Introduction

In modern dairy production systems, the energy requirements of high-producing dairy cattle are usually met by feeding large amounts of rapidly fermentable carbohydrates. However, the inclusion of these carbohydrates can lead to an accumulation of volatile fatty acids (VFA), which causes a drop in the rumen pH. If the increase in VFA production rises above the absorptive capacity of the rumen epithelium, the pH will continue to drop, and this can result in subacute ruminal acidosis (SARA) [1, 2]. SARA, which is characterized by a ruminal pH below 5.6 for 3–5 h per day [3], can result in reductions in feed intake and fiber digestion, as well as decreases in milk yield and declines in the health of the animals [4].

During SARA, the accumulation of VFA decreases ruminal pH and causes a change in the rumen microbiota [5]. Concurrent with the shift in the microbial populations, the concentrations of potentially toxic and inflammatory compounds increase in the rumen. One compound of particular interest is lipopolysaccharide (LPS), a component of the cell walls of the gram-negative bacteria, as it can elicit an inflammatory response in the mammalian cells [6]. When animals are challenged with a SARA-inducing ration, the easy access to fermentable carbohydrates initially results in a logarithmic growth of bacteria. This is later followed by a release of LPS due to massive bacterial lysis that occurs in response to the reduced availability of substrates, the greatly reduced rumen pH, and the accumulation of the end products of rumen fermentation [7,8,9].

Other potentially harmful compounds produced during SARA, in addition to LPS, include biogenic amines, which are believed to associate with lameness [10]. Previous studies showed that, during SARA, a low ruminal pH together with high levels of LPS and biogenic amines, can reduce the barrier function of the rumen epithelium [10, 11]. This barrier damage potentially allows LPS and biogenic amines to translocate into the blood, where they stimulate a systemic inflammatory response [6]. This emphasizes the importance of preventing the accumulation of LPS and bioamines, as well as the decrease in the ruminal pH, in animals experiencing SARA.

Many strategies have been used to improve the ruminal pH during SARA. Ionophores, organic acids, and probiotics based on yeasts (e.g., Saccharomyces cerevisiae) have been explored as stabilizers of the ruminal pH and for improving milk production [12,13,14]. Among these ruminal modulators, fumarate has been proposed as a promising modifier of ruminal fermentation [15]. Fumarate is thought to increase the activity of the succinate–propionate metabolic pathway of several rumen bacteria, which results in increased lactic acid uptake and propionate production [16]. However, the potential for fumarate to serve as a mitigator of the concentration of ruminal LPS and bioamines remains to be established. Besides, previous studies have revealed the fumarate can stimulate the growth of fumarate-utilizing bacteria, including Fibrobacter succinogenes, Selenomonas ruminantium, Veillonella parvula, and Wolinella succinogenes in the rumen [12]; however, information is limited regarding the effects of fumarate on the composition of the ruminal bacterial community.

In this study, we hypothesized that fumarate could positively affect the composition of the ruminal bacterial community, thereby possibly mitigating the accumulation of LPS and bioamines in the rumen. The effects of disodium fumarate (DF) were examined on the changes of the in vitro rumen fermentation profile, the concentration of LPS and bioamines, and the composition of the ruminal microbial community.

Methods

Animals

The experimental design and procedures were approved by the Animal Care and Use Committee of Nanjing Agricultural University. Animals were cared for in accordance with the 1998 guidelines established by the Chinese Science and Technology Committee on Experimental Animal Care and Use.

Three ruminally cannulated Holstein cows (500 ± 23.5 kg BW) served as rumen fluid donors for these experiments. Cattle were fed a 60% alfalfa hay and 40% concentrate (corn grain-based) diet.

Experimental Design

The experimental procedure was as described by Martínez-Fernández et al. [17]. Ruminal contents were obtained immediately before the morning feeding and strained through four layers of cheesecloth. This filtered rumen fluid was mixed with a buffer solution at a ratio of 1:3 (v/v). A 40 mL volume of the mixture was dispensed into a 100-mL bottle containing 1 g substrate (0.49 g soybean meal, 0.21 g maize silage, 0.15 g Chinese wildrye, and 0.15 g alfalfa hay). This substrate was milled through a 1 mm screen before being weighed into the bottles. The final concentration of disodium fumarate (DF) (Sigma) in each group is 0, 4, 8, and 16 mmol/L. The bottles were sealed with rubber stoppers and then aluminum caps, and incubated for 24 h at 39 °C. Each group had four replicates.

Analytical Procedures

Gas production was measured at different time intervals using a pressure transducer and a calibrated syringe [18]. After a 24 h-incubation, the bottles were uncapped and the pH was determined immediately using a pH meter (Ecoscan pH 5, Singapore). The bottles were then immersed in an ice bath to stop the fermentation. The contents of each bottle were homogenized before withdrawing the samples. Short chain fatty acids were determined according to Jin et al. [19], by gas chromatography (GC-14B, Shimadzu, Japan) with a capillary column (Supelco, Bellefonte, USA) at a column temperature of 110 °C; injector temperature of 180 °C; detector temperature of 180 °C. Ammonia–nitrogen (NH3–N) was measured by the indophenol method [20]. Lactic acid was analyzed according to the method of Barker [21]. LPS was measured by a chromogenic end-point Tachypleus amebocyte lysate assay kit (Chinese Horseshoe Crab Reagent Manufactory, Xiamen, China). The concentration of biogenic amine was analyzed according to the method of Wang et al. [22].

DNA Extraction

DNA was extracted using a QIAamp DNA Stool Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. A FastPrep®-24 Instrument (MP Biomedicals, South Florida, USA) was used for processing (bead beating) at a setting of 5 for 2 min. The DNA concentration was measured using a NanoDrop 1000 Spectrophotometer (Nyxor Biotech, Paris, France).

Illumina MiSeq Sequencing and Data Processing

The bacterial communities in the 0 and 16 mmol/L DF groups were determined according to the method of Liu et al. [23]. The V3–V4 regions of the bacterial 16S rRNA genes were amplified using primers 338F (5′-barcode-ACT CCT RCG GGA GGC AGC AG)-3′ and 806R (5′-GGA CTA CCV GGG TAT CTA AT-3′). PCR amplicons were sequenced on an Illumina MiSeq platform according to standard protocols [24], by Majorbio Bio-Pharm Technology Co., Ltd., Shanghai, China. The raw data were processed using the QIIME (version 1.70) software package [25]. The minimum quality score was 20. Sequences with two nucleotide mismatchs and ambiguous characters and those shorter than 50 bp were discarded. Only sequences that overlapped longer than 10 bp were assembled. OTUs were clustered with a 97% similarity cutoff using UPARSE (version 7.1 http://drive5.com/uparse/). Chimeric sequences were identified and removed using UCHIME [26]. The representative sequences were aligned against Greengenes 13.5 using PyNAST [27], with the default parameters set by QIIME. Sequences were classified using the Ribosomal Database Project classifier with a standard minimum support threshold of 80% [28]. Community diversity was estimated using the ACE, Chao1, and Shannon indexes. Principal coordinate analysis (PCoA) was performed by the unweighted UniFrac distance method [29]. The analysis of molecular variance (AMOVA) was conducted based on an unweighted distance to assess the significant differences among the samples using the MOTHUR program (version 1.29) [30].

Statistical Analysis

Fermentation data were analyzed using a one-way analysis of variance (ANOVA) with the statistical software package SPSS vs. 20.0 (SPSS Inc., Chicago, IL, USA). Linear and quadratic effects due to DF addition were determined using polynomial contrasts. The microbial data were analyzed using the nonparametric Kruskal–Wallis test. All P values from the nonparametric Kruskal–Wallis test for microbial data were adjusted by the false discovery rate. Differences were considered statistically significant at P < 0.05.

Results

Effect of DF on the Lipopolysaccharide and Biogenic Amines Contents and The Characteristics of In Vitro Rumen Fermentation

As shown in Table 1, DF addition linearly increased the total gas production, pH, and the concentrations of propionate, valerate, NH3–N, and total volatile fatty acids (TVFA) (P < 0.05), and linearly decreased the concentration of lactic acid and the ratio of acetate to propionate (P < 0.05). No significant differences were observed in the concentrations of acetate, butyrate, isobutyrate, and isovalerate between the control and the DF treatment groups (P > 0.05). DF addition also linearly decreased the concentrations of LPS, methylamine, tryptamine, putrescine, histamine, and tyramine (P < 0.05).

Table 1 Effects of disodium fumarate on the ruminal fermentation profiles, biogenic amines and lipopolysaccharide
Full size table

Effects of DF on The Ruminal Bacterial Community

The Illumina sequencing reads that passed the quality control tests included 133,520 from the control and 138,870 from the DF group. In total, 799 OTUs were formed from the control samples, and 804 OTUs were formed from the DF group samples. Rarefaction curves showed that the sequencing effort covered the majority of bacterial diversity (Supplemental Figure S1). The number of OTUs, Shannon index, Chao1 index, and ACE index did not differ significantly between the control and DF groups (P > 0.05, Supplemental Table S1). Principal coordinate analysis (PCoA) showed that the bacterial communities differed between the control and DF groups (Supplemental Figure S2), as confirmed by the molecular variance analysis (AMOVA, Fs = 2.74668, P = 0.031).

At the phylum level, DF decreased the proportion of Proteobacteria (P < 0.05), and increased the proportions of Spirochaetae and Elusimicrobia (P < 0.05) (Supplemental Table S2). At the genus level, DF addition decreased the proportions of Ruminobacter, Acetitomaculum, Moryella, Bacillus, Staphylococcus, unclassified Christensenellaceae, unclassified Succinivibrionaceae, and unclassified Defluviitaleaceae (P < 0.05), and it increased the proportions of Succinivibrio, Treponema, Blautia, Sutterella, Anaerovorax, Anaerosporobacter, Oligosphaeraceae, Schwartzia, unclassified Bacteroidales, and unclassified Elusimicrobia (P < 0.05, Fig. 1; Supplemental Table S3).

Fig. 1
figure 1

Double dendrogram of the top 30 most abundant genera in the control and disodium fumarate groups

Full size image

The Correlations Between The Dominant Bacterial Phyla and Rumen Fermentation Parameters

As shown in Supplemental Table S4, Pearson correlation analysis revealed that the presence of Proteobacteria was positively correlated with the concentrations of methylamine, tryptamine, and putrescine and negatively associated with the concentrations of acetate, propionate, isovalerate, valerate, and TVFA (P < 0.05). Lentisphaerae were positively correlated with propionate and TVFA and negatively correlated with lactate and putrescine (P < 0.05). Spirochaetae were positively correlated with ruminal pH, propionate, and NH3–N and negatively associated with lactate, putrescine, tyramine, and the ratio of acetate to propionate (P < 0.05). Elusimicrobia were negatively correlated with the level of LPS (P < 0.05). Tenericutes was positively correlated with methylamine and tryptamine (P < 0.05).

Discussion

The addition of DF modified the in vitro rumen fermentation. In general, DF decreased the ruminal lactate concentration and increased the propionate concentration and the pH value, in agreement with previous reports [16, 31]. Fumarate is an intermediate of the citric acid cycle. In the rumen, it is reduced to succinate through succinate–propionate pathway, and then decarboxylated to propionate [32, 33]. Fumarate can also increase the microbial lactate uptake by microbes such as Selenomonas ruminantium [34], which might partly explain the decrease in the lactate concentration observed in the current study. However, malate has been reported to increase lactate uptake by Selenomonas ruminantium, whereas fumarate had no effect [35]. The actual mechanism that leads to the decrease in the lactate concentration therefore remains unclear.

The concentration of the toxic LPS, a component of the cell walls of gram-negative bacteria, was decreased by DF addition. Previous studies showed that feeding high levels of grain to dairy cows led to a decline in the ruminal pH [4, 36]. This, in turn, caused the death and cell lysis of gram-negative bacteria, thereby increasing the concentration of free LPS in the rumen [4]. The concentration of LPS might therefore be closely related to the abundance of gram-negative bacteria. In the current study, the concentration of LPS was negatively correlated with the relative abundance of Elusimicrobia and a group of unclassified bacteria. The cultivated members of Elusimicrobia, namely Elusimicrobium minutum and Endomicrobium proavitum, were gram-negative bacteria [37, 38].

The addition of DF decreased the concentrations of biogenic amines and increased the pH value. Wang et al. [22] reported that the concentrations of biogenic amines were negatively correlated with the ruminal pH value. A decrease in the ruminal pH was considered to increase the activity of bacterial amino acid decarboxylases, as well as promoting the growth of some microbes, such as Streptococcus bovis and Lactobacillus spp., which generate biogenic amines [22, 39]. In the current study, no increase was observed in the relative abundance of Lactobacillus and Streptococcus, which suggested that other bacteria might be involved in the generation of biogenic amines. In the present study, the concentrations of biogenic amines, such as methylamine, tryptamine, and putrescine, were positively correlated with the relative abundance of Proteobacteria. The members of this bacterial phyla were considered to be involved in the ruminal proteolysis.

Disodium fumarate addition increased the percentages of the dominant genera Succinivibrio and Treponema. Succinivibrio spp. are involved in the metabolism of succinate and lactate in the rumen [40]. Mao et al. [31] also reported an increase in Succinivibrio dextrinisolvens in the rumen of goats fed DF. The non pathogenic Treponema spp. are commensal in the gastrointestinal tract of cows [41], and they are also involved in the metabolism of succinate and lactate in the rumen [42]. The relative abundance of Ruminobacter also decreased following the addition of DF. Ruminobacter spp., such as Ruminobacter amylophilus 70, are mainly known for the amylolytic activities [43], which suggests that the degradation rate of starch might decline.

In summary, DF has a beneficial effect on the rumen fermentation of high concentrate feeds by increasing pH and propionate production, enhancing lactate utilization, and reducing the concentrations of LPS and biogenic amines. DF therefore has great potential for use as a feed additive for preventing SARA in ruminants.