DOI: 10.1111/jpn.

12508

ORIGINAL ARTICLE

Effects of a blend of garlic oil, nitrate and fumarate on in vitro

ruminal fermentation and microbial population
D. T. Mbiriri1, S. Cho1,2, C. I. Mamvura1 and N. J. Choi1
1 Department of Animal Science, Chonbuk National University, Jeonju, Korea, and
2 Institute of CALS, CALS Co., Ltd., Gyeonggi, Korea

Summary
Although garlic oil and nitrate can effectively suppress ruminal methane (CH4) production in vitro, the appli-
cation of these compounds is associated with suppressed total volatile fatty acid (VFA) concentration. On the
other hand, the effectiveness of fumarate as a ruminal CH4 mitigating agent is variable but its application
increases total VFA concentration. We therefore hypothesized that the different characteristics of the com-
pounds can compensate for the shortcomings of the other. The objective of this study was to develop an opti-
mal blend of garlic oil, nitrate and fumarate that can suppress in vitro ruminal CH4 without affecting total
VFA concentration. Three ruminal in vitro fermentation experiments were carried out. The first one, a one
factor at a time experiment was employed to investigate the effective concentration of each of the compounds
on CH4 and VFA production by ruminal bacteria. We then applied the fractional factorial design and response
surface methodology in the second experiment to determine optimal concentrations of the compounds in the
blend. The optimal blending of garlic oil, fumarate and nitrate was determined to be 50 mg/l, 15 mM and
20 mM, respectively. This simulated optimal blend was verified in a 48 h in vitro batch fermentation experi-
ment. The blend achieved the intended goal of suppressing CH4 whilst maintaining total VFA concentration.
The blend and nitrate suppressed archaea populations (p < 0.001) but did not affect the total microbial popu-
lation (p = 0.945). The observed results could be explained by additive effects of the agents making up the
blend. Supplementing a high concentrate diet with the blend can significantly decrease ruminal CH4 and
maintain total VFA in vitro. These findings however, need to be verified in vivo using the optimized ratio of
combining the three methane inhibitors as a guide.
Keywords antimethanogen, enteric methane, rumen methanogenic archaea, volatile fatty acid

Correspondence Nag-Jin Choi, Department of Animal Science, Chonbuk National University, Jeonju, 561-756, Republic of Korea;
Tel: +82 63 270 2579; Fax: +82 63 270 2612; E-mail:nagjin@jbnu.ac.kr

D. T. Mbiriri and S. Cho contributed equally to this work.

Received: 28 January 2015; accepted: 27 February 2016

posed inhibitor dietary supplements range from

Introduction
Most of the enteric methane (CH4) produced by
ruminants originates from the rumen (Martin et al.,
2010). It is a normal waste product of rumen micro-
bial fermentation. Ruminal methanogenic archaea
combine CO2 with H2 generated from feed fermenta-
tion, resulting in CH4 formation. Removal of hydro-
gen from the rumen environment is essential for the
continued microbial breakdown of organic matter.
CH4 emitted into the atmosphere is, however, an
environmental concern. An effective enteric methane
inhibitor has to suppress CH4 emissions without neg-
atively affecting ruminal metabolic processes. Pro-
compounds of plant origin, inorganic and organic
substances, alternative H2 sinks to antibiotics
(Kobayashi, 2010). Each of these solutions however,
falls short in some way. We hypothesized that com-
bining certain anti-methanogens can improve their
CH4 inhibitory effect without negatively affecting
metabolic processes in the rumen.
Garlic oil is an essential oil that has been widely
applied as an effective CH4 mitigation agent in vitro
(Busquet et al., 2005a). Its use has, however, been
associated with suppressed nutrient digestibility
(Anassori et al., 2012). Fumarate is an inorganic com-
pound that suppresses enteric methanogenesis

Journal of Animal Physiology and Animal Nutrition © 2016 Blackwell Verlag GmbH 1
Optimized blend of methane inhibitors
through redirecting metabolic H2 from methane pro-
duction towards its reduction to succinate and finally
to propionate (Ungerfeld et al., 2007). When supple-
mented in diets in vitro (Lopez et al., 1999) and in vivo
(Isobe and Shibata, 1993), fumarate can enhance
nutrient digestibility. However, there are variable
results on the effectiveness of fumarate in abating
enteric CH4 (Van Zijderveld et al., 2010a). Nitrate is
effective at inhibiting methanogenesis both in vitro
(Zhou et al., 2012) and in vivo (Van Zijderveld et al.,
2010a). Its drawbacks are suppressing IVDMD (Marais
et al., 1988) and VFA production (Farra and Satter,
1971). To date, most in vivo feeding trials have applied
nitrate at a concentration ranging from 10 to 26 g/kg
DM of feed with consistent enteric CH4 reduction but
varying effects on DM intake and VFA concentration
(Bruning-Fann and Kaneene, 1993; Van Zijderveld
et al., 2010b; Hulshof et al., 2012).
The objective of this study was to formulate a blend
made up of an optimal mixture of garlic oil, fumarate
and nitrate that would effectively suppress CH4 pro-
duction without negatively affecting total VFA con-
centration in vitro.
Materials and methods
Treatments and experimental design
Three in vitro batch experiments were conducted.
Experiments 1 and 2 lasted 24 h whereas Exp. 3 lasted
48 h. In Exp. 1, six treatments; garlic oil (GO),
300 mg/l; fumarate (FM), 21 mM; nitrate (NT),
21 mM; combination of garlic oil and fumarate (GOF)
(GO, 300 mg/l + FM, 21 mM); combination of garlic oil
and nitrate (GON) (GO, 300 mg/l + NT, 21 mM) and a
control (no additive) were applied in a completely ran-
domized design. The purity and sources of GO, FM and
NT used were, garlic oil (artificial; Sigma-Aldrich, St.
Louis, MO, USA), fumaric acid (>98%; Sigma-Aldrich,
St. Louis, MO, USA) and sodium nitrate (99%; Dae-
jung Chemicals and Metals, Siheung, Korea).
In Exp. 2, GO, FM and NT were used as variables.
Each variable was applied at three different inclusion
levels to represent low, medium and high concentra-
tion; GO, 30, 165 and 300 mg/l; FM, 7, 14 and
21 mM; and NT, 7, 14 and 21 mM. The Box Behnken
experimental design (Box and Behnken, 1960) with a
total of 15 runs was employed.
In Exp. 3, five treatments were formulated to verify
effect of the optimized blend; GO, 50 mg/l; GONF
(GO, 50 mg/l + NT, 20 mM + FM, 15 mM); NT, 20 mM;
FM, 15 mM; and a control (no additive). Fermentation
parameters were determined at 0, 3, 6, 9, 12, 24 and
48 h.
2 Journal of Animal Physiology and Animal Nutrition © 2016 Blackwell Verlag GmbH
D. T. Mbiriri et al.
Incubation procedure
Rumen fluid was collected from two Hanwoo steers
(Korean native beef breed), averaging 350
50 kg,
fitted with rumen cannulae. Each steer was fed a
50:50 ration of hay and a commercial beef grower con-
centrate twice a day, at 0900 and 1700 h. The rumen
fluid was collected before the morning feeding. The
fluid was strained through four layers of cheese cloth
and transported to the laboratory in a thermos flask.
The rumen fluid was mixed at a ratio of 1:4 with
McDougall solution (pH 6.9) (McDougall, 1948). Fifty
millilitres of the diluted rumen fluid were then dis-
pensed into each serum bottle under a stream of CO2
to maintain an anaerobic environment. Each serum
bottle in Exp. 1 contained 0.5 g of orchard grass milled
through a 1 mm sieve. In experiments 2 and 3, 0.5 g
of high concentrate diet (orchard grass:concentrate;
20:80) were used. The nutrient content of orchard
grass and the commercial beef grower concentrate are
shown in Table 1. Still under a stream of CO2,
treatments were then applied into respective serum
bottles. Treatments were replicated in triplicate in all
the experiments. Bottles were closed with rubber
stoppers, sealed with aluminum caps and incubated at
39 °C.
Chemical analyses
At the end of each incubation period (except for 0 h
in Exp. 3), total biogas was measured by way of dis-
placing a glass syringe. The collected gas was stored in
aluminium vacutainers for analysis of CH4 and H2.
The analyses of these gases was carried out using a gas
chromatograph (GC-7890A, Agilent Technologies,
2010, 2012 Santa Clara, CA, USA) fitted with a Ther-
mal Conductivity Detector and a capillary column
(NukolTM Fused silica capillary column 30 m 9 0.25
mm 9 0.25 lm film thickness, Supelco, Bellefonte,
PA, USA) according to slightly modified procedures of
Lopez et al. (1999).
Table 1 Chemical and nutrient composition of feed stuffs used in the
study
Commercial beef
Item Orchard grass grower concentrate
Dry matter 92.92 95.46
Crude protein, % DM 7.77 13.88
Ether extract, % DM 1.76 4.23
Crude fibre, % DM 31.21 10.76
Organic matter, % DM 92.71 93.30
D. T. Mbiriri et al.
The fermentation fluid was centrifuged at 3000 9 g
for 20 min at 4 °C and 1.5 ml samples of the super-
natant were stored at 20 °C for VFA and ammonia
nitrogen (N) analyses. Before analysing for VFAs,
1 ml of each sample was pre-treated with 0.2 ml of
25% metaphosphoric acid, allowed to settle for
30 min before centrifuging at 12 300 9 g using a
table-top centrifuge (Gyrozen mini, Seoul, Korea),
then transferred into vials for analysis. VFA were anal-
ysed using a gas chromatograph (GC-7890A, Agilent
Technologies, 2010, 2012) fitted with a Flame Ionizing
Detector and a capillary column (CarbonexTM Fused
silica capillary column 30 m 9 0.25 mm 9 0.25 lm
film thickness, Supelco, Bellefonte, USA) as outlined
by Erwin et al. (1961).
The pH of the fermentation fluid was determined
using a standard pH meter (Mett-Toledo AG, Schw-
erzenbach, Switzerland). Ammonia-N was deter-
mined according to Chaney and Marbach (1962).
IVDMD was determined according to the filter paper
procedure (Vogel et al., 1999).
Microbial population analysis
Metagenomic DNA was extracted from 250 ll of fer-
mentation fluid by bead beating and purified using a
QIAampâ DNA stool mini kit (QIAGEN, Seoul,
Korea). To ensure higher DNA yield, 1 ml of sample
was pipetted into a micro tube using a wide end tip,
centrifuged (16 000 9 g, 4 °C, 5 min) and the super-
natant discarded. Another 1 ml of fermentation fluid
sample was added to the tube with pellet and cen-
trifuged again (16 000 9 g, 4 °C, 5 min). The super-
natant was discarded and 400 ll of fermentation fluid
sample added into the tube with pellet and homoge-
nized. After allowing the mixture to settle for 2 min,
250 ll of the sample was used in the extraction of
DNA. Lysis buffer was added to the sample and the
mixture homogenized for 3 min using a BeadBug
microtube homogenizer (Benchmark Scientific, Edi-
son, NJ, USA). After incubation at 70 °C for 15 min,
with gentle shaking every 5 min, the sample was cen-
trifuged at 4 °C for 5 min at 16 000 9 g. After trans-
ferring the supernatant to a fresh Eppendorfâ tube
(Sigma-Aldrich, Seoul, South Korea), the process was
repeated after adding 300 ll of lysis buffer to the lysis
tube. To precipitate nucleic acids, PPT mix amounting
to 10% of supernatant volume was added to each
tube, inverted 4–5 times and placed on ice for 5 min.
After centrifuging at 4 °C for 10 min at 16 000 9 g,
supernatants were transferred to fresh tubes and equal
volumes of isopropanol added. After mixing well by
pumping, tubes were left on ice for more than
Journal of Animal Physiology and Animal Nutrition © 2016 Blackwell Verlag GmbH
Optimized blend of methane inhibitors
30 min. Samples were then centrifuged (16 000 9 g,
4 °C, 15 min), the supernatant discarded and 1 ml of
70% ethanol added. Samples were centrifuged again
(16 000 9 g, 4 °C, 5 min) and then left to dry for
20 min. The remaining pellet was dissolved in 100 ll
of TE buffer, and the two aliquots pooled. To remove
RNA, 20 ll of (1 mg/ml) DNase-free RNase were
added and incubated at 37 °C for 15 min. After this,
DNA was purified using proteinase K and buffers AL,
AW1, AW2 and AE of QIAGEN (QIAGEN Seoul,
Korea). The DNA yield was quantified using an
Epoch spectrophotometer (BioTek, Winooski, VT,
USA). Standards were created from DNA extracted
from a cultured mixture of rumen microbes. Tenfold
series dilutions of the standards were made and used
to create standard curves. The starting quantities
of target genes for samples were determined from the
respective standard curves of the different microbes.
A CFX96 Real-Time System (BioRad, Foster City,
CA, USA) was used to carry out quantitative real-time
PCR (qPCR) analysis to determine methanogen and
total microbial counts. The following SBYR-based pri-
mers were used; MB1174F, 5ʹ–GAGGAAGGAG
TGGACGACGGTA–3ʹ; Arch 1406-1389R, 5ʹ–ACGGG
CGGTGTGTGCAAG – 3ʹ (methanogens) (Ohene-Adjei
et al., 2008), 341f-GC, 5ʹ–CCTACGGGAGGCAGCAG
3ʹ; 534r-GC, 5ʹ–ATTACCGCGGCTGCTGG-3ʹ (total
bacteria) (Dar et al., 2005). Real-time PCR amplifica-
tion for total bacteria target genes was initiated by a
hot start at 94 °C for 5 min followed by 30 cycles of
80 °C for 60 s, 65 °C for 60 s, 55 °C for 60 s and an
extension of 72 °C for 3 min. As for archaea, amplifi-
cation was initiated by denaturation at 94 °C for
4 min followed by 35 cycles of denaturation (94 °C
for 30 s), annealing (52 °C for 30 s) and extension
(72 °C for 60 s) and then a final extension (72 °C for
7 min).
Statistical analyses
In Exp. 1, data on rumen fermentation parameters
were subjected to a one-way analysis of variance
(ANOVA) as a completely randomized design using the
GLM procedure of SAS 9.2 (SAS Inst., Cary, NC, USA),
with treatment as a fixed effect. In Exp. 2, construc-
tion of the three variable, three-level Box Behnken
design, ANOVA with observed responses and the
response surface model fitting for the calculation of
optimum blending condition were performed using
MINITAB v.16 (Ryan, 2012). Regression analysis was
performed and fitted into the empirical second-order
polynomial model for the experimental data as shown
in the equation below:
3
Optimized blend of methane inhibitors D. T. Mbiriri et al.

X
3 X
3 2 X
X 3 reduced acetate concentration. FM increased propi-
Y ¼ b0 þ bi Xi þ bii Xi2 þ bij Xi Xj onate concentration whereas GO had no effect. NT
i¼1 i¼1 i¼1 j¼1þ1
and GON negatively affected propionate concentra-
tion. NT, FM and GON reduced the production of
where Y = dependent variable (total gas, methane,
butyrate.
total VFA)
b0bibiibij = intercept, regression coefficients of linear,
quadratic and interaction terms of the model, respec- Table 3 Treatment arrangements for the Box Behnken experiment, four
different controls and their responses of VFA, biogas and CH4 after 24 h
tively
of in vitro batch fermentation (Experiment 2)
Xi and Xj = independent variables
Data obtained from Exp. 3 were analysed using a Variable* Response†
two-way ANOVA according to the PROC GLM proce-
Runs GO FM NT Total VFA Total gas CH4
dures of SAS 9.2 (SAS Institute, Carry, NC), with treat-
ment, incubation time and interaction between 1 30 7 14 66.39 75.00 1.34 9 101
treatment and incubation time as fixed effects. Means 2 300 7 14 57.73 68.67 1.65 9 102
3 30 21 14 67.63 86.67 1.19 9 101
in experiments 1 and 3 were separated using the
4 300 21 14 70.20 78.33 4.04 9 102
Duncan’s multiple range test (p < 0.05). A p-value 5 30 14 7 63.90 79.00 3.31 9 101
was interpreted as significant when p < 0.05 and as 6 300 14 7 69.29 73.33 6.74 9 102
tendency when 0.05 ≤ p ≤ 0.100. 7 30 14 21 71.17 71.33 7.00 9 104
8 300 14 21 64.44 68.67 7.00 9 104
9 165 7 7 59.47 63.33 6.00 9 104
Results 10 165 21 7 69.13 80.33 3.86 9 102
11 165 7 21 64.26 55.67 6.00 9 104
Experiment 1
12 165 21 21 72.02 70.33 7.00 9 104
All the measured fermentation parameters were 13 165 14 14 67.21 59.67 6.00 9 104
affected (p < 0.001) by treatments (Table 2). GO, NT, 14 165 14 14 65.44 61.00 6.00 9 104
GOF and GON lowered CH4 production. Fumarate, 15 165 14 14 71.47 60.33 6.00 9 104
when included alone, however did not have an effect Control 1 0 0 0 74.56 100.67 1.31 9 101
on CH4 reduction. Treatments containing garlic oil Control 2 30 0 0 70.50 98.67 1.34 9 101
(blend and GO), with the exception of GON, had Control 3 165 0 0 64.30 87.00 2.23
Control 4 300 0 0 59.55 79.67 6.95 9 101
higher H2 than other treatments. The treatments
affected individual VFA profiles and also total VFA. *Variable: GO, Garlic oil (mg/l); FM, Fumarate (mM); NT, Nitrate (mM).
NT, GO and GON reduced total VFA but FM and GOF †Response: Total VFA, Total volatile fatty acids (mM), Total gas (ml),
did not differ from the control. NT, GO, GON and GOF CH4 (ml).

Table 2 Effect of treatments on in vitro ruminal fermentation parameters after 24 h of in vitro batch fermentation (Experiment 1)

Treatment†

Parameter* CON GO FM NT GOF GON SEM p-value

Total gas (ml) 76.00b 45.67d 89.33a 44.67d 70.67c 36.33e 4.69 <0.001
CH4 (ml) 6.35a 0.40c 4.36b 1.07c 0.28c 0.26c 0.59 <0.001
H2 (ml) 0.06c 1.46a 0.01c 0.08c 1.23b 0.09c 0.15 <0.001
pH 6.62c 6.66b 6.56d 6.72a 6.59cd 6.74a 0.02 <0.001
NH3-N (mg/100 ml) 2.89b 1.04bc 2.24bc 9.13a 0.48c 9.77a 0.84 <0.001
Acetate (mM) 38.46a 26.84c 38.11a 32.67b 26.79c 27.64c 1.30 <0.001
Propionate (mM) 12.64b 11.86b 17.76a 8.72c 17.05a 7.64c 0.93 <0.001
Butyrate (mM) 4.99a 4.85ab 4.38b 1.38c 4.42b 1.41c 0.38 <0.001
Valerate (mM) 1.37a 1.04b 1.27a 0.88c 1.03b 0.87c 0.05 <0.001
Total VFA (mM) 57.46a 44.60b 61.52a 43.65b 49.29b 37.56c 2.09 <0.001
A:P 3.04c 2.26d 2.14e 3.75a 1.57f 3.62b 0.19 <0.001

*NH3-N, ammonia-N; TVFA, total volatile fatty acid; A:P, acetate:propionate.

†Treatments: GO, garlic oil at 300 mg/l; FM, fumarate at 21 mM; NT, nitrate at 21 mM; GON, garlic oil at 300 mg/l + Nitrate at 21 mM; GOF, garlic oil at
300 mg/l + fumarate at 21 mM.
a-f
Within a row, means without a common superscript differ (p < 0.05).

4 Journal of Animal Physiology and Animal Nutrition © 2016 Blackwell Verlag GmbH
D. T. Mbiriri et al. Optimized blend of methane inhibitors

Experiment 2 interaction among treatments were observed in biogas

(p = 0.776) and CH4 (p = 0.259), total VFA tended
The used experimental design layout with four differ-
(p = 0.077) to be affected by interactive effects among
ent controls and responses of total VFA, total gas and
treatments. The regression coefficients obtained are
CH4 production from every run are shown in Table 3.
shown in the models below;
The output from ANOVA of responses from 15 runs
showed significant linear effects on biogas
(p = 0.004), a tendency on CH4 (p = 0.053) but no Ybiogas ¼ 299:43  0:95X1  3:19X2  4:31X3 þ 2:00
linear effects on total VFA (p = 0.483). A significant  103 X12 þ 0:24X22 þ 0:1X32  0:0X1 X2
quadratic effect was observed only in biogas produc-
þ 0:01X1 X3  0:01X2 X3 ðR2 ¼ 0:99Þ
tion (p = 0.001). Although no significant effects of

(a) 300 (b) 50

40
Total gas (ml/g DMD)

200 Methane (ml/g DMD)

30

20
100

10

0 0
6 12 18 24 30 36 42 48 6 12 18 24 30 36 42 48

(c) 2.5 (d) 20

2.0
15
Ammonia N (mg/100 ml)
Hydrogen (ml/g DMD)

1.5

10

1.0

5
0.5

0.0 0
6 12 18 24 30 36 42 48 0 6 12 18 24 30 36 42 48
Incubation time (h) Incubation time (h)

Fig. 1 The effect of Experiment 3 treatments; no additive ( , CON), garlic oil, 50 mg/l ( , GO), fumarate, 15 mM ( , FM), nitrate, 20 mM ( , NT)
and a blend of garlic oil, fumarate and nitrate ( , GONF) on in vitro rumen fermentation parameters; (a) Total gas, (b) CH4, (c) H2 and (d) Ammonia N.

Journal of Animal Physiology and Animal Nutrition © 2016 Blackwell Verlag GmbH 5
Optimized blend of methane inhibitors
Ymethane ¼ 1:44  0:01X1 þ 0:01X2  0:07X3 þ 1:00
 105 X12  2:00  104 X22 þ 8:00
 104 X32 þ 3:00  105 X1 X2 þ 2:00
 104 X1 X3  0:01X2 X3 ðR2 ¼ 0:86Þ
Ytotal VFA ¼ 47:97 þ 0:01X1 þ 1:21X2 þ 0:88X3
 0:0X12  0:04X22  1:00  103 X32
þ 3:00  103 X1 X2  3:00  103 X1 X3
 0:01X2 X3 ðR2 ¼ 0:87Þ
Ybiogas (ml), Ymethane (ml) and Ytotal VFA (mM) are
response variables. X1, X2 and X3 represent garlic oil
(mg/ml), fumarate (mM) and nitrate (mM), respectively.
*Coefficients are significant (p < 0.05).
The major VFA in rumen nutrition were altered
by treatments in various ways. Significant linear
effects were observed in propionate (p = 0.006) and
butyrate (p = 0.001) and tendency was observed in
acetate (p = 0.051). Quadratic effects were seen in
propionate (p = 0.002) and butyrate (p = 0.001)
concentrations but not in acetate (p = 0.266). Inter-
active effects between GO and NT negatively affected
(p = 0.008) on propionate concentration.
The following optimal blend of compounds and
predictions were made after applying the response
surface optimization procedure that maximizes total
VFA and biogas production but minimizes CH4;
Optimal blend = GO, 50 mg/l; FM, 15 mM; NT,
20 mM.
Predictions = biogas, 70 ml, CH4, 0.03 ml, total
VFA, 72 mM.
Experiment 3
Interactions between time and treatment on biogas
production were observed (p < 0.001). The production
Table 4 Effect of experiment treatment on in vitro fermentation parameters (Experiment 3)
Treatment*
Variable CON GO FM
pH 6.69 c
6.70b
6.57 e
IVDMD (%) 51.64a 51.01a 50.99a
Total VFA (mM) 44.66b 41.85d 49.80a
A:P 2.61c 2.41d 2.33e
IVDMD, in vitro dry matter disappearance, VFA, volatile fatty acids, A:P, acetate:propionate.
*Treatments: CON, no additive, GO, garlic oil at 50 mg/l; FM, fumarate at 15 mM; NT, nitrate at 20 mM; GONF, a blend of garlic oil at 50 mg/l + fumarate
at 15 mM + nitrate at 20 mM.
†Significance: probability for, Trt, treatment; Time, incubation time; Trt 9 Time = interaction between treatment and incubation time.
a-e
Means within a row with different superscripts differ (p < 0.05).
6 Journal of Animal Physiology and Animal Nutrition © 2016 Blackwell Verlag GmbH
D. T. Mbiriri et al.
patterns are shown in Fig. 1a. NT and the blend sup-
pressed biogas production significantly when compared
to control treatments (p < 0.0001). GO, on the other
hand, showed a suppression of biogas whereas FM
increased biogas numerically (p = 0.1259). The blend
almost completely inhibited (97%) methanogenesis
when compared to the control (p < 0.0001). It was fol-
lowed by NT (p < 0.0001) which suppressed CH4 pro-
duction by 93%. Although GO (p < 0.001) and FM
(p = 0.0327) also reduced methanogenesis, it was at
lower levels; 32% and 8% suppression, respectively.
As Fig. 1b shows, the CH4 reduction effects of FM
begin to show after 24 h of incubation. Garlic oil
(p < 0.0001) and NT (p = 0.0308) resulted in higher
H2 accumulation. GO addition resulted in a higher
accumulation of H2 output (Fig. 1c). The blend and NT
had similar rises in H2 after the 24 h mark. The ammo-
nia-N concentration in the fermentation fluid was
increased by the addition of NT (p < 0.0001) and blend
(p < 0.0001) as shown in the production trends in
Fig. 1d. FM reduced (p = 0.0121) ammonia-N concen-
tration whereas GO showed diminution in the same.
FM increased total VFA concentration but NT and
GO lowered compared to the control. The blend
showed no difference from the control on the same
variable (Table 4). Fig. 2a–d shows the effects of treat-
ments on molar concentrations of individual VFA.
FM, NT and the blend increased (p < 0.0001),
whereas GO reduced acetate (p < 0.0001). FM and
the blend increased (p < 0.0001) the molar proportion
of propionate by 13% and 11%, respectively. NT
(p = 0.3518) and GO (p = 0.9987) on the other hand
had no effects on propionate. Butyrate and valerate
production were suppressed by NT and the blend
(p < 0.0001). The rest of the treatments did not differ
from the control (p > 0.05). The acetate to propionate
ratio was affected by treatments in the following
ascending order; FM < GO < control < blend < NT
(Table 4).
Significance†
NT GONF SEM Trt Time Trt 9 Time
6.71 a
6.60 d
0.01 <0.001 <0.001 <0.001
46.57b 47.79b 1.29 <0.001 <0.001 <0.001
43.78c 45.27b 1.68 <0.001 <0.001 <0.001
2.79a 2.67b 0.04 <0.001 <0.001 <0.001
D. T. Mbiriri et al. Optimized blend of methane inhibitors

Table 4 also shows the effect of treatments on
IVDMD and pH. The blend and NT suppressed IVDMD
significantly (p < 0.05). Garlic oil and NT registered
pH values greater than that of the control whereas FM
and the blend were lower.
The predicted and expected total gas, CH4 and
total VFA output from the inclusion of the optimized
blend in batch fermentation are shown in Table 5.
The blend suppressed CH4 output without negatively
affecting total VFA concentration, although there
were discrepancies between predicted and observed
values.
The effects of treatments on the microbial commu-
nity are shown in Table 6. Although no significant
treatment effects on methanogens were observed at
time 0 h (p > 0.05), NT and the blend reduced metha-
nogen counts significantly (p < 0.0001) after 48 h
incubation. Total bacterial counts were not affected by
treatments.
Discussion
The observed effect by the blend of suppressing CH4
without affecting total VFA concentration can be

(a) 50 (b) 25

40 20

Propionate (mM)
Acetate (mM)

30 15

20 10

10 5

0 0
0 6 12 18 24 30 36 42 48 0 6 12 18 24 30 36 42 48

(c) 10 (d) 8

8
6
Butyrate (mM)

Valerate (mM)

2
2

0 0
0 6 12 18 24 30 36 42 48 0 6 12 18 24 30 36 42 48
Incubation time (h) Incubation time (h)

Fig. 2 The effect of Experiment 3 treatments; no additive ( , CON), garlic oil, 50 mg/l ( , GO), fumarate, 15 mM ( , FM), nitrate, 20 mM ( , NT) and
a blend of garlic oil, fumarate and nitrate ( , GONF) on in vitro rumen fermentation parameters; (a) Acetate, (b) Propionate, (c) Butyrate and (d) Valerate.

Journal of Animal Physiology and Animal Nutrition © 2016 Blackwell Verlag GmbH 7
Optimized blend of methane inhibitors
Table 5 Rumen fermentation parameter responses predicted by the
response surface model after 24 h in vitro batch fermentation (Experi-
ment 2) vs. the observed responses in a verification trial (Experiment 3)
at the addition of an optimized blend§ of anti-methanogens
Predicted Mean observed
Fermentation Parameter response† response (n = 3)‡
Total gas (ml) 70.00 55.00
Methane (ml) 0.03 0.20
Total VFA (mM) 72.00 66.11
Total VFA, Total volatile fatty acids.
§Blend = a mixture of garlic oil at 50 mg/l + fumarate at 15 mM + nitrate
at 20 mM.
†Predicted responses are what were expected after adding the opti-
mized blend in in vitro fermentation cultures. They were obtained after
inputting the optimized quantities of antimethanogens to formulate a
blend into the following polynomial equations;
Ybiogas ¼ 299:43  0:95X1  3:19X2  4:31X3 þ 2:00  103 X12
þ 0:24X22 þ 0:1X32  0:0X1 X2 þ 0:01X1 X3
 0:01X2 X3 ðR2 ¼ 0:99Þ
Ymethane ¼1:44  0:01X1 þ 0:01X2  0:07X3 þ 1:00  105 X12
2:00  104 X22 þ 8:00  104 X32 þ 3:00  105 X1 X2
þ2:00  104 X1 X3  0:01X2 X3 ðR2 ¼ 0:86Þ
Ytotal VFA ¼ 47:97 þ 0:01X1 þ 1:21X2 þ 0:88X3  0:0X12  0:04X22
 1:00  103 X32 þ 3:00  103 X1 X2  3:00  103 X1 X3
 0:01X2 X3 ðR2 ¼ 0:87Þ
Ybiogas (ml), Ymethane (ml) and Ytotal VFA (mM) are response variables. X1,
X2 and X3 represent garlic oil (mg/ml), fumarate (mM) and nitrate (mM),
respectively.
*Coefficients are significant (p < 0.05).
‡Observed responses were obtained after carrying out an in vitro
fermentation experiment to verify the effect of the optimized blend.
attributed to the additive effects of the three CH4 miti-
gating compounds making up the blend. The close
resemblance in trends of total gas, CH4 and H2
between the blend and NT suggests a higher influence
of nitrate in the blend than the other components.
This is further supported by the negative effects of the
two treatments on methanogen counts. Nitrate has
been observed to lower methanogen counts and the
effect can be explained in three possible ways; the
depletion of H2 to methanogens as it is used up in
nitrate reduction (Zhou et al., 2012), the sensitivity of
methanogenic archaea to nitrate and nitrite (Iwamoto
et al., 2002; Sar et al., 2005b) or the negative impact
on fermentation. Consequently, nitrate supplementa-
tion is an effective CH4 mitigating agent, both in vitro
(Sar et al., 2005a; Zhou et al., 2012) and in vivo (Sar
et al., 2004; Van Zijderveld et al., 2010a; Hulshof
et al., 2012). High ammonia-N concentrations have
also been reported to inhibit methanogenic activity
8 Journal of Animal Physiology and Animal Nutrition © 2016 Blackwell Verlag GmbH
D. T. Mbiriri et al.
(Chen et al., 2008). This could be another means con-
tributing to the reduction of CH4 in NT and the blend,
considering that nitrate addition in all the experi-
ments in the study resulted in increase in ammonia-N
concentration. In vivo studies have made use of nitrate
at concentrations ranging from 10 to 26 g/kg DM of
feed without causing toxicity. However, in some cases
these concentrations reduced or tended to reduce
DMI and VFA concentration (Farra and Satter, 1971;
Bruning-Fann and Kaneene, 1993; Hulshof et al.,
2012). To achieve a nitrate concentration of 20 mM
used in this study in vivo it would require 6.8 g/kg
DM, assuming a cow with a physiological rumen vol-
ume of 80 l and consuming 20 kg DM of feed per day.
Garlic oil, at sufficient concentrations, selectively
inhibits archaea (Busquet et al., 2005a,b; Chaves
et al., 2008). Although at a low inclusion of 50 mg/l,
in Exp. 3, GO did not affect archaea counts, it still low-
ered CH4 production by more than 30%. This seem-
ingly contradicting result can be explained by the
mechanism of rumen microbes adapting to garlic oil
as incubation time increased (Busquet et al., 2005a), a
concentration not high enough or the active agent
was degraded. Fumarate, theoretically suppresses CH4
by taking up H2 in its reduction to succinate and ulti-
mately to propionate. However, according to previous
batch culture studies, the efficacy of fumarate in sup-
pressing CH4 is relatively small as noted in a meta-
analysis by Ungerfeld et al. (2007). Lopez et al. (1999)
had earlier reported that fumarate does not affect total
bacteria and methanogenic archaea populations. Its
application has, however, consistently resulted in
increasing the propionate fraction and or maintaining
total VFA concentration (Ungerfeld et al., 2007). It is
this property of fumarate that ensured the blend did
not suppress total VFA concentration even though
nitrate and garlic oil suppressed VFAs. The negative
effect of nitrate on total VFA and butyrate in vitro have
been reported in earlier studies (Sar et al., 2005a,b;
Takahashi, 1989). In contrast with propionate reduc-
tion reported in in vitro (Sar et al., 2005a,b) and in
in vivo studies (Farra and Satter, 1971; Allison and
Reddy, 1984), the inclusion of nitrate in the present
study did not affect molar proportions of propionate.
This could be partially explained by the high concen-
trate diet which would be fermented to produce high
levels of propionate, masking the effect of nitrate.
Zhou et al. (2012) and Alaboudi and Jones (1985)
reported an increase in the molar proportion of acet-
ate with nitrate supplementation, an observation con-
sistent with findings of this study. The suppressed VFA
production in GO, was also reported by previous
researchers (Busquet et al., 2005b; Cardozo et al.,
D. T. Mbiriri et al. Optimized blend of methane inhibitors

Table 6 Effect of treatments on total bacteria and methanogenic archaea (copies/mg of pellet, log10) at 0 h and after 48 h incubation (Experiment 3)

Treatment

Microbes Incubation time (h) CON GO FM NT GONF SEM p-value

Total bacteria 0 10.86 11.14 10.87 10.79 9.81 4.45 0.6744

48 10.88 10.91 10.96 10.93 11.24 4.39 0.9740
Methanogenic archaea 0 5.98 6.11 5.77 5.84 5.87 1.96 0.9909
48 6.68a 6.34ab 6.57ab 4.81b 4.85b 2.24 0.0117

Treatments: CON, No additive, GO, garlic oil at 50 mg/l; FM, fumarate at 15 mM; NT, nitrate at 20 mM; GONF, a blend of garlic oil at 50 mg/l + fumarate
at 15 mM + nitrate at 20 mM. Data presented as mean
SD. n = 3/treatment.
a,b
Means within a row with different superscripts differ (p < 0.05).

2005), but it was at greater garlic oil concentrations.
These observations are, however, in contrast with
results obtained in vitro by Chaves et al. (2008) and
in vivo by Yang et al. (2007) who did not find any neg-
ative effects of garlic oil on total VFA concentration.
Treatments that had low VFA concentration had high
pH values in contrast to low pH values in treatments
that had high VFA concentration.
Low total gas volume in the blend can be attributed
to garlic oil (Anassori et al., 2012) and nitrate (Zhou
et al., 2012). Gas production is an indicator of dry mat-
ter digestibility. Some researchers have reported that
garlic oil negatively affects nutrient digestibility (Anas-
sori et al., 2012) whereas others, like in this study, did
not observe any effects (Busquet et al., 2005a; Yang
et al., 2007). The explanation to this could be differ-
ences in dosage and diet (Anassori et al., 2012). Nitrite
accumulation has also been observed to reduce
digestibility by impairing cellulose and xylanase
activities (Marais et al., 1988). Increased total gas pro-
duction in FM could be from fumarate reduction itself
and not an increase in organic matter digestion since
IVDMD was not increased by the addition of fumarate.
Carbon dioxide is released when fumarate is reduced to
propionate (Garcıa-Martınez et al., 2005). Lopez et al.
(1999) observed a 6.3% increase in IVDMD after
adding fumarate in a continuous culture system. The
differences in culture system, diet and concentration
could explain the differences in IVDMD results between
the current study and that of Lopez et al. (1999).
Conclusions and recommendations
Supplementing a high concentrate substrate with an
optimized blend of garlic oil, fumarate and nitrate
at 50 mg/l, 15 mM and 20 mM, respectively, mark-
edly decreased CH4 production and maintained total
VFA production in spite of suppressed feed
digestibility. While the inclusion of nitrate or garlic
oil also lowered CH4, they reduced total VFA con-
centration. The optimal ratio formulated here may
be used as a guide in formulating a mix considering
that the diet is nitrogen-rich already. The observed
results, however, need to be verified in vivo for
in vitro studies outcomes do not always translate to
similar outcomes in vivo (Soliva et al., 2011). Cau-
tion will have to be exercised when conducting
in vivo tests to avoid toxicity.
Acknowledgements
This research was supported by The Ministry of Agri-
culture, Food and Rural Affairs, Korea under the pro-
ject, ‘Research on feed additives for reducing the
production of methane by ruminants’.

References
Alaboudi, A.; Jones, G., 1985: Effect of
acclimation to high nitrate intakes on
some rumen fermentation parameters
in sheep. Canadian Journal of Animal
Science 65, 841–849.
Allison, M. J.; Reddy, C. A., 1984: Asapta-
tions of Gastrointestinal Bacteria in Response
to Changes in Dietary Oxalate and Nitrate.
Third International Symposium on
Microbial Ecology, Washington DC, pp.
248–256.
Anassori, E.; Dalir-Naghadeh, B.;
Pirmohammadi, R.; Taghizadeh, A.;
Asri-Rezaei, S.; Farahmand-Azar, S.;
Besharati, M.; Tahmoozi, M., 2012:
In vitro assessment of the digestibility
of forage based sheep diet, supple-
mented with raw garlic, garlic oil and
monensin. Veterinary Research Forum 3,
5–11.
Box, G. E.; Behnken, D. W., 1960: Some
new three level designs for the study of
quantitative variables. Technometrics 2,
455–475.
Bruning-Fann, C. S.; Kaneene, J., 1993:
The effects of nitrate, nitrite, and
N-nitroso compounds on animal health.
Veterinary and Human Toxicology 35, 237–
253.
Busquet, M.; Calsamiglia, S.; Ferret, A.;
Cardozo, P. W.; Kamel, C., 2005a:

Journal of Animal Physiology and Animal Nutrition © 2016 Blackwell Verlag GmbH 9
Optimized blend of methane inhibitors
Effects of cinnamaldehyde and garlic oil
on rumen microbial fermentation in a
dual flow continuous culture. Journal of
Dairy Science 88, 2508–2516.
Busquet, M.; Calsamiglia, S.; Ferret, A.;
Carro, M. D.; Kamel, C., 2005b: Effect of
garlic oil and four of its compounds on
rumen microbial fermentation. Journal
of Dairy Science 88, 4393–4404.
Cardozo, P.; Calsamiglia, S.; Ferret, A.;
Kamel, C., 2005: Screening for the
effects of natural plant extracts at differ-
ent pH on in vitro rumen microbial fer-
mentation of a high-concentrate diet for
beef cattle. Journal of Animal Science 83,
2572–2579.
Chaney, A. L.; Marbach, E. P., 1962: Mod-
ified reagents for determination of urea
and ammonia. Clinical Chemistry 8, 130–
132.
Chaves, A.; He, M.; Yang, W.; Hristov, A.;
McAllister, T.; Benchaar, C., 2008:
Effects of essential oils on proteolytic,
deaminative and methanogenic
activities of mixed ruminal bacteria.
Canadian Journal of Animal Science 88,
117–122.
Chen, Y.; Cheng, J. J.; Creamer, K. S.,
2008: Inhibition of anaerobic digestion
process: a review. Biresource Technology
99, 4044–4064.
Dar, S. A.; Kuenen, J. G.; Muyzer, G.,
2005: Nested PCR-denaturing gradient
gel electrophoresis approach to deter-
mine the diversity of sulfate-reducing
bacteria in complex microbial commu-
nities. Applied and Environmental Microbi-
ology 71, 2325–2330.
Erwin, E.; Marco, G.; Emery, E., 1961:
Volatile fatty acid analyses of blood and
rumen fluid by gas chromatography.
Journal of Dairy Science 44, 1768–1771.
Farra, P.; Satter, L., 1971: Manipulation of
the ruminal fermentation. III. Effect of
nitrate on ruminal volatile fatty acid
production and milk composition. Jour-
nal of Dairy Science 54, 1018–1024.
Garcıa-Martınez, R.; Ranilla, M.; Tejido,
M.; Carro, M., 2005: Effects of disodium
fumarate on in vitro rumen microbial
growth, methane production and fer-
mentation of diets differing in their for-
age: concentrate ratio. British Journal of
Nutrition 94, 71–77.
Hulshof, R.; Berndt, A.; Gerrits, W.; Dijk-
stra, J.; Van Zijderveld, S.; Newbold, J.;
Perdok, H., 2012: Dietary nitrate supple-
mentation reduces methane emission in
beef cattle fed sugarcane-based diets.
Journal of Animal Science 90, 2317–2323.
10
Isobe, Y.; Shibata, F., 1993: Rumen fer-
mentation in goats administered fuma-
ric acid. Animal Science and Technology 64,
1024–1030.
Iwamoto, M.; Asanuma, N.; Hino, T.,
2002: Ability of Selenomonas ruminan-
tium, Veillonella parvula, and Wolinella
succinogenes to reduce nitrate and nitrite
with special reference to the suppression
of ruminal methanogenesis. Anaerobe 8,
209–215.
Kobayashi, Y., 2010: Abatement of
methane production from ruminants:
trends in the manipulation of rumen
fermentation. Asian-Australasian Journal
of Animal Science 23, 410–416.
Lopez, S.; Valdes, C.; Newbold, C. J.; Wal-
lace, R. J., 1999: Influence of sodium
fumarate addition on rumen fermenta-
tion in vitro. British Journal of Nutrition
81, 59–64.
Marais, J. P.; Therion, J. J.; Mackie, R. I.;
Kistner, A.; Dennison, C., 1988: Effect
of nitrate and its reduction products on
the growth and activity of the rumen
microbial population. British Journal of
Nutrition 59, 301–313.
Martin, C.; Morgavi, D.; Doreau, M.,
2010: Methane mitigation in ruminants:
from microbe to the farm scale. Animal
4, 351–365.
McDougall, E., 1948: Studies on ruminant
saliva. 1. The composition and output of
sheep’s saliva. Biochemical Journal 43,
99–109.
Ohene-Adjei, S.; Chaves, A.; McAllister,
T.; Benchaar, C.; Teather, R.; Forster, R.,
2008: Evidence of increased diversity of
methanogenic archaea with plant
extract supplementation. Microbial Ecol-
ogy 56, 234–242.
Ryan, B.; Joiner, B.; Cryer, J., 2012: MINI-
TAB Handbook: update release, 16th
edn. Cengage, BOston, MA, USA.
Sar, C.; Santoso, B.; Gamo, Y.; Kobayashi,
T.; Shiozaki, S.; Kimura, K.; Mizukoshi,
H.; Arai, I.; Takahashi, J., 2004: Effects
of combination of nitrate with beta 1-4
Galacto-oligosaccharides and yeast (Can-
dida kefyr) on methane emission from
sheep. Asian-Australasian Journal of Ani-
mal Science 17, 73–79.
Sar, C.; Mwenya, B.; Pen, B.; Morikawa,
R.; Takaura, K.; Kobayashi, T.; Takahashi,
J., 2005a: Effect of nisin on ruminal
methane production and nitrate/nitrite
reduction in vitro. Crop and Pasture
Science 56, 803–810.
Sar, C.; Mwenya, B.; Santoso, B.; Takaura,
K.; Morikawa, R.; Isogai, N.; Asakura,
Journal of Animal Physiology and Animal Nutrition © 2016 Blackwell Verlag GmbH
D. T. Mbiriri et al.
Y.; Toride, Y.; Takahashi, J., 2005b:
Effect of Escherichia coli wild type or its
derivative with high nitrite reductase
activity on in vitro ruminal methano-
genesis and nitrate/nitrite reduction.
Journal of Animal Science 83, 644–652.
SAS Institute Inc. 2008: SAS/STAT 9.2
user’s guide. SAS Institute Inc., Cary,
NC.
Soliva, C. R.; Amelchanka, S. L.; Duval, S.
M.; Kreuzer, M., 2011: Ruminal
methane inhibition potential of various
pure compounds in comparison with
garlic oil as determined with a rumen
simulation technique (Rusitec). British
Journal of Nutrition 106, 114–122.
Takahashi, J., 1989: Effect of nitrate con-
tent of forage on the production of vola-
tile fatty acids by sheep rumen microbes
in vitro. Japanese Journal of Zootechnical
Science 60, 476–483.
Ungerfeld, E. M.; Kohn, R. A.; Wallace,
R. J.; Newbold, C. J., 2007: A meta-
analysis of fumarate effects on
methane production in ruminal batch
cultures. Journal of Animal Science 85,
2556–2563.
Van Zijderveld, S.; Gerrits, W.; Apajalahti,
J.; Newbold, J.; Dijkstra, J.; Leng, R.;
Perdok, H., 2010a: Nitrate and sulfate:
Effective alternative hydrogen sinks for
mitigation of ruminal methane produc-
tion in sheep. Journal of Dairy Science 93,
5856–5866.
Van Zijderveld, S.; Gerrits, W.; Apajalahti,
J.; Newbold, J.; Dijkstra, J.; Leng, R.;
Perdok, H., 2010b: Nitrate and sulfate:
effective alternative hydrogen sinks for
mitigation of ruminal methane produc-
tion in sheep. Journal of Dairy Science 93,
5856–5866.
Vogel, K. P.; Pedersen, J. F.; Masterson, S.
D., Toy, J. J., 1999: Evaluation of a filter
bag system for NDF, ADF, and IVDMD
forage analysis. Crop Science 39, 276–279.
Yang, W.; Benchaar, C.; Ametaj, B.;
Chaves, A.; He, M.; McAllister, T., 2007:
Effects of garlic and juniper berry essen-
tial oils on ruminal fermentation and on
the site and extent of digestion in lactat-
ing cows. Journal of Dairy Science 90,
5671–5681.
Zhou, Z.; Yu, Z.; Meng, Q., 2012: Effects of
nitrate on methane production, fermen-
tation, and microbial populations in
in vitro ruminal cultures. Bioresource
Technology 103, 173–179.