Effect of phenyllactic acid on silage fermentation and bacterial community of reed canary grass on the Qinghai Tibetan Plateau

This study aimed to investigate the effect of phenyllactic acid as an additive on silage fermentation and bacterial community of reed canary grass (RCG, Phalaris arundinacea L.) on the Qinghai Tibetan Plateau. At the heading stage, RCG was harvested, chopped and ensiled in small bag silos. The silage was treated without (control, 1.0 g/mL sterile water, on a fresh matter basis (FM)) or with phenyllactic acid (PLA, 3 mg/mL, FM), antimicrobial additive (PSB, a mixture of potassium sorbate and sodium benzoate, 2%, FM), lactic acid bacteria inoculant (LABi, L. plantarum + L. curvatus, 1 × 106 cfu/g, FM) and PLA + LABi, and then stored in a dark room at the ambient temperature (5 ~ 15 °C) for 60 days. Compared with control, PLA decreased lactic acid, acetic acid and ammonia-N contents, and subsequently increased CP content of RCG silage. PLA enhanced the growth of lactic acid bacteria and reduced the count of yeasts (P < 0.05) in RCG silage, with reduced bacterial richness index (Chao1), observed operational taxonomic units and diversity index (Simpson). In relative to control, moreover, PLA and PLA + LABi increased the relative abundance of Lactococcus in RCG silage by 27.73 and 16.93%, respectively. Therefore, phenyllactic acid at ensiling improved nutritional quality of RCG silage by advancing the disappearance of yeasts and the dominance of Lactococcus.


Background
The Qinghai-Tibetan Plateau (QTP) is an important alpine grassland livestock production region [1]. However, the unstable weather conditions result in forage shortage for herbivores in winter and early spring [2], which cause yaks' weight loss, low milk production and other problems [3]. Therefore, how to preserve local forages for enhancing milk production and meat quality of yaks is a concerned issue [4].
Ensiling is a good way to preserve the nutritional value of forages, especially in areas where forages are seasonally or regionally unbalanced due to the harsh environment. In cold regions, silage is considered to be a main source of feed for ruminants [5]. Recently, Chen et al. and Zhao et al. have reported that chemical and microbial additives can improve the quality of silage on the Qinghai Tibetan Plateau [6,7]. However, silage also faces challenges in cold regions. For example, silage exposed to the air is easy to deteriorate, and some desirable inherent/exogenous microbes (mainly lactic acid bacteria, LAB) are scarce in a low temperature environment [8,9]. In addition, yeasts and other harmful microorganisms Open Access *Correspondence: lpyzm@sina.cn; baishiqie@126.com 3 Institute of Qinghai-Tibet Plateau Research, Southwest Minzu University, Chengdu 610041, China Full list of author information is available at the end of the article showed a higher activity in silages under cool/cold conditions [10]. These reduce the positive effect from additives on preservation of silage nutrients. Importantly, some studies have reported that the use of some conventional chemical additives such as formic acid can cause different health problems to people [11]. Therefore, it is necessary to explore new alternatives to stabilize silage quality for safe production of milk and meat in cold regions.
Phenyllactic acid (PLA) is a kind of small molecule organic acid, which exists widely in the nature and is harmless to humans and animals [12][13][14]. Rvr et al. reported that PLA has been used as a substitute for antibiotics in livestock feed and dairy products, with a good effect on extending the storage period of dairy products and improving the quality of meat products [11]. However, there is little information on PLA-treated silage [15].
The reed canary grass (RCG, Phalaris arundinacea L.) is a kind of promising high-yield cool-season forage. It can tolerate different environmental conditions, including cold and moist conditions. It also had a high-yield potential during the harvest season. Despite of many reports about RCG silage on the QTP [16][17][18], there is still little information on PLA-treated RCG silage. Therefore, an experiment was designed to investigate the effect of PLA on the fermentation quality and bacterial community of RCG silage on the QTP. We supposed that (1) PLA can improve the fermentation quality of reed canary grass silage. (2) PLA can change the diversity of the bacterial community and the composition of the recombinant bacterial community.

Silage preparation
The experiment was conducted at the experiment base of Sichuan Academy of Grassland Science (N 31°51′-33°33′, E 101°51′-103°22′, altitude 3500 m; Hongyuan, Sichuan, P.R. China). RCG (P. arundinacea L. 'Chuancaoyin No.3' cultivated by Sichuan Academy of Grassland Sciences) at the heading stage was harvested manually, and chopped by the length of 1-3 cm. According to previous report from Chen et al., pre-ensiled RCG featured water soluble carbohydrates (WSC) of 7.88%DM, crude protein (CP) of 12.91%DM, neutral detergent fiber (NDF) 53.96%DM and acid detergent fiber (ADF) of 30.92%DM, lactic acid bacteria (LAB) of < 3.0 log 10 cfu/g FM and yeast and mold of > 4.0 log 10 cfu/g FM [18]. The chopped RCG was randomly divided into five equal parts for the following treatments: no additives (control); lactic acid bacteria inoculant (LABi, a mixture of L. plantarum and L. buchneri, each 10 6 cfu/g fresh matter (FM) base); phenyllactic acid (PLA, 3 mg/kg FM); antimicrobial additive (PSB, a mixture of potassium sorbate and sodium benzoate, each 1.0% FM); and PLA+ LABi. Each treatment was repeated for three times. About 300 g of the mixed-RCG was put into 30 cm × 25 cm polyethylene bag, vacuumed and sealed by a vacuum sealer, and stored at ambient temperature (5 ~ 15 °C) for 60 days.

Chemical analysis
The silage samples were dried at 65 °C for a constant weight to determine the content of DM [19]. And then they were ground with a 1 mm sieve for other nutrient analysis. The CP was determined by the method of Kjeldahl. Both NDF and ADF were determined using an Ankom 2000 fiber analyzer (Ankom Technology, Fairport, NY). The WSC was measured by the method of Murphy [20], weighed 0.2 g of the crushed sample and placed it in a test tube, added 10 mL distilled water and boiled for 30 min, filtered in a 25 mL volumetric flask, pipetted 1 mL of the above sugar extract, put into a test tube, added 5 mL anthrone solution, and boiled for 10 min, took out to cool, then measured the absorbance value (OD) at the 620 nm wavelength, used the standard curve to calculate the sugar extract content, finally, obtained the soluble sugar content by formula.
Note: A represents the volume of extraction liquid. C represents the sugar content in extract (ug/mL). W represents the sample weight (g). (the standard curve line y = 0.0075x + 0.0665 R 2 = 0.9984 y: OD x = C) Fresh sample (20 g) was mixed with 180 mL ultrapure water for 3 min in a Stomacher blender. The pH of filtrate was determined by a pH meter. About 10 mL was filtrated by centrifugation (4500×g, 15 min, 4 °C), and the supernatant was analyzed for contents of lactic acid (LA), acetic acid (AA), propionic acid (PA) and butyric acid (BA) by high performance liquid chromatography [21] (Color spectrum column: Shodex Rspak KC-811 S-DVB gel column, Detector: SPD-M10AVP, the mobile phase used: 3 mmol/L high chloric acid solution, the set flow rate: 1 mL/min, the column temperature: 50 °C, the detection wave: 210 nm, Sample volume: 5 μL, injection time: 25 min). Ammonia nitrogen (AN) was determined by the methods of Broderick and Kang [22], took 1 mL filtered sample solution, added 4 mL of 0.2 mol/L hydrochloric acid solution and mixed well, took 0.2 mL of the mixed solution in another test tube, added 2.5 mL phenol solution and 2.5 mL sodium hypochlorite solution in turn, mixed well, placed it in a 60 °C constant temperature water bath to react for 10 min, Then measured the absorbance value (OD) at the 560 nm wavelength, used the NH 4 Cl standard curve to calculate the NH 4 Cl concentration in the sample solution.
Note: C represents the NH 4 Cl concentration in the sample solution (mg/mL). M represents the NH 4 Cl relative molecular mass. W represents the weight of 20 g sample (g). (the standard curve line y = 1.5907x -0.0087 R 2 = 0.999 y: OD x = C)

Microbial analysis
Microbial population on fresh samples was determined by the method of Cai [23]. Fresh samples (10 g) were put into a sterile glass bottle, and the concentration was diluted from 10 − 1 to 10 −3 with sterile water for serial dilution. Lactic acid bacteria were counted on MRS agar medium (GCM188, Beijing Luqiao Technology Co., Ltd., Beijing, China), yeasts and molds were numbered on malt extract agar (021110, Huankai Microbial Technology Co., Ltd., Guangzhou, China). The microbial count was expressed as log 10 cfu/g FM.
The DNA of bacteria in the silage sample was extracted by the method of Li et al. [19]. And the Phusion ® highfidelity PCR master mix (New England Biolabs) was used for the PCR reaction. 515F (5′-CCT ACG GGA GGC AGCAG-3′) and 907R (5′-TTA CCG CGG CTG CTGGC-3) were taken as primers to amplify the 16S rRNA gene. The PCR products were purified using Qiagen Gel Extraction Kit (Qiagen, Germany), and then sequenced at the paired ends (250 bp) of the Illumina MiSeq PE2500 platform at Novogene Company. All barcodes and primers were discarded to obtain high-quality sequences. After sequencing, trimmomatic was used to process the original sequence, and the PE reading was overlapped with FLASH (V 1.

Statistical analysis
Data of silages was subjected to one-way analysis ANOVA. Significant differences between means were considered significant for P < 0.05 by Tukey's Studentized range test. This analysis was performed using the SPSS 19.0 program.

Chemical composition and microbial population of reed canary grass silage
The chemical composition of RCG silage was shown in Table 1. Compared with control, PLA decreased (P < 0.05) the WSC and ADF contents, and increased (P < 0.05) the CP content of RCG silage; LABi increased (P < 0.05) the WSC and CP contents of silage; PSB, PLA and LABi increased the count of LAB. In addition, yeasts were not observed in silage treated with PLA and PLA + LABi.
As shown in Table 2, all additives reduced (P < 0.05) the lactic acid content of silage as compared to control. Silages treated with LABi and PLA + LABi had lower (P < 0.05) pH value, acetic acid content and ammonia-N ratio of total N.

Bacterial composition of reed canary grass silage
The bacterial alpha diversity of silage is shown in Table 3. Compared with control, PLA and LABi reduced (P < 0.05) the bacterial Chao1, observed OTUs and Simpson indices in silage. In addition, PLA and LABi significantly reduced (P < 0.05) the observed OTUs and Simpson indices as compared with PSB.
The bacterial community of silage was shown in Fig. 1. At the genus level, Carnobacterium (54.94%) and Lactobacillus (35.45%) were observed in control silage (Fig. 1a). PLA increased the relative abundance  (Fig.1b). Low dominance of Lactobacillus species (L. curvatus and L. buchneri) were observed in PLA silage. Lactobacillus species (L. plantarum and L. curvatus) became a dominant flora in LABi silage. PLA + LABi silage increased the relative abundance of L. plantarum and L. buchneri as compared with control silage. The changes in fermentation characteristics during ensiling are inseparable from the metabolism of microbial flora. As shown in Fig. 2, Carnobacterium was positively correlated (P < 0.05) with pH value and lactic acid content in silage. Ralstonia was highly positively correlated (P < 0.01) with acetic acid content. The relative abundance of L. plantarum was negatively correlated (P < 0.05) with pH value. The relative abundance of L. curvatus was negatively correlated (P < 0.05) with acetic acid content. The count of Erwinia persicina was negatively correlated with propionic acid content (P < 0.05).

Chemical composition and microbial population of reed canary grass silage
The CP is an important nutritional indictor in silage. In this study, higher CP content was observed in PLA and LABi silages. PLA can inhibit the growth of yeasts to reduce the consumption of nutrients [24]. In addition, LAB can secrete antibacterial substances to inhibit the growth of harmful microorganisms [25]. PLA reduced ADF content in RCG silage. Similar results were from Li et al. who reported that the fiber content of silage reduced as the ratio of crude protein increased [26]. Therefore, to some extent, PLA increased the nutritional value of the silage. The beneficial bacteria, LAB, played a significant part in the silage fermentation process [27]. Under anaerobic conditions, LAB can reduce pH value by generating organic acids (mainly lactic acid) and secrete antibacterial substances to inhibit the growth of undesirable microorganisms [19,23]. In this study, applications of PSB, PLA and LABi at ensiling promoted the growth of LAB in silage. In this study, as expected, PLA alone or in combine with LABi decreased the count of yeasts in silage.
The pH value is one of the important indicators for evaluating the quality of silage [28]. Too high pH value will promote the growth of acid-intolerant and harmful microorganisms such as Clostridium, making the silage   [19]. It will promote LAB to produce lactic acid and acetic acid for reducing the pH value in silage. In this study, PLA + LABi silage had a lower pH value than control silage, which may be related to the combined effect of phenyllactic acid and exogenous LAB. Low lactic acid content was observed in PSB and PLAtreated silages. A high pH value reduced the antibacterial effects of potassium sorbate and sodium benzoate. A study from Weinberg & Muck showed that potassium sorbate and sodium benzoate had a poor antibacterial effect inhibiting the growth of undesirable microorganisms, such as yeasts and molds at the pH value > 5 [29]. Previous studies showed that PLA could inhibit the fermentation of LAB [30]. Meanwhile, exogenous LAB produced biological antagonism against inherent LAB adhered to the raw materials [31]. Acetic acid showed the second highest concentration of organic acid in silage. The appropriate concentration of acetic acid was beneficial to the fermentation of silage. Low acetic acid content may increase the potential of aerobic deterioration in silage [8]. In this study, treatments with LABi and PLA + LABi had low acetic acid content. This was attributed to the dry matter content [8]. In this study, higher butyric acid content was observed in PLA silage. This may be related to that both Clostridia and plant proteolytic enzymes were active at the pH value > 5 [32]. This also indicated that ammonia-N was not significantly reduced in PLA silage. Ammonia-N is also an important index for evaluating the quality of silage, which reflected the activity of plant proteases or the degree of protein degradation based on clostridial fermentation [33]. In this study, treatments with PSB, LABi and PLA + LABi had lower ammonia-N contents than the control, which indicated that PSB, LABi and PLA + LABi inhibited the

Microbial composition of reed canary grass silage
In this study, the goods coverage was > 0.99, indicating that sequencing depth covered the microbial composition in silage. PLA-treated silage had lower Chao1, observed OTUs and Simpson indices, which suggested that PLA reduced the microbial diversity in silage. The possible reason was that LAB became the dominant group in PLA silage and inhibited the growth of other bacteria [34]. However, silage treated with PLA + LABi showed a higher diversity as compared with control. This may be because the organic acid content was too low to inhibit bacteria in PLA + LABi silage. Moreover, we found that PLA silage had lower observed OTUs and Simpson indexes than PSB silage, which may be attributed to the growth and propagation of LAB in silage. Previous studies showed that LAB can produce some antibacterial substances in the metabolic process to effectively inhibit the growth of harmful microorganisms, such as yeasts, molds and clostridiums [25]. Carnobacterium and Lactobacillus were dominant genera in control silage. Genus of Carnobacterium was considered as harmful bacteria, which may cause the poor quality of silage [35]. PLA silage increased the abundance of Carnobacterium and Lactococcus. Furthermore, PLA silage decreased the relative abundance of Lactobacillus. This may be attributed to the high pH value being observed in PLA silage. Previous studies revealed that it was more conducive to the growth of Carnobacterium and Lactococcus in a high-pH environment, but was not good for the growth of more acid-tolerant Lactobacillus [36][37][38]. Lactococcus is one of the ideal functional bacteria in silage fermentation process. During fermentation, they secrete lactic acid to lower the pH value, thereby improving the quality of silage [39,40]. Therefore, they played an important role in improving the quality of silage. In addition, Lactobacillus became a dominant genus in LABitreated silage, which was consistent to the results of Bai et al. who reported that the inoculation of LAB can increase the abundance of Lactobacillus in silage [34]. The dominant genus was Lactobacillus in PLA + LABi silage, which may be attributable to the fact that PLA and LAB had the joint broad-spectrum antibacterial effect for inhibiting the growth of undesirable microorganisms [41].
L. curvatus, as a facultative hetero-fermentative LAB, dominated in control silage. Similar results were reported by Terán et al. [42]. In addition, it had better lactic acid productivity than L. plantarum under low temperature conditions, which may indicate the high lactic acid content in control silage [43]. Previous studies showed that potassium sorbate and other inorganic salt additives can generate corresponding organic acids when dissolved in water. These organic acids can penetrate into the cell membranes of LAB, destroying their biological activity and inhibiting their growth [44]. PLA silage decreased the abundance of Sugarcane phytoplasma (uncharacterized), L. plantarum and L. curvatus, which indicated that PLA can inhibit both undesirable microorganisms and LAB due to the high concentration of propionic acid in PLA silage, which can inhibit the activity of microorganisms including LAB [19]. However, a study from Jung et al. showed that PLA can promote the growth of LAB. In addition, Wu's findings showed that PLA had no effect on the growth of LAB [15]. Therefore, the effect of PLA on the growth of LAB needed to be further studied. For example, we can explore the mechanism of action of PLA on the growth of LAB through macro-genomics and metabonomics.
Carnobacterium is gram-positive bacteria, which is able to produce lactic acid from glucose and closely related to Enterococcus [36]. However, it is often found in poor-fermentation silage with a high pH value [45]. Many studies proved that L. plantarum can effectively reduce the pH value [46,47]. This suggested the relationship between L. plantarum and pH value. Previous studies showed that bacteria of the genus Ralstonia can secrete short-chain fatty acids to alleviate the inflammatory response in intestine [48]. Acetic acid is a kind of short-chain fatty acid, so the increased count of Ralstonia can increase acetic acid content. Species of L. buchneri is a hetero-fermentative LAB, which is able to metabolize 1,2-propanediol to produce propionic acid [49]. Propionic acid can inhibit the activity of acetic acid-producing microorganisms, thereby reducing acetic acid content [50]. It is well known that propionic acid can inhibit the growth of undesirable microorganisms, such as Erwiniapersicina [51,52].

Conclusion
Application of PLA during ensiling can improve the nutritional quality of RCG silage. PLA promoted the growth of LAB and reduced the bacterial alpha diversity in silage. The combination of PLA and LABi increased the relative abundance of Lactobacillus and shaped the bacterial community of silage. In conclusion, this study confirmed that PLA, as an antifungal silage additive, exerted a good effect in improving the quality of silage. However, a further study will conduct to explore the effect of PLA-producing bacteria on silage quality.