Isolation, identification and utilization of lactic acid bacteria from silage in a warm and humid climate area

The study aimed to isolate and identify lactic acid bacteria (LAB) from silages and their application to improve the fermentation quality of alfalfa. Forty-nine LAB strains were isolated from silages, and two strains were screened for growth and acid production rates. Then two strains were selected for Physiological and morphological tests and 16S rRNA sequencing. They were Gram-positive and Catalase-negative and were able to grow at pH 3.5 and at 45 °C, were unable to grow different NaCl concentrations as 3.0% and 6.5%. Strain BDy3-10 was identified as Lactobacillus rhamnosus, while TSy1-3 was identified as L. buchneri. The selected strains were evaluated on fermentation of alfalfa silage. The highest crude protein content occurred in the BDy3-10 treatment group. The contents of neutral detergent fiber and acid detergent fiber in the TSy1-3 treatment were significantly lower than other treatment (P < 0.05). Compared to the control treatment, inoculation treatments deceased pH during ensiling (P < 0.001) and provided the most increased lactic acid content after ensiling for 10 days (P < 0.001). The acetic acid contents of all the inoculation groups were significantly increased (P < 0.001) during ensiling, and were lower than that of control group (P < 0.001). So, the TSy1-3 treatment most effectively improved the fermentation quality of alfalfa silage in warm and humid climate area.

Silage and ensiling. Alfalfa (Medicago sativa L.) at early budding was harvested from Guizhou University West Campus experimental site (Guiyang, Guizhou, China) on January 1, 2019, and wilted for 48 h. The chemical and microbial compositions after wilting are shown in Table 2. Two selected strains (BDy3-10 and TSy1-3) were used as inoculants for silage preparation. Each strain was dissolved to approximately 10 6 colony-forming units (cfu) g −1 FM, and 100 mL of inoculant was dissolved and sprayed on 3.2 kg of chopped forage (1-2 cm), which was then mixed thoroughly. The chopped forage was then treated with the same amount of (1) distilled water (control), (2) L. rhamnosus (BDy3-10), (3) L. buchneri (TSy1-3), or (4) BDy3-10 + TSy1-3 at a ratio of 1:1. All the treated forages were packed into polyethylene plastic bags (dimensions 16 × 25 cm; Embossed Food saver bag; Taizou Wenbwu Soft-Packing Color-Printing Co. Ltd, Zhejiang, China), and approximately 200 g of wilted forage was packed in each polyethylene bag and then vacuum-sealed, with three replicates for each treatment. The bags were stored at room temperature and opened after 1, 6, 10, 20 and 40 days of storage, their chemical composition, fermentation quality and aerobic stability were analyzed.
Chemical and fermentation analysis. The dry matter (DM) contents of fresh and ensiled forages were determined by drying the sample in a forced-air oven at 65 °C for 48 h. The dried samples were ground to pass a 1 mm screen by a laboratory knife mill (FW100, Taisite Instrument Co., Ltd., Tianjin, China). Crude protein (CP) was analyzed using a Kjeldahl nitrogen analyzer (Kjeltec 2300 Auto-Analyzer, FOSS Analytical AB, Hoganas, Sweden) and crude fat (EE) was determined by an extraction method 21 . Crude ash content (Ash) was detected in an ash furnace by burning at 550 °C for 4 h. Crude fiber (CF), neutral detergent fiber (NDF) and acid detergent fiber (ADF) contents were measured by an A220 Fiber Analyzer (ANKOM Technology Corp., Macedon, NY, USA) 22 . Water soluble carbohydrate (WSC) was determined using the thracenone-sulphuric acid method 23 .
Twenty grams of each silage sample was mixed with 180 mL of distilled water, stored at 4 °C for 18 h, and then filtered. The pH of this filtrate was measured by a glass electrode pH meter (PHS-3C, INESA Scientific Instrument Co., Ltd, Shanghai, China), and ammonia-N was determined by steam distillation of the filtrates. The concentration of organic acids (lactic acid, acetic acid, propionic acid and butyric acid) was measured using high performance liquid chromatography (column, Shodex RSpak KC-811S-DVB gel C; 8.0 mm × 30 cm; Shimadzu, Tokyo, Japan); oven temperature, 50 °C; mobile phase, 3 mmol/L HClO4; flow rate, 1.0 mL/min; injection volume, 5 μL; and a SPD-M10AVP detector 24 .  Table 3, forty-nine strains of LAB were isolated from silages. We initially screened two strains by their high growth and acid-producing rates for 24 h at 37 °C, the OD value of strain TSy1-3 at 24 h was significantly largest that other strains (P < 0.05), at 3.02. The lowest pH value occurred in the BDy3-10 treatment group (P < 0.05), at 3.75. The characteristics and type strains of screened LAB strains are shown in Table 4. Two strains were Grampositive and Catalase-negative. Strain BDy3-10 was a homofermentative LAB, and TSy1-3 was a heterofermentative LAB. The two strains grew normally in the range of 20-45 °C, but grew weakly at 10 °C and 50 °C. They were able to grow at pH values ranging from 3.5 to 7.0, and grew weakly at pH 3.0 tolerating salt (MRS with 3.0% and 6.5% NaCl concentrations, respectively) which limited their growth.

Screening of LAB. As shown in
16S rRNA analysis of screened LAB. The 16S rRNA sequences were compared using the NCBI database (see Table 5), and the results showed that the similarities between all sequences obtained here and several the known 16S rRNA gene sequences in the database were 99.0-100.0%.
The phylogenetic tree of the partial 16S rDNA sequence of BDy3-10 and TSy1-3 are presented in Fig. 2. The sequence of strain BDy3-10 is closely related to that of L. rhamnosus, with 99.7% similarity in their 16S rRNA gene sequences. The 16S rRNA gene sequence of the TSy1-3 strain showed 98.8% similarity to the corresponding sequence, in L. buchneri, confirming its identity.
Effect of screened strains on the chemical compositions and aerobic stability of alfalfa silage. The results of the alfalfa silage chemical composition analysis are shown in Table 6. The DM and WSC contents of the TSy1-3 group treatment were significantly higher than that of the CK group (P > 0.05). The CP contents of all additive treatment groups were significantly higher than that of the CK group (P > 0.05), while the BDy3-10 treatment had the highest CP content. The CF content of the combination of TSy1-3 strain and BDy3-10 strain treatments was lower than the CK group, but the alone inoculation of TSy1-3 strain and BDy3-10 strain were higher than the CK group. The NDF and ADF contents in the TSy1-3 treatment group were www.nature.com/scientificreports/ significantly lower than those in the other groups (P < 0.05), and the EE and Ash contents in all treatments were not significantly different (P > 0.05). After 40 d of ensiling, the aerobic stability of all additive treatment groups was significantly higher than the CK treatment group (P < 0.05), and the TSy1-3 was the highest, at 173 h.
Effects of isolated strains on the fermentation quality of alfalfa during ensiling. The changes in LA, AA and PA that occurred during ensiling are shown in Table 7. The LA content reached a maximum value for T 10 silages, and then decreased dramatically until T 40 in all treatments (P < 0.001). AA and PA contents increased with prolonged fermentation (P < 0.05). The content of LA was generally higher in the BDy3-10 treatment than in the other treatments during ensiling, except at T 40 (P < 0.05), and reached a maximum value (57.03 mmol) for the T 10 silage. The AA content was significantly higher in the CK silage than in the inoculated silages during ensiling (P < 0.05). The PA contents of all treatment groups were significantly lower than that of the CK treatment group (P < 0.05).
The pH values at T 1 ensiling time were approximately 6.0 for all treatments and decreased progressively from T 1 to T 40 . The lowest pH value was recorded for the TSy1-3 treatment at T 20 (4.99). The pH value of the CK treatment at T 40 was 5.58, which was significantly higher than those of all inoculant treatments (P < 0.001). All inoculant treatments were significantly similar in terms of pH values at the T 40 ensiling time, and all values tended to be 5.00.

Discussion
In the present study, biochemical and phylogenetic analyses revealed that all of the characterized LAB belong to the genus Lactobacillus. However, previous studies found that the natural fermentation processes in forage crop and grass silages were dominated by Leuconostoc, Lactococcus, Enterococcus, and Pediococcus, not Lactobacillus species [25][26][27] . A plausible reason may be that the bacterial colonization of fresh crops and plants is controlled by many factors, such as the plant material. Shah et al. found 28 that king grass silage was dominated by P. acidilactici and L. plantarum, while Italian ryegrass (Lolium multiflorum Lam.) silage was dominated by P. acidilactici and L. rhamnosus 12 . However, Ennahar et al. reported 15 that the LAB species of paddy rice silage in Japan, included P.     16 that the LAB species of rice silage in Henan, China include P. pentosaceus, Enterococcus mundtii and L. garvieae. Hence, the LAB species of the same material in different regions were also different. The LAB characteristics and storage temperature exerted strong effects on the fermentation quality 29 . The strain BDy3-10 (L. rhamnosus) had a high acid production rate that could result in the inhibition of aerobic microorganism activities and the reduction of fermentation substrates, while the strain TSy1-3 (L. buchneri) had the fastest growth, which that could shorten fermentation time, decrease nutritional loss and improved silage quality. The temperature of silage may rise to above 40 °C at the beginning of ensiling, which is due to the continuous plant respiration and activity of aerobic microorganisms when air still exists in the plant gap, particularly in the tropics and subtropics 30 . These results were consistent with those of our study, in which the screened strains BDy3-10 and TSy1-3 could grow at 45 °C, but they were both unable to tolerate salt. These results implied that they are very tolerant of acid and high temperature, satisfying the demands for growth in low-pH and high-temperature environments. This is a suitable method for silage preparation in warm and humid areas. Previous studies reported that LAB inoculation could reduce DM loss during ensiling 31,32 . Higher residual WSC contents indicate smaller DM losses during fermentation that result in silage with a higher nutritive value 33 . In our study, lower DM losses and higher residual WSC contents occurred in the TSy1-3 treatment than in the other groups. Because the TSy1-3 strain grew fast in 24 h, it ensured rapid and vigorous LA fermentation and faster reduction of silage pH at earlier stages, which depressed the loss of WSC by fermentation via undesirable bacteria. The low NDF and ADF contents had a positive effect on the silage nutritive value and enhanced digestibility. In our study, the NDF and NDF contents in all inoculant treatments were lower than those in the control. This was similar to the result showing that inoculating LAB results in the highest decline of ADF and NDF, which improves feed intake and digestibility 34 . CP is the critical factor affecting the quality of commercial feed and roughage for ruminants 35 . Higher CP contents were obtained by all inoculants compared to the control. This could be related to the rapid reduction in pH caused by the addition of inoculants, which inhibited the growth and proteolytic activity of microorganisms such as Clostridia 36 . The highest CP content was found in the BDy3-10 treatment. BDy3-10 is a homofermentative LAB that is more efficient in lactic acid production than heterofermentative LAB and can ferment a wide variety of substrates and quickly produce large amounts of lactic acid 37 .
The pH value is considered a very important indicator for the fermentation profile and fermentation quality of ensiled materials 38 . The pH values of the control silages were generally higher than those of the inoculated silages. It is generally desired that the pH value is approximately 3.8-4.2 for any high-quality silage. In this study, the measured pH values reached approximately 5.00 at T 40 with inoculation. This was because alfalfa is a high-protein Table 7. Contents of LA, AA, PA and pH dynamic changes during ensiling. a-c Means having different letter superscripts within column are significantly different (P < 0.05). A-E Means having different letter superscripts within row are significantly different (P < 0.05). LA lactic acid, AA acetic acid, PA propionic acid.