Native grasslands are widely distributed in the north and west of China, including the Mongolian Plateau, Qinghai, and the Tibet Plateau, are an essential resource in animal production, and grow well in fall and autumn, providing sufficient nutrition and biomass for animals (Du et al., 2019). Hay is a traditional method for maintaining forage, while the disadvantages of native grass hay, including hay quality and palatability, make it hard to shake off the seasonal and yearly imbalance of available forage (Yan et al., 2019). Ensiling is a traditional way for preservation of animal feed and green forage crops because it can supply forages for animals year-round, effectively reduce the nutrition loss of forages, and prolong storage time (Pahlow et al., 2003). However, the moisture and water-soluble carbohydrate (WSC) content and the number of lactic acid bacteria (LAB) were lower than the requirement for a well-preserved silage (You et al., 2021). Therefore, it is difficult to produce high-quality silage of native grass using natural fermentation.

Generally, LAB additives are a practical method for improving fermentation quality, which is widely used for silage preparation. The LAB group has been selected from various forages, including alfalfa (Ogunade et al., 2018), Elymus nutans (Xu et al., 2018), oat (Romero et al., 2017), King grass (Shah et al., 2018), Moringa oleifera (Wang Y. et al., 2018), native grass (You et al., 2021), and Teff (Tilahun et al., 2018). Ensiling involves complex microbial interactions, and the microbial ecology associated with silages was conducted using classical microbial techniques (Keshri et al., 2019). Studies on the microbial composition during silage fermentation have shown that Lactobacillus often plays an important role during the later stages of ensiling and manufactures the amount of lactic acid to improve silage fermentation (Ding et al., 2019). Previous studies also found that Lactobacillus, Weissella, Pseudomonas, and Leuconostoc spp. in silage samples were the dominant genera (Guan et al., 2018; Keshri et al., 2019; Yang et al., 2019; Bai et al., 2020). Although the isolation, selection, and application of LAB are critical for silage fermentation and bacterial communities, the LAB was isolated from various forages and grasses with different effects on silage fermentation by various environments (Yang et al., 2010; Zhang et al., 2015; Yan et al., 2019), but few LAB have been isolated and applied on native grass in the Mongolian Plateau. Some LAB strains were isolated from the native grass and naturally fermented silage in the Mongolian Plateau in our research. These strains grew under low-pH conditions in an anaerobic environment, producing more lactic acid, and widely used carbohydrates. Still, the effects of these strains on silage fermentation of native grass remain unclear.

Consequently, this study aimed to select and identify LAB strains using physiological and morphological tests and molecular methods from native grass, and its effects on the bacterial community and the silage quality of native grass were also studied.

Forage crops, silages, and dairy products have been found in various LAB species, and many isolates have been identified as the Lactobacillus group (You et al., 2021). Previously published studies also showed that Lactobacillus is the primary microorganism on forages and silages (Cai et al., 1999c). However, it is challenging to identify the species among differentiated species by morphological, physiological, and biochemical tests (Chen et al., 2013). The 16S rDNA sequence analysis identifies organisms by genus and species (Ennahar et al., 2003).

Similarly, strains XM2, 265, and 842 at pH 4.0 could grow normally and strain XM2 weakly grew at pH 3.0. These results showed that strain XM2 had a high tolerance to acidity and the ability to grow and thrive in low-pH environments, which agree with the findings that L. plantarum displayed high resistance to a low-pH environment (Wang S. et al., 2018). Additionally, these isolated strains grew normally from 5°C to 30°C. While significant differences were observed among these strains, strains XM2 and 842 grew normally at 10°C, and strain 867 grew well at 50°C, which could be attributed to the unique environment, long periods of natural selection, and evolution on the Inner Mongolian Plateau (You et al., 2021). Therefore, the unique traits of these isolated LAB strains could have applications as additives on silage fermentation. What is more, strain XM2 could ferment more substrates than the other strains, which is consistent with the previous report that shows that the L. plantarum group could ferment a wide variety of carbohydrates (McDonald et al., 1991).

The moisture content of the grass is one of the essential factors that can directly influence the silage fermentation quality. In this study, the moisture content of native grass agrees with the previous report that shows the range of 44.93–47.51% in meadow steppe on the Mongolian Plateau (Hou et al., 2017). The plant diversity and environment may contribute to the difference. The WSC content was lower than an adequate WSC concentration that good silage needs WSC content higher than 5% on DM for LAB fermentation (Amer et al., 2012). This result follows the previous study on the low WSC content in native grass (Hou et al., 2017). The CP, ADF, and NDF contents followed those of Du et al. (2019), who discovered the native grass with a lower CP content and higher NDF and ADF contents.

Generally, the silage fermentation process and fermentation quality were determined by the counts and species of epiphytic LAB (Yan et al., 2019). A previously published study showed that the well-preserved silages need the numbers of Lactobacilli to be at least 10⁵ cfu/g FM of ensiling (Cai et al., 1999c). However, in this study, the numbers of LAB and undesirable microorganisms were lower (10⁵ cfu/g) and higher (10⁴–10⁸ cfu/g) in FM than the required values, which could lead to poor fermentation quality. Additionally, LAB groups also included other species that may have few effects on fermentation. Consequently, the use of LAB additives is necessary for native grass silage; not only the low moisture, WSC content, and LAB counts, but also the growth of undesirable microorganisms was inhibited during the early stages of fermentation.

The DM content was lower in the inoculated groups than in the control in this study. Besides, the WSC content was significantly reduced in inoculated silages compared to the control silage, which agrees with Kleinschmit and Kung (2006), who found that grass silages with inoculant had a lower WSC content than that in the control. The WSC contents were fermented by LAB and transformed into organic acids, ethanol, and carbon dioxide by microorganisms during silage fermentation (Muck et al., 2017; Wang S. et al., 2018). The acid hydrolysis during ensiling also reduced the DM content (Zhao et al., 2018). Therefore, the DM, WSC, NDF, ADF, and pH decreased in inoculated silages. The CP and NH₃-N contents in control were lower and greater than those in the inoculated silages, reflecting the growth of undesirable microorganisms and the accumulation of NH₃-N during ensiling (Dong et al., 2020). The higher CP and lower NH₃-N contents in inoculated silages may be attributed to the lower pH, which inhibits the growth and activities of undesirable microorganisms (Arriola et al., 2011; Heinritz et al., 2012).

The pH value in inoculated silages was lower than that in the control, while the lactic acid in the control was lower than that in L, XM2, and 265 groups. These results may be attributed to a high amount of LA with reducing pH values in the anaerobic environment (Fiyla, 2003). A previous study indicated that a pH of less than 4.20 could inhibit the growth of harmful bacteria and ensure fermentation quality (Wang et al., 2019). The pH value of the control was significantly higher than that of inoculated silages, and the pH value was 4.42. The lower number of LAB and higher numbers of yeasts and aerobic bacteria may be the main reason because the LA and other nutrients used by yeasts, molds, and other aerobic microorganisms produce large quantities of metabolism, resulting in pH increase (Woolford, 1990). In this study, the coliform bacteria and mold were undetectable after 30 days of fermentation, which is in agreement with the report that coliform bacteria and mold were not detected after ensiling for 30 days (Dong et al., 2020). The growth of coliform bacteria and mold inhibited by lower pH was the main reason (Lima et al., 2010; Reich and Kung, 2010).

Amplicon sequencing of bacteria in fresh native grass and silage samples was performed. Good’s coverage values in fresh matter and silages were higher than 99%, suggesting that sequencing could adequately reflect the dynamic change in the bacterial community (Yang et al., 2019). Compared to FM, the diversity and richness were reduced in inoculated silages, especially in inoculated silages, which follow Ogunade et al. (2018), who found that LAB additives could reduce the bacterial diversity due to the increase in the abundance of the predominant genus. The antibacterial activity that affects the composition of different bacteria and the lower pH inhibiting the growth of the desirable bacteria in this study may be the main reason (Bai et al., 2020). Compared to the control, the diversity and species richness of the microbial communities in the LAB-treated silages decreased, which agrees with previous studies that show that the addition of LAB could reduce diversity and richness (Dong et al., 2019).

The PCoA plot showed a clear separation of bacteria in FM and silages, which indicated that ensiling reconstructs the microbial community. These results followed the work of Zeng et al. (2020) that showed that the bacterial communities were distinguished in FM and silages. Besides, compared to the control, the PCoA of LAB-treated silages was also separate, which showed that the additives significantly influenced the microbial community.

This study showed that Proteobacteria was the most abundant phylum in FM, comprising more than 80% of the microflora, which agrees with reports that Proteobacteria was the dominant phylum in fresh forage (Dong et al., 2019). After 30 days of ensiling, the abundance of Firmicutes increased and dominated the fermentation in the native grass silages (80%), and the abundance of Proteobacteria decreased. These results were similar to the reports of Yang et al. (2019) and Bai et al. (2020). Compared to the control, the abundance of Firmicutes and Proteobacteria was significantly increased and lower in L and XM2 groups, whereas no significant differences were observed among control, 265, and 842 groups, which the higher microbial populations of LAB could explain.

This study showed that Pantoea was the major facultative aerobic genera in FM. Pantoea has been discovered in fresh alfalfa (Ogunade et al., 2018) and soybean (Ni et al., 2017). After 30 days of fermentation, drops in Pantoea abundances could be attributed to their high sensitivity to pH decline (Ogunade et al., 2018). Lactobacillus, Pedicoccus, Weissella, and Leuconostoc are considered as the four most predominant LAB genera responsible for driving lactic fermentation during ensiling (Pang et al., 2012; Ni et al., 2017; Guan et al., 2018; Liu et al., 2019). During ensiling, a significant shift in the bacterial community from Proteobacteria to Firmicutes could be explained by increased abundance of genera Lactobacillus, Weissella, and Pediococcus, which flourished in the environmental conditions developed during ensiling (Keshri et al., 2019). Therefore, the abundance of Lactobacillus, Enterobacter, Pediococcus, and Weissella increased and dominated the silage fermentation among silages (80%). In this study, the LAB-treated silages exhibited an increase in Lactobacillus than in the control, which agrees with the findings that LAB-treated silages could increase the abundance of Lactobacillus (Liu et al., 2019; Yan et al., 2019; Bai et al., 2020). Additionally, the highest abundance of Lactobacillus was found in the XM2 group; strain XM2 isolated from native grass and the higher microbial populations in the XM2 group may be the main reason. Weissella was the other predominant microbe in all silages throughout the fermentation and was the early colonizer (Graf et al., 2016; Guan et al., 2018); Pediococcus contributed to an initial decline in silage pH, resulting in an anaerobic environment suitable for developing Lactobacillus (Yang et al., 2019). The previously published studies showed that Weissella was heterofermentative. These two LAB genera could not thrive at a pH environment lower than 4.5 and were thus active only during the early stages of ensiling. The follow-up lactic acid production mainly depends on Lactobacillus, which becomes active and thrives as pH decreases (Cai et al., 1998). However, the abundance of Weissella was significantly reduced in the 842 group, which could be explained by adding L. graminis, which could inhibit the growth of Weissella. The abundance of Pediococcus significantly increased in the 265 group, which could contribute to the addition of P. acidilactici. Lactobacillus was enriched in the XM2 group, and Pediococcus was enriched in the 265 group, which is in agreement with the addition of LAB in the XM2 and 265 groups.

The bacterial community of fresh native grass was discovered to be dominated by Proteobacteria, Actinobacteriota, and Cyanobacteria. This study demonstrated that the addition of LAB could influence silage fermentation by reconstructing microbiota. Lactobacillus was the dominant genus in the native grass silages, followed by Enterobacter and Pediococcus. Strain XM2 exhibited the potential possibility to respond to improving silage fermentation in native grass. Further results indicated that strain XM2 could effectively improve the silage quality, and it was proposed to be a potential starter culture for native grass silage.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

SY: investigation, methodology, visualization, validation, data curation, writing—original draft, and conceptualization. SD: investigation, software, formal analysis, and writing—review and editing. GG: conceptualization, funding acquisition, supervision, and writing—review and editing. TW: project administration and supervision. YJ: conceptualization, funding acquisition, project administration, and supervision. All authors contributed intellectual input and assisted with this study and manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.