Two kinds of roughages (CWR vs CS) were used as roughage substrates. They were analyzed for dry matter (DM), crude protein (CP), ash (AOAC, 1997), neutral detergent fiber (NDF) and acid detergent fiber (ADF) expressed inclusive of residual ash (Van Soest et al., 1991). Chemical composition of the two roughages is presented in Table 1. Each roughage substrate was assigned to all the treatments in a 2×3 factorial arrangement with AOC and HMB as main effects. The AOC (Amaferm, Biozyme, Inc., St. Joseph, MO, USA), was added into substrates at the level of 0, 3, or 6 mg. The HMB (MFP, Novus Int. Inc. St. Louis, MO, USA), was added into substrates at the level of 0 or 15 mg together with AOC. The treatments included: control (no AOC or HMB), AOC group without HMB (3 mg or 6 mg), HMB group without AOC (15 mg) and AOC plus HMB group (3 mg AOC+15 mg HMB; 6 mg AOC+15 mg HMB). Each treatment was performed in triplicate.

The substrates in the in vitro test consisted of finely ground corn (100 mg) and one kind of roughage (100 mg), CWR or CS. The gas production (GP) was determined according to the method of Menke and Steingass (1988). Rumen fluids were collected before morning feeding from three rumen-fistulated cows fed twice daily on a mixed diet. Ingredients and chemical composition (% DM) of the diet are shown in Table 2.

Rumen fluid was strained through four layers of gauze into a pre-warmed syringes flushed with CO₂ at 39°C. Syringes were filled with 30 mL medium consisting of 10 mL rumen fluid and 20 mL artificial saliva. The syringes were incubated in a water bath (39°C) for 24 h. The GP in vitro was recorded at 3, 6, 9, 12, and 24 h incubation times.

The incubation was stopped at 24 h fermentation and the pH of rumen liquor was determined immediately using a pH meter (S40 SevenMulti, Mettler Toledo, Shanghai, China). The incubated inoculants were used for analysis of ammonia-N, VFA, microbial protein (MCP) and microbial communities. Ammonia-N, VFA and MCP concentrations were determined using methods of Hu et al. (2005).

Total DNA was isolated using a genomic DNA kit (Axygen Biosciences, Union City, CA, USA) following the manufacturer’s instructions. The amplifying primer sets of total bacteria, protozoa, fungi, F. succinogenes, R. flavefaciens and R. albus were designed as described by Denman and McSweeney (2006), Denman et al. (2007), and Koike and Kobayashi (2001), respectively as in Table 3. The species-specific real-time qPCR was performed using the ABI 7500 real time PCR system (Applied Biosystems, Foster City, CA, USA) with fluorescence detection of SYBR green dye. Samples were run under the amplification conditions as described by Mao et al. (2010).

To describe the dynamics of GP over time, the following equation (Ørskov and McDonald, 1979) was chosen:

where, GP = cumulative GP (mL) at t time, t = incubation time (h), a = the gas production from immediately soluble fraction (mL), b = the gas production from immediately insoluble fraction (mL), (a+b) = potential GP (mL), c = rate constant of GP (%/h), and a, b and c are constants.

The following equation of Menke and Steingass (1988) was used to calculate the in vitro organic matter digestibility (IVOMD):

where, GP = cumulative GP (mL) at 24 h, CP = crude protein content of substrates (g/kg).

Populations of protozoa, fungi, F. succinogenes, R. albus, and R. flavefaciens, were expressed as a proportion of total rumen bacterial 16S rDNA according to the equation of Chen et al. (2008):

where, Ct represented threshold cycle.

All of the statistical analyses were conducted using the general linear models (GLM) procedures of SAS software (SAS Institute, 1999). Data from CWR and CS were analyzed separately as a completely randomized design with HMB, AOC and their interaction included. Effect of AOC supplementation was determined by the “contrast” option of the GLM procedure. When this effect was significant (p<0.05), orthogonal polynomial (ORPOLY) contrasts using contrast coefficients that were obtained by ORPOLY macro were used to determine linear and quadratic responses to the concentrations of the AOC treatment. Significance was declared at p<0.05.

In vitro rumen GP, ammonia-N and MCP at 24 h incubation with substrates of CWR and CS are given in Table 4.

Effects of HMB and AOC on total VFA and molar proportion of each fatty acid of two forages are presented in Table 5.

Microbial populations (% of total bacterial 16S rDNA) of two forages are given in Table 6.

The pH, GP, ammonia-N and MCP are important parameters reflecting ruminal environment. In the current study, terminal pH of two roughage sources was not influenced by HMB and AOC and all pHs were within the normal range. Addition of AOC and HMB caused marked increases of GP, IVOMD, and MCP which meant the two additives promoted rumen fermentations. The responses to AOC were very similar to the results of Lee et al. (2004) who found the addition of anaerobic fungal culture caused a marked increase in GP, cellulose digestion and bacterial population. There are few reports of HMB on GP in the in vitro test. In addition, for CS there were HMB×AOC interactions in GP and IVOMD which meant that the two additives interacted with each other and showed that adding HMB and AOC simultaneously to CS had a significant effect on GP and IVOMD.

Ammonia concentration in the rumen is a balance between degradation of feed protein and uptake of ammonia for synthesis of microbial protein. Ammonia-N decreasing effect was only shown with HMB for CWR, which means HMB was more efficient than AOC in MCP production and a part of HMB seemed to be efficiently utilized as a nitrogen source. This finding is consistent with the result of Jouany et al. (1998), who reported that adding AOC had no effect on in vitro ammonia-N. The result with HMB agreed with the report of Vázquez-Añón et al. (2001), who found that addition of HMB in the feed improved microbial protein synthesis and reduced rumen ammonia-N. However, both AOC and HMB reduced ammonia-N and enhanced MCP for CS. In addition, a HMB×AOC interaction in MCP for the two roughages was observed. The reason for a different effect of AOC on ammonia-N of CWR and CS is not clear. This phenomenon was also further enhanced by a larger HMB×AOC effect on CS ammonia-N. The reason for this difference may be due to the high nitrogen content of CS.

VFA is the product of microbial fermentation of carbohydrates in the rumen; therefore, increased ruminal VFA concentrations are often assumed to be a result of this fermentation. From our study, total VFA of CWR and CS was increased by supplementing AOC and HMB. Frumholtz et al. (1989) reported the addition of AOC resulted in an increase of total VFA. Martin et al. (2013) showed HMB and isopropyl ester of HMB (HMBi) supplementation increased VFA concentrations in the rumen. However, Whelan et al. (2013) showed supplementary HMBi had no effect on rumen pH, VFA or ammonia-N in lactating dairy cows offered a low crude protein diet. Though both HMB and AOC increased total VFA, their effects on molar proportions of each VFA were different with the two roughage sources. That different fermentation patterns were shown by the two forages was probably due to their different chemical compositions. As shown in Table 1, CS had more protein, less organic matter, NDF and ADF than CWR which led to its easily being digested by microbes in the rumen. Noftsger et al. (2003) showed supplementation of HMB did not affect total VFA but observed differences in concentration of individual VFA which was different from our findings. In addition, AOC showed better effect on propionate feasibly due to enhanced soluble carbohydrate digestion by its α-amylase. This finding was inconsistent with Carton et al. (1993), who concluded supplemental AOC did not affect total VFA and VFA proportions. There were HMB×AOC interactions in total VFA for CS. In combination, the two additives had significant effects on total VFA.

In the current study, adding HMB and AOC had different effects on protozoa and fungi. Despite HMB supplementation significantly enhancing GP and total VFA of CWR, it did not increase the population of three cellulolytic species. An increase of F. succinogenes and R. albus was only observed with HMB for CS. Blake et al. (1986) suggested that HMB might stimulate a more metabolically active bacterial population. Martin et al. (2013) reported HMB supplementation had no effect on the total protozoa concentration, increased rumimal abundance of F. succinogenes and R. flavefaciens, but did not improve rumen fibrolytic activity. Supplementing AOC significantly increased the populations of fungi and F. succinogenes in the two roughage sources. In addition, AOC only significantly increased R. albus in CWR compared with CS because CWR had a greater amount of fiber substrate to act upon. It may be that these results are related to the NDF and/or starch proportion of the substrates. Although AOC does not produce the enzymatic machinery to completely depolymerize structural carbohydrates to simple sugars, it does produce enzymes that cause partial depolymerization (Boing, 1983) and it aids rumen cellulytic bacteria in completing the depolymerization of cellulosic material to simple sugars (Autrey et al., 1975). Newbold (1997) also presented data indicating that AOC stimulated the attachment of rumen microbes to plant fiber, which may explain how small quantities of fungi can have a significant effect on fiber degradation. In addition, there were interactions between HMB and AOC in fungi and R. flavefaciens of CWR and in R. albus of CS. The larger HMB×AOC effects on fungi and R. flavefaciens with AOC could partly be explained by Schmidt et al. (2004) who found that AOC accelerated the production and maturation of zoospores of Neocallimastix frontalis EB 188, along with an elevated production of protein and carboxymethy cellulase (CMCase). The mechanisms by which these supplements could impact microbial digestion in the rumen have not yet to be clarified. Nevertheless further research is warranted to understand their exact mode of action.