Lactiplantibacillus plantarum LM0901	7.26 ± 0.06	3.60 ± 0.02	49.59 ± 0.19		8.36 ± 0.01	115.16 ± 1.15	n.d.	57.19 ± 4.95
Lacticaseibacillus casei Shirota	7.23 ± 0.04	3.11 ± 0.07	43.01 ± 1.03		8.35 ± 0.03	115.49 ± 0.31	0.06 ± 0.02	55.73 ± 1.68

Values are the mean of triplicate ± SD. The statistically significant difference from L. casei Shirota (LcS) with the p < 0.01 was indicated by one asterisk (*). n.d. not detected.

Long-term colonization in the human gut is a prerequisite for selecting a potential probiotic candidate. This characteristic can be commonly assessed by the presence of marker genes coding for bacterial adhesion proteins or other in vitro assays, such as hydrophobicity of the bacterial cell surface and adherence to human intestinal epithelial cells (Choudhary et al., 2019; Fonseca et al., 2021; Parlindungan et al., 2021). The performance of these in vitro tests may not reflect the true conditions in the gastrointestinal tract, in which multiple adverse factors such as trypsin in pancreatic juice, peristaltic flow, competitive interactions with other fecal bacteria, and host immune responses are present (Haller et al., 2001; Jensen et al., 2012). However, previous comparative studies showed a good correlation between the adhesion capacity to human intestinal epithelial cells and the in vivo persistence of probiotic bacteria (Crociani et al., 1995; Jacobsen et al., 1999). Of five selected strains in the present study, four LAB demonstrated the ability to adhere HT-29 cells, with the adhesion capacity ranging from 0.08% to 3.1% (Table 3). Enterococcus faecium LM0204 had the highest adhesion capacity (3.1%), followed by Lactiplantibacillus plantarum LM0705 (2.41%), Lactococcus lactis LM0302 (0.87%), and Latilactobacillus sakei LM0405 (0.48%), which were greater than that of the reference LcS strain (p < 0.05). Compared to other commercial and potential probiotic LAB strains reported in the previous studies (Jensen et al., 2012; Fonseca et al., 2021), our selected LAB strains had a lower adhesion capacity. We found a similar observation that the adhesion capacity of Latilactobacillus sakei strains was lower than that of Lactiplantibacillus plantarum strains (Jensen et al., 2012). Together with the persistent colonization in the human gut, the adhesion property of probiotic LAB could reduce the adherence of enteropathogens to intestinal epithelial cells (Ruiz-Moyano et al., 2009; Fonseca et al., 2021).

Many LAB species have the scavenging capacities of free radicals such as DPPH, OH, and O^-2 (Chen et al., 2015; Zhang et al., 2017). Overproduction of free radicals from biological oxidation results in oxidative stress, which can damage biomolecules and cause degenerative diseases such as cancer, atherosclerosis, and cirrhosis (Pieniz et al., 2015). Thus, antioxidant activity is a potential health benefit characteristic of probiotic LAB (Ji et al., 2015), and the scavenging activity is commonly determined using a stable synthetic free radical of DPPH (Oh and Jung, 2015; Zhang et al., 2017). Previous studies showed that LAB strains with great antioxidant activity could be applied as starter cultures to enhance the sensory characteristics and the safety of fermented meats (Zhang et al., 2017). In this present study, E. faecium LM0204 had the highest DPPH scavenging activity (62.23%), followed by Lactococcus lactis LM0302 (60%), Lactiplantibacillus plantarum LM0901 (57.19%), Lactiplantibacillus plantarum LM0705 (55.64%), and Latilactobacillus sakei LM0405 (49.47%). Except for Latilactobacillus sakei LM0405, the DPPH scavenging activity of tested LAB strains was higher than the reference LcS strain (Table 3) and relatively equal to that of Latilactobacillus curvatus SR6 (46.48%) and Lacticaseibacillus paracasei SR10-1 (59.67%) in the previous study of Chinese traditional fermented sour meat products (Zhang et al., 2017).

The ability to inhibit the growth of enteric pathogens or spoilage bacteria by producing compounds of organic acids, hydrogen peroxide, diacetyl, or bacteriocins is considered one of the most desirable properties of probiotic strains and functional starter cultures (Babić et al., 2011; Parlindungan et al., 2021). Non-neutralized CFSs of Latilactobacillus sakei LM0405, Lactiplantibacillus plantarum LM0705, and Lactiplantibacillus plantarum LM0901 were able to inhibit L. monocytogenes and V. vulnificus strains. In addition, Lactiplantibacillus plantarum LM0705 and Lactiplantibacillus plantarum LM0901 were able to inhibit A. dhakensis. None of the strains were active against E. faecalis, S. enterica, and S. aureus (Table 4). After neutralizing the CFSs, all of the tested LAB strains did not show antibacterial activity against these pathogens (data not shown). The data indicated that the antimicrobial activity of the LAB strains in the present study is likely due to organic acids rather than antimicrobial peptides or bacteriocin-like compounds (Argyri et al., 2013; Casarotti et al., 2017; Parlindungan et al., 2021). Organic acids produced by LAB are considered natural preservatives and play an essential role in food safety and protection (Parlindungan et al., 2021).

Table 4 Antagonistic activity of cell-free supernatant of the potential LAB strains against foodborne pathogens

Pathogen	Zone of inhibition*
	Enterococcus faecium LM0204	Lactococcus lactis LM0302	Latilactobacillus sakei LM0405	Lactiplantibacillus plantarum LM0705	Lactiplantibacillus plantarum LM0901	Lacticaseibacillus casei Shirota
Aeromonas dhakensis	-	-	-	++	+++	++
Campylobacter jejuni	-	-	-	-	++	-
Enterococcus faecalis	-	-	-	-	-	-
Escherichia coli	-	-	-	-	-	+
Listeria monocytogenes	-	-	+	++	+++	++
Salmonella enterica	-	-	-	-	-	-
Staphylococcus aureus	-	-	-	-	-	-
Vibrio vulnificus	-	-	++	++	+++	+++

*Zone of inhibition included 5 mm agar well; - no inhibition; + inhibition zone ≤ 7 mm; ++ inhibition zone 7–9 mm; +++ inhibition zone ≥ 10 mm.

Technological characteristics of selected LAB strains

Five selected LAB strains were homofermentative because fermentation of glucose did not produce gas (Table 5). This feature is a criterion for selecting starter cultures for fermenting meats because heterofermentative LAB strains may produce a large amount of carbon dioxide, leading to hole formation in the final product (Ammor and Mayo, 2007).

Table 5 Technological characteristics and safety aspects of selected LAB strains

Characteristics	Strains
	Enterococcus faecium LM0204	Lactococcus lactis LM0302	Latilactobacillus sakei LM0405	Lactiplantibacillus plantarum LM0705	Lactiplantibacillus plantarum LM0901
Gas productiona	-	-	-	-	-
Growth at (OD600 nm) after 48 h
22°C	7.87 ± 0.23	6.99 ± 0.54	14.47 ± 1.01	14.19 ± 0.47	20.30 ± 0.08
30°C	6.54 ± 0.23	6.82 ± 0.31	12.05 ± 0.86	11.28 ± 1.17	18.92 ± 1.09
37°C	5.89 ± 0.08	5.45 ± 0.39	10.51 ± 0.86	9.19 ± 0.54	18.10 ± 0.86
43°C	n.d.	n.d.	n.d.	n.d.	n.d.
Lactic acid production (g/L)
30°C	11.23 ± 0.17	11.29 ± 0.08	14.64 ± 2.48	19.48 ± 0.50	17.02 ± 0.54
Medium pH after 48 h	4.53	4.27	4.15	3.84	3.83
Growth at simulated sausage condition (log CFU/ml)
Fermentation	n.d.	n.d.	7.94 ± 0.41	7.95 ± 0.01	9.12 ± 0.03
Ripening	8.45 ± 0.04	8.79 ± 0.03	7.97 ± 0.58	7.88 ± 0.20	9.22 ± 0.15
Biogenic aminesb
L-arginine	-	-	-	-	-
L-histidine	-	-	-	-	-
L-lysine	-	-	-	-	-
L-tyrosine	+	-	-	-	-
L-tryptophan	-	-	-	-	-
Antibioticsc
Ampicillin (10 µg)	-	-	S	S	S
Ciprofloxacin (5 µg)	-	-	R	R	R
Chloramphenicol (30 µg)	-	-	S	S	S
Clindamycin (2 µg)	-	-	S	S	S
Erythromycin (15 µg)	-	-	S	S	S
Gentamicin (10 µg)	-	-	R	S	R
Streptomycin (10 µg)	-	-	R	R	R
Tetracycline (30 µg)	-	-	MS	S	S
Vancomycin (30 µg)	-	-	R	R	R

n.d. not detected.^{a, b} -, negative; +, positive.

^c R, resistant; MS, moderately susceptible; S, susceptible; -, test not conducted.

The growth ability of LAB at different temperatures is an essential physiological characteristic for selecting starter cultures because the production of sausage has to undergo fermentation and ripening at various conditions (Rant, 2005). Five tested LAB strains had growth at 22°C and 30°C and tended to decrease slightly at 37°C (Table 5). These results are in agreement with those obtained by (Babić et al., 2011) and (Todorov et al., 2017), who reported that Lactiplantibacillus plantarum strains grew well in a range from 22°C to 37°C. At 43°C, all tested strains did not grow. In contrast, previous studies showed that Lactiplantibacillus plantarum strains could grow at a higher temperature of 40°C or 45°C (Papamanoli et al., 2003; Todorov et al., 2017).

At 30°C, the LAB strains produced lactic acid with concentrations ranging from 11.23 g/L to 19.48 g/L. The production of lactic acid decreased the pH values of MRS medium to a range from 3.83 to 4.53. The highest lactic acid production was observed in Lactiplantibacillus plantarum LM0705 (19.48 g/L), followed by Lactiplantibacillus plantarum LM0901 (17.02 g/L), and Latilactobacillus sakei LM0405 (14.64 g/L). E. faecium LM0204 and Lactococcus lactis LM0302 produced a similar amount of lactic acid (11.23 and 11.29 g/L, respectively). The results are in accordance with Babić et al. (2011), who reported that LAB strains produced a significant level of lactic acid, ranging from 14.76 to 26.95 g/L (Babić et al., 2011). Lactic acid production of LAB can contribute to the development of texture, color, and taste and improve the safety of the final sausage product (Hugas and Monfort, 1997; Drosinos et al., 2007; dos Santos Cruxen et al., 2019).

The tested strains showed different adaptability to fermentation and ripening conditions (Table 5). Latilactobacillus sakei LM0405, Lactiplantibacillus plantarum LM0705, and Lactiplantibacillus plantarum LM0901 were adapted well to both conditions, while E. faecium LM0204 and Lactococcus lactis LM0302 did not grow at fermentation conditions. This result was similar to that reported by Agüero et al. (2020), who observed a different adaptive response of studied probiotic LAB strains to fermentation and ripening conditions of the sausages (Agüero et al., 2020). Typically, sausage batter is added with 2–3% of salt. During fermentation, the initial pH of the batter (pH 6.0) decreases between 4.6 and 5.1 after 48 h stuffing and subsequently increases to a range from 5.1 to 5.5 by yeast activity (Ammor and Mayo, 2007). Sodium chloride, nitrite, and low pH are considered intrinsic and extrinsic factors of food matrices, which can affect the growth of potential starter culture (Agüero et al., 2020). Therefore, resistance to salt, nitrite, and acidic pH is required for probiotic strains in functional starter culture. Under such conditions in the batter, the selected strains have to proliferate, compete with the natural microbiota to maintain sufficient quantities in the final products, and exert their beneficial properties on the host at the time of consumption (Agüero et al., 2020).

Safety assessment of selected LAB strains

Among the five LAB strains, only E. faecium LM0204 possessed decarboxylase activity on tyrosine (Table 5). Tyrosine is a precursor for tyramine, the most prevalent biogenic amines (BAs) in fermented sausages. Other studies also showed that E. faecium strains isolated from fermented sausages exhibited commonly the ability to produce tyramine (Landeta et al., 2013; Žugić Petrović et al., 2020). In addition, many authors mentioned the formation of BAs by other LAB species, such as Limosilactobacillus fermentum, Limosilactobacillus reuteri, Lacticaseibacillus rhamnosus, and Latilactobacillus sakei (Ruiz-Moyano et al., 2009; Landeta et al., 2013; Agüero et al., 2020). Our Latilactobacillus sakei LM0405 did not produce BAs, indicating that decarboxylase activity was strain-specific (Landeta et al., 2013; Rubio et al., 2014).

BAs are undesirable metabolic products derived from microbial decarboxylation of amino acids (Ammor and Mayo, 2007). BAs like histamine, tyramine, putrescine, cadaverine, and tryptamine are commonly found in foods and beverages. Excessive consumption of these amines could cause health problems, including headache, nausea, hypo- or hypertension, migraine, skin allergy and digestive problems of food poisoning (Papageorgiou et al., 2018). Therefore, no amino acid decarboxylase activity has been proposed as a critical requirement for selecting LAB as an appropriate starter strain to produce healthy and safe sausages (Ammor and Mayo, 2007). Among the tested LAB strains in the present study, E. faecium LM0204 demonstrated the best probiotic properties because of its remarkable ability to survive in gastrointestinal tract conditions and adhere to the intestinal cell line. However, the strain did not grow in stimulated fermentation conditions, generated a low lactic acid amount in the medium, and produced BA of tyramine from tyrosine. The undesirable technological properties suggested that E. faecium LM0204 was not suitable for selecting starter cultures of sausage production. Lactococcus lactis LM0302 did not produce BAs, but the strain could not grow in stimulated fermentation conditions which might cause the overgrowth of other pathogenic and spoilage bacteria during the initial sausage fermentation. Together with low lactic acid production, Lactococcus lactis LM0302 was not suggested for sausage starter cultures.

Three LAB strains of Latilactobacillus sakei LM0405, Lactiplantibacillus plantarum LM0705, and Lactiplantibacillus plantarum LM0901 were selected based on their suitable technological properties. The three LAB strains showed susceptibility to ampicillin, chloramphenicol, clindamycin, erythromycin, and tetracycline but resistance to ciprofloxacin, streptomycin, and vancomycin. Lactiplantibacillus plantarum LM0705 was susceptible to gentamicin, while other strains displayed resistance to this antibiotic. The results are in agreement with previous studies which reported on lactobacilli species (Ammor and Mayo, 2007; Landeta et al., 2013). It is known that the resistance of lactobacilli species to gentamicin, streptomycin, and vancomycin is considered to be intrinsic (natural) (Gueimonde et al., 2013; Parlindungan et al., 2021), while genetic determinants of resistance are commonly found in conjugative plasmids or transposons of the enterococci (Žugić Petrović et al., 2020). Intrinsic resistance exhibits a lower risk for spread than acquired resistance mediated by plasmids or transposons (Ruiz-Moyano et al., 2009). Because traditionally fermented products are usually not heat-treated before consumption, the products might serve as a transfer route of antibiotic-resistant genes by horizontal transfer among commensal or pathogenic bacteria (Landeta et al., 2013) Therefore, the selection of starter cultures with LAB strains sensitive to clinically relevant antibiotics is a particular concern. Among two Lactiplantibacillus plantarum strains tested, LM0705 showed more antibiotic sensitivity than LM0901 (Table 5).

This study demonstrated the heterogeneity of LAB species in the traditional fermented sausage of lap xuong in northwestern Vietnam. Of the collected LAB strains, Latilactobacillussakei was the most common species, followed by Lactiplantibacillus plantarum, Lactococcus lactis, and P. pentosaceus. Bacterial characterization showed that the probiotic, safe, and technological properties of LAB species were strain-specific. Enterococcus faecium LM0204 was the greatest probiotic candidate, but the strain produced tyramine from tyrosine. Additionally, E. faecium LM0204 and Lactococcus lactis LM0302 did not grow under the simulated fermentation conditions of sausage, which could not suggest the usage of the strains for starter cultures. Compared to E. faecium LM0204, Lactiplantibacillus plantarum LM0705 and Latilactobacillus sakei LM0405 exhibited moderate and weak probiotic properties, respectively. However, both strains showed the greatest technological properties, such as growth at a wide range of temperatures and in the simulated fermentation and ripening stages of sausage, production of a high lactic acid amount without gas formation, and no BA production. Together with the susceptibility to antibiotics and the common LAB species in lap xuong, both Lactiplantibacillus plantarum LM0705 and Latilactobacillus sakei LM0405 are preliminarily proposed as promising functional starter cultures for the production of fermented sausage in northwestern Vietnam. Further studies are needed to confirm the sensory quality and safety of the starter cultures in a fermented sausage model.

This research has been done under the research project QG “Conservation and maintaining microbial bioresources 2022” of Vietnam National University, Hanoi.

The authors declare no conflict of interest.