Evaluation of Different Yeast Species for Improving In vitro Fermentation of Cereal Straws

Abstract

Information on the effects of different yeast species on ruminal fermentation is limited. This experiment was conducted in a 3×4 factorial arrangement to explore and compare the effects of addition of three different live yeast species (Candida utilis 1314, Saccharomyces cerevisiae 1355, and Candida tropicalis 1254) at four doses (0, 0.25×10⁷, 0.50×10⁷, and 0.75×10⁷ colony-forming unit [cfu]) on in vitro gas production kinetics, fiber degradation, methane production and ruminal fermentation characteristics of maize stover, and rice straw by mixed rumen microorganisms in dairy cows. The maximum gas production (Vf), dry matter disappearance (IVDMD), neutral detergent fiber disappearance (IVNDFD), and methane production in C. utilis group were less (p<0.01) than other two live yeast supplemented groups. The inclusion of S. cerevisiae reduced (p<0.01) the concentrations of ammonia nitrogen (NH₃-N), isobutyrate, and isovalerate compared to the other two yeast groups. C. tropicalis addition generally enhanced (p<0.05) IVDMD and IVNDFD. The NH₃-N concentration and CH₄ production were increased (p<0.05) by the addition of S. cerevisiae and C. tropicalis compared with the control. Supplementation of three yeast species decreased (p<0.05) or numerically decreased the ratio of acetate to propionate. The current results indicate that C. tropicalis is more preferred as yeast culture supplements, and its optimal dose should be 0.25×10⁷ cfu/500 mg substrates in vitro.

INTRODUCTION

Supplementation with yeast products in the diets has become a common practice in improving the efficiency of feed utilization and the performance of ruminants for over 20 years (Moallem et al., 2009). It has been confirmed that yeast culture supplementation benefits digestion and metabolism of ruminants in several aspects, such as the improvement of nutrient digestibility, optimization of the proportion of volatile fatty acids (VFA) in the rumen, decrease in the ruminal ammonia nitrogen (NH₃-N), alleviation of pH fluctuation, and stimulation of ruminal microorganism population (Chaucheyras-Durand et al., 2008). Furthermore, it has been verified that yeast culture inclusion in the diets of ruminants can provide various growth factors, pro-vitamins and other stimulants to rumen microorganisms, and balance the ruminal fluid redox potential to create the optimal fermentation conditions for the rumen bacterial microflora (Jouany, 2001).

In the past few years, it had also been demonstrated that dietary live yeast supplementation plays a beneficial role in the improvement of ruminant’s productivity. Holtshausen and Beauchemin (2010) reported that live yeast (Levucell SC-1077) supplementation had a positive effect on milk yield and milk efficiency in cows fed a barley-based diet. In another study, dietary live yeast supplementation in dairy cows during the hot season in Israel improved the rumen environment by enhancing the ruminal pH and ammonia utilization, and in consequence improved dry matter (DM) intake, productivity and conversion efficiency of feeds (Moallem et al., 2009).

However, as observed in many studies so far, the effectiveness of dietary yeast products inclusion are variable, which might be ascribed to variation between animals, experimental diets fed, method of feeding, strains of yeasts, and their viability as well. For instance, supplementing beef cattle with Saccharomyces cerevisiae could raise the live weight by 7.5% depending on the type of diet tested, while improvement reached 13% in feedlot for diets rich in starch and sugars (Estefan, 1999). Supplementation of yeast culture improved the rate of gas production (GP), DM and organic matter disappearances for rice straw, wheat straw and maize stover (Tang et al., 2008). Meanwhile, dietary inclusion of S. cerevisiae NCYC 240, NCYC 1026 and Yea-Sacc stimulated total and cellulolytic bacterial numbers, while S. cerevisiae NCYC 694 and NCYC 1088 exerted no influence on the numbers of bacteria in vitro (Newbold et al., 1995).

Up to now, there is still little available information about the variation in the ruminal fermentation resulting from the addition of different live yeast species. Hence, the objectives of this study were to explore and compare the effects of three different species of yeasts (Candida utilis 1314, S. cerevisiae 1355, and Candida tropicalis 1254) on the in vitro ruminal fermentation characteristics of rice straw and maize stover by ruminal microorganisms from dairy cows, to further understand the mode of action of live yeast species in the rumen, and to provide more valuable information on the live yeast application to ruminants’ diets in practice.

MATERIALS AND METHODS

Crop straws, yeasts, and experimental design

Two types of crop straws which are most commonly used as roughage in diets for dairy cows in south China, i.e., maize stover (Zea mays, variety Kexiang Sweet Corn No. 1, Changsha, China) and rice straw (Oryza sativa, variety Xiang 125S/BAR-1, Changsha, China) were selected as in vitro fermentation substrates in this study. They were dried at 65°C for 24 h, and then ground through a 1 mm sieve and stored in plastic bags for assay. Maize stover and rice straw contained (DM basis): 52.3 and 62.4 g crude protein (CP)/kg, 636 and 632 g neutral detergent fiber (NDF)/kg, and 386 and 434 g acid detergent fiber (ADF)/kg, respectively.

Three different species of yeasts (C. utilis 1314, S. cerevisiae 1355, and C. tropicalis 1254) originally used as feed additives, were purchased from and reactivated by the China Center of Industrial Culture Collection. Yeasts were cultured and amplified using liquid malt extract medium (130 g malt extract and 0.1 g chloramphenicol/L), then their total viable numbers were counted in the form of colony-forming units (cfu) by employing the spread plate method. Afterwards, the yeasts were preserved at 4°C until the in vitro fermentation was started. The experiment was conducted in a 3×4 factorial arrangement, factors included yeast (three yeast species) and dose (0×10⁷cfu [without addition of yeast], 0.25×10⁷cfu, 0.50×10⁷cfu, and 0.75×10⁷cfu).

In vitro gas production and sampling

Culture solutions, i.e., macroelement solution, buffered solution and reducing solution used for in vitro fermentation were prepared to form artificial saliva according to the procedures modified by Tang et al. (2006). The artificial saliva was kept anaerobic by continuously pumping carbon dioxide for 2 h.

Rumen fluids were obtained from three rumen-cannulated Holstein dairy cows fed ad libitum a mixed diet of rice straw and concentrate (60:40, weight/weight) offered twice daily at 07:00 and 19:00 h. Concentrate contained (per 1,000 g DM): 396 g ground maize, 181 g soybean meal, 10 g CaHPO₄, 3 g limestone meal and 10 g premix. The rumen-cannulated Holstein dairy cows were managed according to the protocols approved by the Animal Care and Use Guidelines of the Animal Care Committee, Institute of Subtropical Agriculture, The Chinese Academy of Sciences, Changsha, China. Rumen contents of each dairy cow were obtained from various locations within the rumen immediately before the morning feeding, mixed and strained through four layers of cheesecloth under a continuous CO₂ stream. The obtained rumen fluids were then anaerobically combined with artificial saliva in the proportion of 1 to 9 at 39°C.

Samples of strawor stoverin an amount of 500±10 mg was accurately weighed into 100-mL fermentation bottles (Wanhong Glass Instrument Factory, Haimen, China) prewarmed at 39°C, then 50 mL of the mixed fluids (artificial saliva plus rumen fluids) were introduced into each bottle using a dispenser (Varispenser 4960000.060; Eppendorf, Wesseling-Berzdorf, Germany). After that, the yeasts were respectively added according to the above-mentioned different doses when the in vitro fermentation was started. Blanks containing only mixed fluids, mixed fluids and substrates, mixed fluids and different doses of yeasts were all incubated together with the treated bottles.

All fermentation bottles were connected with pressure sensors (CYG130-12; SQsensor, Kunshan, China) and incubated at 39°C. The pressure in all the bottles was recorded at 0, 1, 2, 4, 6, 12, 24, and 48 hours during the process of in vitro fermentation. Each time at 12, 24, and 48 h, three bottles for each treatment were respectively taken out from the incubator to stop the incubation. After termination of incubation, a 5 mL gas sample was collected into the vacuum flask (LabcoExetainer; Labco, High Wycombe, UK) with plastic syringe for CH4 determination, and then undegraded residues were immediately filtered through 2 layers of nylon cloth (40-um pore size). The incubation solutions of each treatment were sampled for determination of NH₃-N and VFA concentrations at 12, 24, and 48 h, respectively. In vitro fermentation was separately run three times on different days to result in nine analytical replicates (i.e., three analytical replicates per run).

Chemical analysis

The DM (method 930.15) and CP (6.25×N, method 990.03) were analyzed using the procedures of the Association of Official Analytical Chemists (AOAC, 1999). The NDF and ADF content were determined using a Fibretherm Fiber Analyzer (Gerhardt, Bonn, Germany) according to Van Soest et al. (1991) with addition of sodium sulphite and alpha-amylase in the NDF analysis. The filtered residues were dried at 105°C for 12 h and weighed for in vitro dry matter disappearance (IVDMD) determination. The NDF content in the dried residues was determined to calculate in vitro NDF disappearance (IVNDFD).

Two mL of incubation solution was centrifuged at 10,000×g at 4°C for 15 min, then 1.5 mL of supernatant solution was taken and 0.15 mL of metaphosphoric acid was added and homogenized. The mixed solution was centrifuged at 10,000×g at 4°C for 15 min again, and the supernatant solution was used to determine VFA content with a gas chromatograph (HP5890, Agilent 5890; Agilent Technologies, Palo Alto, CA, USA). A DB-FFAP column (30 m in length with a 0.25 mm inside diameter [i.d.]) was used for the separation. The attenuation was set at a nitrogen diffluent ratio of 1:50, hydrogen flow 30 mL/min, airflow 365 mL/min, injector temperature 250°C, column temperature 150°C, and detector temperature 220°C. The N₂ was used as carrier gas at a flow rate of 0.8 mL/min. The relative response factor, representing the peak of each VFA, was calculated using the standard VFA mixture, which was chromatographed with each group of 10 samples. Total molar concentration was calculated by taking the sum of individual VFA as 1.

For the determination of NH₃-N, 5 mL of incubation solution was centrifuged at 4,000×g and 4°C for 10 min, then 2 mL of the supernatant solution was taken and mixed with 8 mL 0.2 M HCl into a tube followed by homogenization. Subsequently, 0.4 mL of the mixed solution was taken and mixed with 2 mL of sodium nitroprusside solution (0.08 g sodium nitroprusside dissolved in 100 mL of 0.14 natrium salicylicum) and 2 mL of prepared solution (2 mL sodium hypochlorite solution mixed with 100 mL 0.3 M sodium hydroxide solution), then transferred into a tube followed by homogenization and equilibrated at room temperature for 10 min. The light absorption value was recorded at 700 nm using spectrophotometer (UV-2300; Shimadzu, Kyoto, Japan).

The CH₄ analysis was performed by gas chromatography (GC)-flame ionization detection using GC (GC7890A; Agilent Technologies, USA) equipped with a Hayesep Q packing column (2.44 M×1/8 in.×2.0 mm i.d.). The temperature of column and injector was respectively set at 60°C and 100°C for 3 min. The N₂ was used as carrier gas at a flow rate of 21 mL/min.

Calculation and statistical analysis

During the initial stages of this work, the correlation between the pressure in bottle and gas volume was measured at 39°C, and the regression equation was then established:

$$\text{y}=1.506\text{x\hspace{0.17em}}(\text{n}=20,{\text{r}}^2=0.999,\text{p}<0.0001)$$ (1)

Where y represents gas volume (mL), x is the pressure in bottle (kPa), 1.506 is a constant. The measured pressure was then converted to GP (mL). In vitro GP at 0, 1, 2, 4, 6, 12, 24, and 48 hours was fitted to Logistic-Exponential (Wang et al., 2011):

$$\text{GP}=\text{Vf\hspace{0.17em}}(1-\text{exp}(\text{d}-\text{t}\times \text{k}))/(1+\text{exp}(\text{b}-\text{k}\times \text{t}))$$ (2)

Where GP represents GP at t time, Vf means the maximum GP (mL), k represents GP fraction (/h), b and d represent the shapes of the GP curve. The time (t_0.5, h) when half of the maximum GP was achieved and the initial fractional rate of degradation (/h) were calculated by respectively employing the following two equations (Wang et al., 2011; Wang et al., 2013):

$${\text{t}}_{0.5}=\text{In}(\text{exp}(\text{b})+2\text{exp}(\text{d}))/\text{k}$$ (3)

$${\text{FRD}}_0=\text{k}/(1+\text{exp}(\text{b}))$$ (4)

GP, IVDMD, and IVNDFD were corrected by subtracting the values obtained for the blanks. Data were analyzed by two-way analysis of variance in the MIXED procedure of SAS (SAS Institute Inc., Cary, NC, USA) (SAS Institute Inc., 2001). For GP parameters, the model included species, dose, and species×dose as fixed effects. For pH, NH₃-N, CH₄ production, VFAs, IVDMD and IVNDFD, the fixed effects of species, dose, and species×dose were included in the model, with incubation time as a repeated effect. The bottle was used as the experimental unit, and run and bottle were considered as random effects in the entire study. Linear and quadratic effects of dose were analyzed using orthogonal polynomial contrasts. Cubic effects of dose were not examined for inexplicability in biology. Least squares means are reported throughout the text, and significance was declared at p<0.05.

RESULTS

In vitro gas production parameters

For maize stover, in vitro GP parameters generally were not affected by yeast species except for Vf, which was respectively 7% and 8% higher (p<0.01) for S. cerevisiae and C. tropicalis than the C. utilis (Table 1). All the parameters, except FRD₀, were influenced to a certain extent by yeast dose being dependent on yeast species. The C. utilis addition quadratically decreased (p<0.05) Vf, while linearly reduced (p<0.01) k. The addition of C. tropicalis showed a quadratic decreasing (p<0.05) effect on t_0.5. In comprehensive consideration of the effectiveness of improving GP and rate, the optimum supplemental dose of S. cerevisiae and C. tropicalis would be 0.25×10⁷ cfu and 0.75×10⁷ cfu, respectively.

Table 1.

Effects of different yeast species addition on in vitro gas production kinetics for maize stover

Item	Species	Dose (×107 colony-forming unit)					SEM2	Significance (>p)3
		Mean 1	0	0.25	0.50	0.75		Species	Dose	Species×dose
Vf (mL)	Candida utilis	67.97f	71.33a	65.50ab	64.25b	70.80a	2.013	<0.01	Q (<0.05)	NS
	Saccharomyces cerevisiae	72.85e	71.33	75.91	72.16	71.99			NS
	Candida tropicalis	73.12e	71.33	72.45	72.45	76.24			NS
	SEM 4	1.006
k (/h)	Candida utilis	0.087	0.105a	0.108a	0.091a	0.043b	0.0113	NS	L (<0.01)	NS
	Saccharomyces cerevisiae	0.010	0.105	0.010	0.096	0.098			NS
	Candida tropicalis	0.097	0.105	0.102	0.100	0.081			NS
	SEM	0.0056
FRD0 (/h)	Candida utilis	0.027	0.024	0.025	0.026	0.034	0.0038	NS	NS	NS
	Saccharomyces cerevisiae	0.023	0.024	0.026	0.021	0.021			NS
	Candida tropicalis	0.027	0.024	0.028	0.026	0.031			NS
	SEM	0.0019
t0.5 (h)	Candida utilis	17.10e	16.84	15.42	16.67	19.49	0.729	NS	NS	NS
	Saccharomyces cerevisiae	16.96ef	16.84	15.67	17.75	17.58			NS
	Candida tropicalis	15.91f	16.84	15.11	15.72	15.98			Q (<0.05)
	SEM	0.364

SEM, standard error of the mean; NS, not significant.

¹Mean for individual species across doses including the dose of 0.

²SEM for strain×dose.

³NS (p>0.05), L = linear effect of dose, Q = quadratic effect of dose.

⁴SEM for pooled mean of species including the dose of 0.

^{a, b}Means within a row for doses that do not have a common superscript differ (p<0.05).

^{e, f}Means within a column for species that do not have a common superscript differ (p<0.05).

For rice straw, the yeast species exerted significant effects (p<0.01) on Vf and FRD₀ (Table 2). Compared with the addition of C. utilis, Vf for S. cerevisiae and C. tropicalis were respectively 16% and 19% higher. Besides, FRD₀ was 43% higher for the C. tropicalis treatment than for the S. cerevisiae treatment. The dose effects of yeast addition on in vitro GP parameters were dependent on yeast species. The addition of C. utilis decreased k and t_0.5 in the same manner (linear, p<0.05), but increased FRD₀ (linear, p<0.05). Moreover, a quadratic (p<0.05) dose response to S. cerevisiae addition for Vf was positively observed. The profitable effects of C. tropicalis addition on FRD₀ (linear, p<0.01) and t_0.5 (linear, p<0.05) were also noted. The interactive effects of species and dose on Vf and t_0.5 were observed (p<0.05). Generally considering in vitro GP and rate, the optimum supplemental doses of S. cerevisiae and C. tropicalis might both be 0.25×10⁷ cfu.

Table 2.

Effects of different yeast species addition on in vitro gas production kinetics for rice straw

Item	Species	Dose (×107 colony-forming unit)					SEM2	Significance (>p)3
		Mean 1	0	0.25	0.50	0.75		Species	Dose	Species×dose
Vf (mL)	Candida utilis	53.23f	58.46	53.12	50.31	51.05	2.564	<0.01	NS	<0.05
	Saccharomyces cerevisiae	61.68e	58.46b	69.13a	60.30b	58.82b			Q (<0.05)
	Candida tropicalis	63.55e	58.46	67.48	62.86	65.41			NS
	SEM 4	1.282
k (/h)	Candida utilis	0.101ef	0.119a	0.110ab	0.100ab	0.074b	0.0122	NS	L (<0.05)	NS
	Saccharomyces cerevisiae	0.115e	0.119	0.103	0.121	0.117			NS
	Candida tropicalis	0.093f	0.119	0.078	0.099	0.077			NS
	SEM	0.0061
FRD0 (/h)	Candida utilis	0.018ef	0.013b	0.014b	0.020ab	0.024a	0.0023	<0.01	L (<0.05)	NS
	Saccharomyces cerevisiae	0.014f	0.013	0.016	0.014	0.013			NS
	Candida tropicalis	0.020e	0.013b	0.021a	0.021a	0.025a			L (<0.01)
	SEM	0.0011
t0.5 (h)	Candida utilis	19.31	19.98a	19.67a	18.25b	19.34ab	0.412	NS	L (<0.05)	<0.05
	Saccharomyces cerevisiae	19.36	19.98	19.20	18.65	19.63			NS
	Candida tropicalis	19.07	19.98a	20.07a	17.97b	18.26b			L (<0.05)
	SEM	0.206

SEM, standard error of the mean; NS, not significant.

¹Mean for individual species across doses including the dose of 0.

²SEM for strain×dose.

³NS (p>0.05), L = linear effect of dose, Q = quadratic effect of dose.

⁴SEM for pooled mean of species including the dose of 0.

^{a, b}Means within a row for doses that do not have a common superscript differ (p<0.05).

^{e, f}Means within a column for species that do not have a common superscript differ (p<0.05).

In vitro dry matter and neutral detergent fiber disappearance

For maize stover, the yeast species affected IVDMD and IVNDFD (p<0.01) (Table 3). The lowest IVDMD and IVNDFD were observed in the C. utilis treatment, the former was 5% and 13% less, while the latter was 11% and 21% lower, when compared with S. cerevisiae and C. tropicalis treatment, respectively. A linear decrease (p<0.01) both in IVDMD and in IVNDFD was noted in response to the increase of C. utilis addition, while the C. tropicalis addition linearly increased IVDMD and IVNDFD (p<0.05), and the maximum IVDMD and IVNDFD were both achieved at the supplemental dose of 0.75×10⁷ cfu, which were respectively 12% and 22% greater than the control.

Table 3.

Effects of different yeast species addition on IVDMD and IVNDFD of maize stover and rice straw

Item	Species	Dose (×107 colony-forming unit)					SEM2	Significance (>p)3
		Mean 1	0	0.25	0.50	0.75		Species	Dose	Species×dose
Maize stover
IVDMD (mg/mg)	Candida utilis	0.499g	0.531a	0.505ab	0.483b	0.478b	0.0002	<0.01	L (<0.01)	<0.05
	Saccharomyces cerevisiae	0.526f	0.531	0.529	0.497	0.548			NS
	Candida tropicalis	0.574e	0.531b	0.589a	0.580a	0.597a			L (<0.05)
	SEM 4	0.0001
IVNDFD (mg/mg)	Candida utilis	0.380g	0.424a	0.356b	0.380ab	0.359b	0.0002	<0.01	L (<0.01)	<0.01
	Saccharomyces cerevisiae	0.428f	0.424	0.431	0.426	0.430			NS
	Candida tropicalis	0.484e	0.424b	0.500a	0.493a	0.518a			L (<0.05)
	SEM	0.0001
Rice straw
IVDMD (mg/mg)	Candida utilis	0.442g	0.482a	0.453ab	0.424bc	0.408c	0.0001	<0.01	L (<0.01)	<0.01
	Saccharomyces cerevisiae	0.475f	0.482	0.471	0.472	0.473			NS
	Candida tropicalis	0.509e	0.482	0.527	0.523	0.503			NS
	SEM	0.0001
IVNDFD (mg/mg)	Candida utilis	0.339f	0.402a	0.347b	0.298b	0.310b	0.0002	<0.01	Q (<0.01)	<0.01
	Saccharomyces cerevisiae	0.400e	0.402	0.398	0.395	0.405			NS
	Candida tropicalis	0.423e	0.402b	0.467a	0.464a	0.359b			Q (<0.01)
	SEM	0.0001

IVDMD, in vitro dry matter disappearance; IVNDFD, in vitro neutral detergent fiber; SEM, standard error of the mean; NS, not significant.

¹Mean for individual species across doses including the dose of 0.

²SEM for strain×dose.

³NS (p>0.05); L = linear effect of does; Q = quadratic effect of dose.

⁴SEM for pooled mean of species including the dose of 0.

^a–cMeans within a row for doses that do not have a common superscript differ (p<0.05).

^e–gMeans within a column for species that do not have a common superscript differ (p<0.05).

As for rice straw, IVDMD and IVNDFD were influenced (p<0.01) by the yeast species, and the least IVDMD and IVNDFD were observed in the C. utilis treatment, the former was 7% and 13% less, while the latter was 15% and 20% less than those of S. cerevisiae and C. tropicalis treatment, respectively. The C. utilis supplementation decreased IVDMD (linear, p<0.01) and IVNDFD (quadratic, p<0.01). A quadratic (p<0.01) response in IVNDFD to C. tropicalis addition was noted, and the greatest IVNDFD occurred at the supplemental dose of 0.25×10⁷ cfu, which was improved by 16% compared with the control.

pH, NH₃-N and CH₄ production

For maize stover, yeast species affected (p<0.01) the NH₃-N concentration of in vitro fermentation liquors and CH₄ production/g IVDMD (Table 4). The NH₃-N concentration of the S. cerevisiae treatment was respectively 6% and 8% less than those of the C. utilis and C. tropicalis treatments, while the CH₄ production for C. utilis addition was respectively 12% and 10% lower than the addition of S. cerevisiae and C. tropicalis. A linear reduction (p<0.05) in pH value was noted for both the C. utilis and C. tropicalis treatments, while a quadratic response and a linear increase in the concentration of NH₃-N were observed for the S. cerevisiae (p<0.05) and C. tropicalis (p<0.01) treatments, respectively. Besides, the addition of S. cerevisiae and C. tropicalis caused overall increases in CH₄ production, which were quadratic (p<0.01) and linear (p<0.01), respectively. Moreover, the greatest CH₄ production in those two yeast treatments were both achieved at the supplemental dose of 0.50×10⁷cfu, which were respectively 33% and 26% greater than the control. There was an interactive effect (p<0.05) of species and dose on the NH₃-N concentration.

Table 4.

Effects of different yeast species addition on pH, NH₃-N concentration and CH₄ production of maize stover and rice straw in vitro

	Item	Species	Dose (×107 colony-forming unit)					SEM2	Significance (>p)3
			Mean 1	0	0.25	0.50	0.75		Species	Dose	Species ×dose
Maize stover	pH	Candida utilis	6.46	6.50ab	6.51a	6.43ab	6.41b	0.033	NS	L (<0.05)	NS
		Saccharomyces cerevisiae	6.47	6.50	6.49	6.46	6.45			NS
		Candida tropicalis	6.44	6.50a	6.44ab	6.41ab	6.40b			L (<0.05)
		SEM4	0.017
	NH3-N (mg/dL)	Candida utilis	7.57e	7.30	7.92	7.41	7.65	0.269	<0.01	NS	<0.05
		Saccharomyces cerevisiae	7.14f	7.30ab	6.68b	7.10ab	7.50a			Q (<0.05)
		Candida tropicalis	7.77e	7.30b	7.25b	7.70b	8.83a			L (<0.01)
		SEM	0.135
	CH4 (mmol/g) IVDMD	Candida utilis	0.44f	0.43	0.38	0.49	0.46	0.024	<0.01	NS	NS
		Saccharomyces cerevisiae	0.50e	0.43c	0.53ab	0.57a	0.48bc			Q (<0.01)
		Candida tropicalis	0.49e	0.43b	0.49ab	0.54a	0.52a			L (<0.01)
		SEM	0.013
Rice straw	pH	Candida utilis	6.49	6.57a	6.51b	6.46bc	6.42c	0.019	NS	L (<0.01)	<0.05
		Saccharomyces cerevisiae	6.50	6.57a	6.49b	6.48b	6.46b			Q (<0.05)
		Candida tropicalis	6.50	6.57a	6.48bc	6.44c	6.52ab			Q (<0.01)
		SEM	0.010
	NH3-N (mg/dL)	Candida utilis	6.30f	6.33	5.91	6.64	6.30	0.276	<0.01	NS	<0.01
		Saccharomyces cerevisiae	5.87g	6.33a	5.73ab	5.24b	6.16a			Q (<0.01)
		Candida tropicalis	7.14e	6.33b	7.14a	7.58a	7.51a			L (<0.01)
		SEM	0.138
	CH4 (mmol/g) IVDMD	Candida utilis	0.35f	0.34b	0.28c	0.36b	0.41a	0.018	<0.01	Q (<0.01)	<0.01
		Saccharomyces cerevisiae	0.38e	0.34b	0.40a	0.40a	0.39ab			Q (< 0.05)
		Candida tropicalis	0.39e	0.34b	0.40a	0.41a	0.42a			L (< 0.01)
		SEM	0.009

SEM, standard error of the mean; NS, not significant.

¹Mean for individual species across doses including the dose of 0.

²SEM for strain×dose.

³NS (p>0.05); L = linear effect of dose, Q = quadratic effect of dose.

⁴SEM for pooled mean of species including the dose of 0.

^a–cMeans within a row for doses that do not have a common superscript differ (p<0.05).

^e–gMeans within a column for species that do not have a common superscript differ (p<0.05).

As regards rice straw, the yeast species influenced the NH₃-N concentration of fermentation liquors (p<0.01) and CH₄ production/g IVDMD (p<0.01). The NH₃-N concentration in the S. cerevisiae treatment was respectively 7% and 18% lower than those of C. utilis and C. tropicalis treatment, while CH₄ production supplemented with C. utilis was 8% less than that of S. cerevisiae, and 10% less than that of C. tropicalis. The pH value decreased in response to the addition of C. utilis (linear, p<0.01), S. cerevisiae (quadratic, p<0.05), and C. tropicalis (quadratic, p<0.01). As for the concentration of NH₃-N, a quadratic reduction and a linear increase (p<0.01) were observed with the increasing doses of S. cerevisiae and C. tropicalis, respectively. The CH₄ production was increased by the addition of C. utilis (quadratic, p<0.01), S. cerevisiae (quadratic, p<0.05), and C. tropicalis (linear, p<0.01). Addition of C. utilis at the dose of 0.25×10⁷cfu decreased CH₄ production by 18% compared with the control. There were interactive effects of species and dose on pH (p<0.05), NH₃-N (p<0.01) and CH₄ production/g IVDMD (p<0.01).

Volatile fatty acid

The yeast species influenced (p<0.01) the concentrations of isobutyrate, butyrate, isovalerate, and ratio of acetate to propionate (A:P) in incubation fluids when maize stover was used as substrate (Table 5). For the addition of S. cerevisiae, the concentration of isobutyrate was 10% and 16% lower, while the concentration of butyrate was 20% and 12% higher, when compared to those of C. utilis and C. tropicalis treatments respectively. The maximum concentration of isovalerate and A:P were observed in C. tropicalis treatment, which were respectively 20% and 37%, and 9% and 6% greater than those of C. utilis and S. cerevisiae treatments. The addition of C. utilis linearly decreased (p<0.01) A:P, and it obtained the numerical greatest concentrations of acetate, propionate, isovalerate, and total volatile fatty acids (TVFA) at the dose of 0.50×10⁷cfu, which were respectively 25%, 26%, 18%, and 27% greater than the control. The S. cerevisiae addition increased the propionate concentration (linear, p<0.05), but decreased the concentrations of isobutyrate (quadratic, p<0.05), isovalerate (quadratic, p<0.05), valerate (linear, p<0.05), and A:P (linear, p<0.05). Additionally, the C. tropicalis addition linearly increased (p<0.01) the concentration of isovalerate, reaching the maximum which was 36% greater than that of the control at the dose of 0.75×10⁷ cfu. The interactive effects of species and dose on the concentrations of acetate (p<0.05), butyrate (p<0.05), isovalerate (p<0.01), TVFA (p<0.05), and A:P (p<0.05) were noted respectively.

Table 5.

Effects of different yeast species addition on volatile fatty acids concentration of in vitro incubation fluids for maize stover

Item 1	Species	Dose (×107 colony-forming unit)					SEM3	Significance (>p)4
		Mean 2	0	0.25	0.50	0.75		Species	Dose	Species×dose
Acetate (mM)	Candida utilis	18.08	17.97	16.56	22.39	15.40	1.260	NS	NS	<0.05
	Saccharomyces cerevisiae	18.55	17.97	17.36	18.91	19.96			NS
	Candida tropicalis	18.79	17.97	19.43	18.03	19.75			NS
	SEM 5	0.630
Propionate (mM)	Candida utilis	8.51	8.07	7.90	10.20	7.86	0.561	NS	NS	NS
	Saccharomyces cerevisiae	8.79	8.07b	8.44ab	8.95ab	9.72a			L (<0.05)
	Candida tropicalis	8.38	8.07	8.48	8.20	8.77			NS
	SEM	0.280
Isobutyrate (mM)	Candida utilis	0.30e	0.31	0.26	0.33	0.29	0.020	<0.01	NS	NS
	Saccharomyces cerevisiae	0.27f	0.31a	0.24b	0.25b	0.27ab			Q (<0.05)
	Candida tropicalis	0.32e	0.31	0.32	0.32	0.34			NS
	SEM	0.010
Butyrate (mM)	Candida utilis	1.73f	1.83	1.52	2.18	1.39	0.150	<0.01	NS	<0.05
	Saccharomyces cerevisiae	2.08e	1.83	2.14	2.18	2.18			NS
	Candida tropicalis	1.85f	1.83	1.93	1.73	1.91			NS
	SEM	0.075
Isovalerate (mM)	Candida utilis	0.40f	0.39	0.35	0.46	0.40	0.022	<0.01	NS	<0.01
	Saccharomyces cerevisiae	0.35g	0.39a	0.32b	0.33ab	0.35ab			Q (<0.05)
	Candida tropicalis	0.48e	0.39b	0.50a	0.50a	0.53a			L (<0.01)
	SEM	0.011
Valerate (mM)	Candida utilis	0.46	0.56	0.36	0.48	0.46	0.069	NS	NS	NS
	Saccharomyces cerevisiae	0.37	0.56a	0.30b	0.30b	0.32b			L (<0.05)
	Candida tropicalis	0.41	0.56	0.34	0.35	0.38			NS
	SEM	0.034
A:P	Candida utilis	2.03f	2.21a	2.04b	1.95bc	1.90c	0.048	<0.01	L (<0.01)	<0.05
	Saccharomyces cerevisiae	2.09f	2.21a	2.05b	2.09ab	2.02b			L (<0.05)
	Candida tropicalis	2.21e	2.21	2.27	2.14	2.23			NS
	SEM	0.024
TVFA (mM)	Candida utilis	29.70	29.11	26.95	36.87	25.84	1.903	NS	NS	<0.05
	Saccharomyces cerevisiae	30.40	29.11	28.78	30.92	32.79			NS
	Candida tropicalis	30.22	29.11	31.00	29.12	31.67			NS
	SEM	0.962

SEM, standard error of the mean; NS, not significant.

¹TVFA = total short chain fatty acids, A:P = ratio of acetate to propionate.

²Mean for individual species across doses including the dose of 0.

³SEM for strain×dose.

⁴NS (p>0.05), L = linear effect of dose, Q = quadratic effect of dose.

⁵SEM for pooled mean of species including the dose of 0.

^a–cMeans within a row for doses that do not have a common superscript differ (p<0.05).

^e–gMeans within a column for species that do not have a common superscript differ (p<0.05).

Regarding to rice straw, the concentrations of isobutyrate, isovalerate, and valerate of incubation fluids were affected (p<0.01) by yeast species (Table 6). For the S. cerevisiae addition, the isobutyrate concentration was respectively 15% and 18% less than the addition of C. utilis and C. tropicalis, while the isovalerate concentration in the C. tropicalis treatment was 19% and 39% higher compared to C. utilis and S. cerevisiae. In addition, the valerate and 23% greater than that in the S. cerevisiae and C. tropicalis treatments. The C. utilis addition linearly increased (p<0.05) the concentration of isobutyrate and reached the peak which was 16% greater than that of the control, while the concentrations of isovalerate and valerate were both quadratically reduced (p<0.05) by adding S. cerevisiae. The C. tropicalis addition linearly increased the concentrations of isobutyrate (p<0.05) and isovalerate (p<0.01) reaching the maximum at the dose of 0.75×10⁷cfu, but quadratically decreased (p<0.05) the valerate concentration. Besides, the A:P reached the numerical minimum with the addition of C. utilis and S. cerevisiae both at the doses of 0.25×10⁷cfu, while it was obtained at the dose of 0.50×10⁷cfu in the C. tropicalis treatment, which were respectively 14%, 9%, and 11% less than that of the control. The interactive actions of species and dose on the concentrations of isobutyrate (p<0.05) and concentration in the C. utilis treatment was respectively 41% isovelerate (p<0.01) were also observed.

DISCUSSION

As a matter of fact, the effectiveness of yeast addition on in vitro GP parameters is somehow inconsistent in some previous studies. Mutsvangwa et al. (1992) reported that in vitro GP of a barley diet for beef cattle supplemented with yeast culture (Yea-Sacc1026) was on average less than that in the control, while Tang et al. (2008) found that supplementation of yeast culture (Original XP; Diamond V Mills Inc., Cedar Rapids, IA, USA) increased the cumulative GP, theoretical maximum of GP and the rate of GP of low quality roughages. This disparity might be caused by the difference in the yeast species used in their studies, fermentation substrate and experimental conditions. In the present study, adding C. utilis at all the designated doses decreased in vitro GP compared to the control, which was in agreement with the results obtained by Mutsvangwa et al. (1992). Meanwhile, maize stover or rice straw supplemented with S. cerevisiae and C. tropicalis achieved greater GP than that supplemented with C. utilis, suggesting that the selection of yeast species should be taken into consideration when live yeast was applied to improve in vitro fermentation efficiency of forages. Indexes of FRD₀ and T_0.5 usually reflect the rate of degradation at early incubation stages of ‘<12 h’ and the incubation time of reaching half of the maximum GP, respectively. In general, the FRD₀ is inversely proportional to t_0.5. The addition of C. utilis and C. tropicalis decreased t_0.5 but increased FRD₀ of rice straw fermentation, indicating that the rate of degradation would be faster at the early stage of in vitro fermentation. Moreover, the two reverse responses of t_0.5 in S. cerevisiae and C. tropicalis treatments for maize stover indicated that the influence on the rate of degradation would be dependent on the yeast species, but this hypothesis required further research to be conducted. The alteration in the rate of degradation in response to yeast culture addition has also been verified in some previous studies. For instance, Newbold et al. (1995) suggested that S. cerevisiae culture stimulated the rate rather than the extent of degradation by ruminal micro-organisms. Sullivan and Martin (1999) found S. cerevisiae culture filtrate stimulated the initial rate of cellulose degradation. In addition, the decrease of Vf and k caused by C. utilis addition in comparison with control suggested that this yeast species might not be suitable for dietary supplement.

It was noted in the study that the increase or decrease of IVDMD and IVNDFD of maize stover and rice straw was depended upon different yeast species, as C. utilis reduced both IVDMD and IVNDFD while C. tropicalis improved IVDMD and IVNDFD, and S. cerevisiae did not affect IVDMD and IVNDFD being dose-dependent. Furthermore, we found that the two higher supplemental doses of C. utilis did not always ensure the higher IVDMD and IVNDFD compared to the dose of 0.25×10⁷cfu/500 mg, which was similar to the findings of Tang et al. (2008). In the study of Tang et al. (2008), the greatest values of IVDMD occurred for maize stover, maize stover silage, and wheat straw when yeast culture was supplemented at the level of 5.0 g/kg rather than the higher level of 7.5 g/kg. However, this phenomenon lacked sufficient explanation and needs to be fully studied in further research. Considering both IVDMD and IVNDFD could be more closely related to in vivo conditions, it is suggested that C. tropicalis should be more appropriate for supplements at the dose of 0.25×10⁷cfu/500 mg substrates.

As pH value is a main index reflecting the internal homeostasis of rumen environment, therefore maintaining a relatively stable ruminal pH is vital to assuring efficient rumen fermentation. Ruminants usually possess highly developed systems to maintain ruminal pH within a physiological range of about 5.5 to 7.0 (Krause and Oetzel, 2006). In this study, although adding C. utilis, S. cerevisiae, and C. tropicalis, respectively lowered pH value to different extents, whilst pH value across all treatments ranged from 6.40 to 6.57, which still kept a suitable condition for fermentation, growth of microorganism, and fiber degradation in the rumen (Stewart et al., 1997). Satter and Slyter (1974) suggested that the lowest NH₃-N concentration of rumen liquor should not be less than 5 mg/dL to maintain the higher growth rate of bacteria. Deficiency of NH₃-N restricts the microbial protein synthesis, while an overly high NH₃-N concentration also inhibits the microbial utilization of NH₃-N (Hristov et al., 2002). Concentration of NH₃-N across three yeast treatments ranged from 5.24 to 8.83 mg/dL in this study, indicating that the growth and protein synthesis of microorganisms was not restricted. Fadel Elseed et al. (2007) reported yeast (S. cerevisiae) supplementation resulted in a numerical increase in ammonia-N concentration in rumen fluid of Nubian goat’s kids. Similarly, the inclusion of C. utilis and C. tropicalis could enhance NH₃-N concentration to different extents with maize stover as fermented substrate, while for rice straw, only C. tropicalis addition elevated NH₃-N concentration of incubation fluids.

Methanogenesis is an essential metabolic pathway for hydrogen elimination and subsequently for efficient degradation of plant cell wall carbohydrates in the rumen (Wolin et al., 1997). Since yeast, especially S. cerevisiae, is the most frequently used direct-fed microbial in ruminant production, its influence on methanogenesis has been investigated in a few studies both in vitro and in vivo, but the results of these studies are inconsistent. In present study, the addition of S. cerevisiae and C. tropicalis respectively increased CH₄ production/g IVDMD of crop straws. The elevation of CH₄ production might be due to the increased disappearance of fiber under the in vitro closed anaerobic environment. Qiao and Shan (2006) found that addition of S. cerevisiae and Saccharomycopsis fibuligera also increased in vitro methane production, while C. tropicalis addition decreased CH₄ production with cornstarch, soybean, and wheat bran plus concentrate (3:1) mixture as the fermented substrates. It was inferred that the reduced CH₄ production might be attributed to the suppressed methanogens in the rumen, while the enhanced CH₄ production could be caused by the stimulated methanogens. Nevertheless, the inconsistency in results from different studies necessitates further research on this topic.

A number of trials have been conducted to examine the effects of yeast culture supplements on VFA in the rumen. Dawson et al. (1990) reported that the VFA patterns were not altered by yeast supplement (S. cerevisiae) in either rumen-simulating cultures or in the rumens of steers, while Mutsvangwa et al. (1992) found the addition of yeast culture (Yea-Sacc1026) increased the concentration of acetate and TVFA both in vitro and in vivo. In this study, not only Vf, k, IVDMD, and IVNDFD, but also the production of VFA was decreased by supplementing C. utilis, which indicated that C. utilis might be unsuitable as an additive for enhancing in vitro fermentation of cereal straws. Besides, supplementing S. cerevisiae and C. tropicalis elevated CH₄ production without significantly increasing VFA concentrations could be regarded as a disadvantage for in vitro fermentation of cereal straws. In addition, it was found that S. cerevisiae and C. tropicalis addition increased the propionate concentration, but decreased the concentrations of isovalerate and valeratein vitro with the increasing dose with maize stover and rice straw as fermented substrates. The decline of isovalerate and valerate concentrations suggested that S. cerevisiae and C. tropicalis addition have potential to stimulate plant cell wall digestion and ammonia utilization by mixed ruminal bacteria, as it was found that cell wall digestion and ammonia utilization were increased by low concentrations of isovalerate and valerate (Gorosito et al., 1985). Additionally, our results showed that the addition of these three yeasts decreased or numerically decreased A:P, this was in agreement with the in vitro finding of Martin et al. (1989). In contrast, Mutsvangwa et al. (1992) pointed out that yeast culture addition did not alter A:P in vitro and in vivo, whereas Arambel et al. (1987) reported that A:P in the in vitro rumen fermentation supplemented with a yeast culture increased. This variation could be attributed to different species or strains of yeast used in different studies or the distinction between live yeast and yeast culture, and it needs to be investigated via further experiments.

CONCLUSION

In conclusion, both S. cerevisiae and C. tropicalis are more desirable than C. utilis as yeast culture supplements. Further, C. tropicalis is preferred compared to S. cerevisiae and its optimal dose should be 0.25×10⁷ cfu/500 mg substrates, as C. tropicalis enhanced IVDMD and IVNDFD, and digestibility was viewed as the most fundamental parameter reflecting in vitro fermentation of cereal straws. The finding obtained from this study provides the dairy farmers with a practicable reference on the selection of live yeast species as feed additives. However, the verification of effects on these three live yeast additives in vivo, the evaluation of more live yeast additives for improving rumen fermentation, and the explanation of mechanisms of live yeast additives in ruminants all require further research in future.