Journal of Integrative Agriculture 2017, 16(7): 1566–1575

Available online at www.sciencedirect.com

ScienceDirect

RESEARCH ARTICLE

In vitro investigation of the effect of dairy propionibacteria on rumen

pH, lactic acid and volatile fatty acids

Jianbiao Luo1, Chaminda Senaka Ranadheera1, 2, Stuart King1, Craig Evans1, Surinder Baines3

1
School of Environmental & Life Sciences, The University of Newcastle, Callaghan NSW 2308, Australia
2
Advanced Food Systems Research Unit, College of Health and Biomedicine, Victoria University, Melbourne VIC 3030, Australia
3
School of Health Sciences, The University of Newcastle, Callaghan NSW 2308, Australia

Abstract
Ruminal acidosis is a prevalent disorder in ruminants such as dairy cows and feedlot beef cattle, caused primarily by the
inclusion of a high percentage of readily fermentable concentrates in the diet. The disorder presents as an accumulation of
lactic acid, a decrease of pH in the rumen and a subsequent imbalance of the rumen fermentation process with detrimental
impacts on the animal’s health and productivity. Dairy propionibacteria, a group of bacteria characterised by utilization of lactic
acid as the favoured carbon source, with propionic acid produced as a by-product, were evaluated in this study as potential
direct-fed microbials for use in controlling ruminal acidosis. Acidosis was simulated by introduction of high concentrations
of lactic acid into rumen fluid samples and a multi-strain in vitro analysis was conducted, whereby changes in pH and lactic
acid metabolism were compared in identical acidified rumen samples, following inoculation with various propionibacteria.
This was followed by a study to evaluate the effect of bacterial inoculation dosage on acid metabolism. The results indicated
that lactic acid levels in the rumen fluid were significantly reduced, and propionic acid and acetic acid concentrations both
significantly increased, following addition of propionibacteria. Significant ‘between strains’ differences were observed, with
Propionibacterium acidopropionici 341, Propionibacterium freudenreichii CSCC 2207, Propionibacterium jensenii NCFB
572 and P. jensenii 702 each producing more rapid reduction of lactic acid concentration than P. freudenreichii CSCC 2206,
P. acidopropionici ATCC 25562 and Propionibacterium thoenii ATCC 4874. Furthermore, the efficacy of this application
was dosage related, with the rates of reduction in lactic acid levels and production of propionic acid, both significantly
greater for the higher (1010 cfu mL–1) compared with lower (105 cfu mL–1) dosage inoculation. The results confirmed that
the introduction of propionibacteria could promote more rapid reduction of lactic acid levels than would occur without their
addition, demonstrating their potential in controlling ruminal acidosis.

Keywords: probiotics, ruminal acidosis, lactic acid, propionic acid, dairy Propionibacterium

Received 1 June, 2016 Accepted 12 December, 2016
Correspondence Jianbiao Luo, Mobile: +61-0404724160, E-mail:
bjastin@hotmail.com
© 2017 CAAS. Publishing services by Elsevier B.V. All rights
reserved.
doi: 10.1016/S2095-3119(16)61556-3
1. Introduction
Lactic acid accumulation in the rumen is characteristic of
ruminal acidosis in lactating cows and intensively reared
beef cattle, as well as in other ruminants such as sheep.
The main reason for the accumulation of lactic acid in the
rumen is feeding on highly fermentable carbohydrates
 Jianbiao Luo et al. Journal of Integrative Agriculture 2017, 16(7): 1566–1575
(such as a high grain diet). The available carbohydrates
act as stimulation agents for the rapid growth of lactic
acid bacteria such as Streptococcus bovis in the rumen
(Elghandour et al. 2015). These bacteria are usually
present as part of the normal rumen micro-flora, but
during rapid growth they can produce large quantities of
lactic acid which may imbalance the buffering capacity
of the rumen, resulting in acidosis in healthy adult cattle
(Owens et al. 1998; Parrot et al. 2001; Nagaraja and
Titgemeyer 2007; Hutton et al. 2012). In healthy adult
cattle, the lactic acid concentration in the rumen does not
typically exceed 5 mmol L–1, but it may increase as much as
40 mmol L–1 during acute acidosis (Dawson et al. 1997).
The rapid accumulation of lactic acid can result in a
significant decrease in pH within the rumen. When ruminal
pH drops to a level between 5.0 and 5.5, fermentation
within the rumen is disturbed, upsetting the animal’s
capacity to digest food, leading to a reduction in feed
intake, energy absorption, and reduced meat and milk
production (Dirksen 1985; Enemark et al. 2004; Plaizier
et al. 2008). The continued production of lactic acid in the
rumen can lead to further reductions in pH, which can be
fatal to the animal if treatment is delayed.
There are a number of established strategies for
treatment and prevention of ruminal acidosis, such as
neutralization using chemicals like bicarbonate soda,
feed management (improved fibre content) and antibiotic
treatment (Enemark 2008). Although these methods
can be effective in controlling and preventing ruminal
acidosis, there are shortcomings and limitations to their
application, such as development of resistant strains of
bacteria through the long term use of antibiotics. Therefore
the use of direct-fed microbials (DFM) for the treatment
of ruminal acidosis has been investigated (Kung and
Hession 1995; Martin and Streeter 1995; Nocek et al. 2002;
Elghandour et al. 2015). Several lactobacilli including
Lactobacillus acidophilus and Lactobacillus casei, as
well as Enterococcus diacetylactis and Bacillus subtilis,
are commonly available as DFM products for ruminants.
These microorganisms may aid in providing of a constant
lactic acid supply, adapting of overall rumen microbiota to
the lactic acid accumulation, stimulation of lactate utilizing
bacteria and stabilizing of ruminal pH (Seo et al. 2010).
Compared with the prevention methods outlined above,
the DFM approach does not encourage proliferation of
antibiotic resistant strains; avoids the potential problem
of overdosing of buffering chemicals and may have better
cost-efficiency than strict feed management. However,
acid metabolism patterns can vary significantly between
bacterial species or strains, therefore appropriate selection
of DFM candidates as a treatment for ruminal acidosis
is critical. For the treatment and prevention of ruminal
1567
acidosis, the capacity for lactic acid consumption is the
primary selection criteria. It is also important that the
selected organisms do not produce any by-products during
the application that might be harmful to either the recipient
animal or the environment.
Lactic acid metabolism by propionibacteria results in
production of acetic and propionic acid which can be readily
absorbed through the ruminal wall as energy sources for
the animal (Huntington et al. 1983; Peters et al. 1990;
Pascal 1999). Propionibacteria such as Propionibacterium
acnes belong to the indigenous micro-flora of the rumen
but are generally present in low numbers under normal
conditions (Mackie and White 1990; Koniarova 1993),
possibly due to their less competitive nature. There are
no reports to indicate that substantial increases in the
number of propionibacteria occur in the rumen in response
to rising lactic acid concentrations during ruminal acidosis.
However, several previous studies have assessed the
effects of direct supplementation with dairy propionibacteria
on rumen fermentation processes in cattle. Lehloenya
et al. (2008) reported that feeding Propionibacterium
strain P169 and yeast culture to Angus×Hereford steers
for 21 days increased the molar proportion of propionate
by 9.7% compared with a control group fed a diet without
supplementation of propionibacteria or yeast. Previous
studies have also shown that the introduction of certain
strains of Propionibacterium helped to improve calf weight
gain (Adams et al. 2008) and milk quality (Francisco et al.
2002; Stein et al. 2006) without any adverse effect on the
animal, and that direct feeding with Propionibacterium
strain P169 was responsible for higher production of
ruminal propionate in Holstein cows (Stein 2006). Another
advantage in using dairy propionibacterium as DFM is their
natural ability to survive in low pH conditions (Cousin et al.
2011). Certain propionibacteria strains have demonstrated
better survival in acidic conditions around pH 3–4 (Huang
and Adams 2004; Ranadheera et al. 2012). Theoretically,
the application of DFM such as dairy propionibacteria in
ruminal acidosis is based on a potential ability to consume
and reduce the lactic acid accumulated in the rumen and
redirect the imbalance of disturbed fermentation back to
normal. To comprehend how these bacteria behave and
metabolize in rumen fluid is important for their potential
application in cattle. Not only would it provide detailed
evidence to support the theory of their application in
treating ruminal acidosis, but would also provide a valuable
guideline for future research.
The objective of the present study was to investigate
the feasibility and efficacy of using dairy propionibacteria
as DFM to treat ruminal acidosis, by examining changes
in acid levels in rumen fluid following introduction of
propionibacteria, under simulated acidosis conditions.
1568 Jianbiao Luo et al. Journal of Integrative Agriculture 2017, 16(7): 1566–1575

2. Materials and methods
2.1. Experimental design and rumen sample prepa-
ration
Intact fresh whole cattle rumens (both oesophagus and small
intestine ends were sealed to prevent leakage during trans-
portation) were obtained from a local abattoir (Kurri Meats,
Kurri Kurri, NSW Australia). All rumen were from pasture fed
Hereford beef cattle, each approximately one year old. The in-
dividual weights of the cattle ranged between 250 and 300 kg,
and each rumen weighed approximately 30 kg. All cattle
were healthy and ready for the slaughter process for human
consumption. Once dissected from the body all rumen were
stored on ice for transportation to the microbiology laboratory
at The University of Newcastle, Australia, within 3 h. Treat-
ments in each experiment were arranged in a completely
randomized design. In each replication, one whole rumen
was used. The rumen content was collected into a sterile
bucket and well mixed. 300 g of bulk rumen content was
weighed and transferred to a sterile sealed bottle as one
rumen sample. Since the acid tolerance of dairy propioni-
bacteria is strain-specific (Huang and Adams 2004), seven
different strains of propionibacteria were used in this study.
In each replication, a total of eight rumen samples were pre-
pared for a multi-strain study (control+seven strains) of the
effect of various propionibacteria on lactic acid levels, and
six for a study of the effect of inoculum dosage level. Due to
the high viscosity of the rumen content which is a mixture of
digesta and the rumen fluid, consistency between replicates
was maintained by measuring weight rather than volume. All
rumen sample preparation processes were performed under
aseptic conditions to minimise the loss of indigenous micro-
organisms and the introduction of foreign microorganisms.
2.2. Preparation of bacteria and growth conditions
Seven strains of major dairy Propionibacterium species were
used for this study (Table 1). All strains were maintained
as stock cultures stored at –80°C. A sequential process of
recovery and selection were applied to each strain prior to
use in the study. For recovery, cells from the respective
stock cultures were inoculated directly into 25 mL of SLB
(sodium lactate broth: 1% sodium lactate, 1% tryptone, 1%
yeast extract, 0.025% KH2PO4 and 0.005% MnSO4) and
incubated anaerobically at 30°C which considered to be the
optimum growth temperature for culturing propionibacteria
for 72 h. The anaerobic condition was achieved by
incubating samples in a sealed anaerobic jar (Anaerojar,
Theromoscientific, AG0025) including anaerobic sachet
(Anaerogen, Theromoscientific, AN0025) for consumption
of residual oxygen. 100 µL of the recovered cell culture
was inoculated into a fresh 25-mL aliquot of SLB and
incubated anaerobically at 30°C for 40 h. The purpose of
this procedure was to select the cells from the exponential
growth phase which had been established as occurring
from 24 to 48 h after inoculation. The final experimental cell
culture was generated by transferring 100 µL of the 40 h
cell culture to a fresh 25-mL aliquot of SLB and incubated
anaerobically at 30°C for 96 h to achieve a cell density of
approximately 109 cfu mL–1. The bacterial cells were washed
by centrifugation at 4 000 r min–1 for 15 min, discarding the
supernatants, then re-suspending the cell pellets with Milli
Q water to the same initial volume. This washing process
was repeated thrice in order to eliminate any carryover of
the culture medium.
2.3. Propionibacterium inoculation
Multi-strain study For the multi-strain study, one bottle of
300 g rumen sample was prepared for each of the seven
strains of propionibacteria. In order to simulate severe
acidosis conditions (i.e., lactate concentration approximately
80 mmol L–1), 30 mmol L–1 of (80%) lactic acid (Sigma-
Aldrich Pty Ltd., Australia) was added to the rumen samples
and well mixed. Following the addition of lactate, each of
the rumen samples were inoculated with 1 mL of a 109 cfu
mL–1 preparation of a single strain of Propionibacterium
(7 strains). The inoculated rumen sample was then shaken
manually to ensure thorough distribution.
Dosage study Based on the results of the multi-strain
study, P. jensenii 702 (PJ 702) was selected for the dosage
study (due to its more rapid utilisation of lactic acid and
efficient production of propionic acid). For this study, the

Table 1 The sources and strains of dairy propionibacteria strains used in the study
Strain name Abbreviated name Culture source
Propionibacterium jensenii 702 PJ 702 The University of Newcastle, Australia
P. jensenii NCFB 572 PJ 572 University of Verona, Italy
Propionibacterium freudenreichii CSCC 2206 PF 2206 CSIRO Melbourne, Australia
P. freudenreichii CSCC 2207 PF 2207 CSIRO Melbourne, Australia
Propionibacterium acidopropionici ATCC 25562 PA 25562 American Type Culture Collection, USA
P. acidopropionici 341 PA 341 University of Melbourne, Australia
Propionibacterium thoenii ATCC 4874 PT 4874 American Type Culture Collection, USA
 Jianbiao Luo et al. Journal of Integrative Agriculture 2017, 16(7): 1566–1575 1569

same method as per multi-strain study was applied except
that the concentrations of the PJ 702 bacterial solution
were varied as detailed in Table 2. For the higher dosage
bacterial inoculation, the washed cell solutions (109 cfu mL–1)
were centrifuged again at 4 000 r min–1 for 15 min. The
supernatants were discarded and the bacterial pellets were
resuspended in 1/10 the original volume of Milli Q water to
provide a 10-fold concentration of the bacterial solution,
resulting in a final solution concentration of 1010 cfu mL–1.
For the lower dosage inoculation, the washed cell solutions
were serially diluted in Milli Q water at a ratio of 1 to 10 to
achieve a final concentration of 105 cfu mL–1.
All treated rumen samples in both studies were incubated
in an anaerobic environment (10% CO2) at 37°C for 72 h
immediately after the inoculation.
2.4. Measurements of pH and organic acid concen-
trations
Samples (10 g) were taken for analysis from each Propioni-
bacterium inoculated rumen culture at 0, 4, 6, 12, 24, 48 and
72 h. Half of the sample (5 g) was used for pH measurement,
and the other half (5 g) used for HPLC analysis of organic
acids. The pH of the rumen samples were directly mea-
sured and recorded by pH meter (Syberscan 510, EUTEOH
Instruments, Singapore) at room temperature.
The concentrations of lactate, propionate and acetate were
determined using a high pressure liquid chromatography
(HPLC) system (Hewlett-Packard 1100 DAD, Santa Clara,
CA, USA) fitted with a Pyrospher RP-18 (125 mm×4 mm,
5 μm) column (Hewlett-Packard). Sample preparation
consisted of dilution with MIlli Q water and centrifugation
at 4 000 r min–1 for 20 min to remove rumen debris. The
supernatant was then filtered through a 0.22-µm sterile
filter for further elimination of small particles. For the HPLC
analysis, the mobile phase comprised acetonitrile 45%
(solvent A) and 2.5 mmol L–1 methanesulfonic acid (solvent
B) (Sigma-Aldrich Pty Ltd., Australia), with the temperature
set at 30°C, utilizing the following gradient over a total run
time of 16 min: initially 100% solvent B holding for 1 min
followed by an increase to 45/55% solvent A/B over 11 min
and held isocratically for a further 4 min at a flow rate of
1.0 mL min–1. Compounds in the eluent were detected with
a UV-diode-array detector (HP 1040 M, Hewlett-Packard,
USA), set at 210 nm. The acid concentrations within the
samples were determined by extrapolation of chromatogram
peak areas against calibration curves generated from
duplicate analysis of certified analytical standards. In all
cases instrument function was verified by inclusion of adipic
acid (hexanedioic acid) (Sigma-Aldrich Pty Ltd., Australia)
as an internal standard.
2.5. Statistical analysis
Data were analysed using a Generalized Linear Model pro-
cedure in SPSS/PASW statistical software ver. 18 (SPSS
Inc., Chicago, Il, USA). One-way ANOVA was used for
comparison of the pH between control and treatments.
One-way ANOVA was also applied to comparison of the
difference between control and treatments for the ratios of
acetic and propionic acid production, and the ratio of these
products to lactic acid consumption. The changes in acid
concentrations during the incubation period (0–72 h) were
analysed for main effects using repeated measure ANOVA.
In all cases, the Bonferroni post hoc test was performed
for means comparison. T-tests were also applied for com-
parison between initial and final average pH value in the
multi-strains study. For all statistical analyses, probability
values less than 0.05 were considered significant.
3. Results
3.1. Multi-strain study
The aims of the multi-strain comparison, in which separate
preparations of rumen fluid were fortified with lactic acid to
simulate a condition of acidosis and then each inoculated
with one of seven different strains of Propionibacterium,
were to determine: whether inoculation with propionibacteria
would result in a significant reduction in lactic acid concen-

Table 2 Sample sets prepared for analysis of the effect of the dosage level of propionibacteria on lactic acid metabolism, under
simulated rumen conditions
Set Treatment1)
Set 1 (non-acidified rumen sample)
A Control, 300 g rumen fluid sample+1 mL Milli Q water only
B 1 mL 105 cfu mL–1 PJ 702+300 g rumen fluid sample
C 1 mL 1010 cfu mL–1 PJ 702+300 g rumen fluid sample
Set 2 (acidified rumen sample)
D 1 mL 105 cfu mL–1 PJ 702+30 mmol L–1 lactic acid+300 g rumen fluid sample
E 1 mL 1010 cfu mL–1 PJ 702+30 mmol L–1 lactic acid+300 g rumen fluid sample
F Control, 30 mmol L–1 lactic acid+300 g rumen fluid sample
1)
PJ 702, Propionibacterium jensenii 702.
1570 Jianbiao Luo et al. Journal of Integrative Agriculture 2017, 16(7): 1566–1575

tration and elevation in pH, and if so, whether all strains of Control PJ 702 PA 341
Propionibacterium would be similarly effective. PA 25562 PT 4874 PF 2206
Immediately following the addition of lactic acid and PF 2207 PJ 572
A 90
inoculation with propionibacteria, the average pH of the
rumen samples was found to be decreased from 7.53 to 75

Lactic acid (mmol L–1)

6.47. The initial and final (post incubation) pH of each
60
rumen sample is presented in Fig. 1, showing an increase
in pH (P<0.001) for each of the strain preparations, with the 45

overall final average pH being 6.70. No significant difference 30

between strain preparations was observed either in the initial
15
or final pH measurement.
Prior to the experimental addition of lactic acid, no lactate 0
0 20 40 60 80
was detected in any of the rumen samples. After inoculation,
Time (h)
the initial (Time 0) average concentration of lactic acid across B 50
all preparations was found to be 75.92 mmol L–1. Variation

Propionic acid (mmol L–1)

between preparations for the initial lactic acid concentration 40
was not statistically significant. However, in addition to
30
significant (P<0.001) reductions in lactic acid concentration
in all preparations during the 72 h incubation period, a 20
significant difference (P<0.001) between the trajectories
of lactic acid reduction curves of the various preparations 10
was also evident (Fig. 2-A). All inoculated rumen samples
0
exhibited a more rapid reduction of lactic acid concentration 0 10 20 30 40 50 60 70 80
by comparison with the control. Samples containing PJ 702, Time (h)
PA 341, PF 2207 and PJ 572 exhibited a similar pattern of C 45
lactate consumption, with lower concentrations recorded 40
Acetic acid (mmol L–1)

than for the other preparations at each measurement point.

35
In these four samples complete consumption of lactic acid
had been reached after 48 h, while this situation was not 30

apparent for the PA 25562, PT 4874 and PF 2206 samples 25

20
7.5
Initial pH 15
Final pH 10
7.0
0 10 20 30 40 50 60 70 80
Time (h)
pH

6.5
Fig. 2 Change in lactic (A), propionic (B) and acetic (C) acid
6.0 concentrations in the various simulated rumen samples during
72 h incubation. PJ 702, Propionibacterium jensenii 702;
PA 341, Propionibacterium acidopropionici 341; PA 25562,
5.5
P. acidopropionici ATCC 25562; PT 4874, Propionibacterium
l

74

06

07

2
tro

thoenii ATCC 4874; PF 2206, Propionibacterium freudenreichii

70

34

56

57
48

22

22
on

25
PA
PJ

PJ

CSCC 2206; PF 2207, P. freudenreichii CSCC 2207; PJ 572,

PT
C

PF

PF
PA

Inoculated Propionibacterium strain P. jensenii NCFB 572. Each point represents the mean of the
measured values of the replicates (n=3).

Fig. 1 Average initial and final pH values for each of the

rumen samples in each group (n=3). pH measurements
were taken at t=0 and t=72 h after inoculation. PJ 702,
Propionibacterium jensenii 702; PA 341, Propionibacterium
acidopropionici 341; PA 25562, P. acidopropionici ATCC
25562; PT 4874, Propionibacterium thoenii ATCC 4874; PF
2206, Propionibacterium freudenreichii CSCC 2206; PF 2207,
P. freudenreichii CSCC 2207; PJ 572, P. jensenii NCFB 572.
Data are means±SD.
until the 72 h measurement. In the control sample, the
lactic acid concentration had reduced by 64.88% after 72 h.
When compared at the 24 h time point, the PJ 702, PA 341,
PF 2207 and PJ 572 preparations showed a significantly
greater reduction in lactate levels (72% on average) than
that observed for the PA 25562 (43.2%), PF 2206 (14.1% )
 Jianbiao Luo et al. Journal of Integrative Agriculture 2017, 16(7): 1566–1575 1571

and PT 4874 (8.0%) preparations.
The initial average concentration of propionic and acetic
acid across all preparations was 1.54 and 15.06 mmol L–1
respectively. During the incubation period, marked increase
in propionic and acetic acid concentration were evident in
all rumen samples (P<0.001) with all inoculated samples
exhibiting higher concentrations of propionic and acetic
acid by comparison with the control. For each preparation,
both the propionic and acetic acid concentration curves
(Fig. 2-B and C) were found to exhibit a similar pattern,
however, significant differences (P<0.001) were evident
between the concentration curves of the various treatments.
Samples inoculated with PJ 702, PA 341, PF 2207 and
PJ 572 all reached a maximum propionic and acetic acid
concentration at 48 h at an average of 41.46 and 38.17 mmol
L–1, respectively. Samples inoculated with PA 25562,
PF 2206 and PT 4874 exhibited lower concentrations during
the incubation, with all reaching the maximum concentration
at the final measurement of 34.80, 35.29 and 28.46 mmol
L–1 for propionic acid respectively, and 33.40, 32.50 and
31.48 mmol L–1 for acetic acid. The control sample produced
least propionic and acetic acid during the incubation with
the maximum concentrations of 13.15 and 27.96 mmol L–1
respectively.
The relationships between the production of acetic
and propionic acid and consumption of lactic acid in each
treatment were also investigated (Table 3). For each sample
treatment the maximum production of acetic and propionic
acid was calculated by subtracting the initial concentration
from the maximum concentration of the acids reached during
the incubation period. Similarly, the maximum consumption
of lactic acid was calculated by subtracting the lowest (final)
concentrations from the highest (initial) for each treatment.
Reduction in lactic acid levels and production of acetic and
propionic acid was greater in all inoculated treatments than
in the control. Between Propionibacterium treatments, there
were significant differences in the production of acetic acid
(P=0.04) and propionic acid (P=0.04), but not for total lactic
acid consumption (P=0.92). In the control, more acetic acid
was produced than propionic acid while the opposite was
observed for the Propionibacterium treatments. The ratio
of the maximum acetic acid production to propionic acid
production (A/P) ranged from 0.50 to 0.63. The maximum
conversion ratio of acetic acid and propionic acid to lactic
acid ((A+P)/L) can be represented in terms of the ratio of the
production of acetic or propionic acid, or their sum, to the
maximum consumption of lactic acid, as indicated in Table 3.
A significant difference between groups was observed for the
maximum conversion ratio of acetic acid to lactic acid (A/L,
P=0.002), propionic acid to lactic acid (P/L, P<0.001) and
their sum to lactic acid ((A+P)/L, P<0.001). The strain PJ 702
preparation exhibited the highest conversion ratio of lactic
acid to both propionic and acetic acid, and the control the
lowest conversion ratio of P/L and (A+P)/L. Strain PT 4874
showed the lowest conversion ratio of lactic acid among the
Propionibacterium treatments.
3.2. Dosage study
In order to determine the potential impact of varying
inoculum dose level on the rate of consumption of lactic acid,
acid levels were monitored following inoculation of acidified
and non-acidified rumen fluid samples with PJ 702 at two
different dose levels. Strain PJ 702 was chosen for this
study on the basis that it produced the most rapid reduction
of lactic acid concentration and highest concentration of
propionic acid in the multi-strain study.
As expected, the pH of samples (treatments D, E, and
F, all treatments shown in Table 2) fortified with lactic acid
was considerably lower throughout the experiment (initial
average pH 6.5, and final average 6.7) than the non-acidified
sample (treatments A, B, and C), for which the average initial

Table 3 Maximum production of acetic and propionic acid, reduction of lactic acid concentration, and the relative ratios of each
in the various inoculated treatments1)
Inoculums2) Amax (mmol L–1) Pmax (mmol L–1) Lmax (mmol L–1) A/P A/L P/L (A+P)/L
Control 13.62++ 11.19++ 48.13++ 1.22+ 0.28 0.23++ 0.52++
PJ 702 24.63+ 39.71+ 73.90 0.62 0.33+ 0.54+ 0.87+
PA 341 21.04 39.26 74.04 0.52 0.28 0.53 0.81
PA 25562 17.72 32.56 78.90+ 0.54 0.23 0.41 0.64
PT 4874 16.71 28.47 77.62 0.59 0.22++ 0.37 0.58
PF 2206 17.23 34.50 78.68 0.50++ 0.22 0.44 0.66
PF 2207 23.27 38.81 74.74 0.57 0.31 0.52 0.83
PJ 572 23.70 37.76 74.80 0.63 0.32 0.50 0.82
1)
Amax, the maximum production of acetic acid; Pmax, the maximum production of propionic acid; Lmax, the maximum reduction of lactic
acid; A/P, ration of the maximum acetic acid production to propionic acid production; A/L, conversion ratio of acetic acid to lactic acid;
P/L, conversion ratio of propionic acid to lactic acid; (A+P)/L, total conversion ratio of acetic and propionic acid to lactic acid.
2)
PJ 702, Propionibacterium jensenii 702; PA 341, Propionibacterium acidopropionici 341; PA 25562, P. acidopropionici ATCC 25562;
PT 4874, Propionibacterium thoenii ATCC 4874; PF 2206, Propionibacterium freudenreichii CSCC 2206; PF 2207, P. freudenreichii
CSCC 2207; PJ 572, P. jensenii NCFB 572.
+
, the highest value between groups; ++, the lowest value between groups.
1572 Jianbiao Luo et al. Journal of Integrative Agriculture 2017, 16(7): 1566–1575

pH was 7.68, decreasing slightly to a final average pH of A 100 D

7.49 (results not shown). No lactic acid was detected in any E

Lactic acid (mmol L–1)

80
of the non-acidified preparations (treatments A, B, and C) at F
any time point during the experiment. Among the acidified 60
preparations (treatments D, E, and F), the average initial
concentrations of lactic acid were 95.05, 90.26 and 94.20 40

mmol L–1, respectively. A significant (P<0.001) decrease in 20

lactic acid concentration across the experimental period was
observed in all three cases (Fig. 3-A). Treatments D and E 0
0 20 40 60 80
exhibited more rapid decline in lactic acid concentration than Time (h)
the control (treatment F) (P<0.001). The overall decrease in B 40
C
the lactate concentration of the control was 49.65%, while

Propionic acid (mmol L–1)

D
lactate could not be detected at all in the lower dosage 30
E
treatment (D) after 72 h, or in the higher dosage treatment
F
(E) after 48 h. 20
Propionic acid was not detected at Time 0 in any of the
preparations, and no detectable propionic acid appeared 10
at any point of time during the experimental period in
0
preparations A and B. However, steady increases in 0 20 40 60 80
propionic acid concentrations were observed in treatments Time (h)
C, D, E and F (Fig. 3-B). Significant differences (P<0.001) C 50
A D
were found between the propionic acid accumulation curves B E
Acetic acid (mmol L–1)

of these preparations. Preparations D and E showed more 40 C F

rapid increases in propionic acid concentration than the
control (F), with the high dose inoculation (E) producing more
propionic acid than the low dose treatment (D). Treatment
C which had no addition of lactic acid but a high dosage
inoculation of PJ 702 exhibited an increase in propionic
acid concentration, albeit lower than in the lactate enriched
treatments (D, E, and F).
The average initial acetic acid concentration for
preparations A, B and C was 12.57 mmol L–1. There was
a continuous reduction of acetic acid concentration in this
group of preparations during the incubation (Fig. 3-C)
and the final reduction rates were: 69.70, 71.78 and
74.87%, respectively. Differences between the acetic
acid concentration curves of these preparations was not
statistically significant (P=0.971). In contrast, PJ 702
preparations which had lactic acid fortification (treatments
D and E), exhibited marked increases in acetic acid
concentrations (Fig. 3-C). Significant differences in the
acetic acid concentration curves (P<0.001) were evident
between these three preparations, with the higher dose
inoculation treatment (E) exhibiting higher acetic acid
concentrations than the lower dose treatment (D).
4. Discussion
4.1. Survival and metabolic activity of propionibacb-
teria in rumen fluid
A successful DFM candidate must both survive within the
30
20
10
0
0 20 40 60 80
Time (h)
Fig. 3 Change in lactic (A), propionic (B) and acetic (C) acid
concentration of the various simulated rumen samples in the
direct-fed microbials (DFM) dosage study during the 72 h
incubation period. Treatments: A, control, 300 g rumen fluid
sample+1 mL Milli Q water only; B, 1 mL 105 cfu mL–1 PJ
702+300 g rumen fluid sample; C, 1 mL 1010 cfu mL–1 PJ
702+300 g rumen fluid sample; D, 1 mL 105 cfu mL–1 PJ 702+
30 mmol L–1 lactic acid+300 g rumen fluid sample; E, 1 mL
1010 cfu mL–1 PJ 702+30 mmol L–1 lactic acid+300 g rumen fluid
sample; F, control, 30 mmol L–1 lactic acid+300 g rumen fluid
sample. Each point represents the mean value of replicate
measurements (n=3).
target environment, and be able to carry out the metabolic
activities necessary to provide the host animal with the de-
sired benefits (Ranadheera et al. 2014). Convenient and
efficient methods for isolation, enumeration and cultivation of
live propionibacteria in the rumen are still in the developing
stage. The large number and variety of microorganisms in
the rumen, and the sensitivity of propionibacteria to many
conventional antibiotics, has made the possibility of design-
ing and using selective media for this task extremely difficult.
 Jianbiao Luo et al. Journal of Integrative Agriculture 2017, 16(7): 1566–1575
Therefore, alternative methods must be used to examine
the survivability and activity of inoculated propionibacteria
in the rumen environment. In light of this, the metabolic
substrates and products which propionibacteria readily use
and express, currently represent the reasonable indicators
of their survival and metabolic activity.
The results of the current study have shown that the
concentrations of such substrates (lactic acid) and products
(propionic acid) in the rumen samples were significantly
decreased or increased following inoculation with propi-
onibacteria, when compared with control preparations.
Furthermore, a strong correlation between dosage level and
alteration of the acid profile was observed. Similar evidence
of the active metabolism of propionibacteria, in the form of
modifications to the fatty acid profile and other alterations to
rumen fermentation dynamics, has been reported in other
studies following either direct feeding to cattle, or in vitro
application of different strains of Propionibacterium to rumen
fluid (Francisco et al. 2002; Ghorbani et al. 2002; Stein et al.
2006; Raeth-Knight et al. 2007; Lehloenya et al. 2008).
In this context, the current findings reinforce the potential
capacity of these bacteria to survive and actively influence
metabolic equilibria within the rumen.
Although variation in the metabolic capacity of different
strains of Propionibacterium can be significant (Narvaez
et al. 2014), a common feature of most previous studies
has been the focus on one particular strain rather than com-
parisons of performance between strains. In the complex
rumen environment, metabolic activity would be likely to
vary greatly between different strains. The results presented
here demonstrate clear variation in metabolism between
different strains of Propionibacterium in the simulated
rumen samples, with several strains producing more rapid
consumption of lactic acid and higher rates of production
of propionic and acetic acid than others. It is logical that
such metabolic variation between strains would also occur
in vivo, thus careful investigations and comparisons of strain
physiology under targeted environmental conditions are
likely to be necessary in identifying strains most appropriate
for application in the rumen.
4.2. Influence of propionibacteria inoculation on
rumen pH
While the pH levels of the simulated rumen samples in the
present study (ca. pH 6.5) were not fully representative of a
severe ruminal acidosis condition (i.e., pH<5.5) even after
addition of lactic acid, elevations of pH following inoculation
with dairy propionibacteria were observed. However, the
results indicated that there was neither any significant
difference in pH between the treatment samples and the
control, nor any effect on pH with variation in inoculum dosage
1573
level, suggesting that the observed effect on pH was not
specifically related to enhanced conversion of lactic acid to
propionic and acetic acid by the introduced propionibacteria.
4.3. In vitro acid metabolism in rumen fluid samples
Lactic, acetic and propionic acid In a healthy rumen
system, lactic acid produced by microorganisms during
fermentation is immediately removed by lactic acid utilising
bacteria. Thus, the amount of lactic acid accumulated in the
rumen is low. In treating ruminal lactic acidosis, reduction
of lactic acid levels is a priority, which was achieved here
following inoculation with dairy propionibacteria. Importantly,
significant differences were observed between the capacities
of the various strains examined to reduce lactic acid levels
in the rumen fluid medium (Fig. 2-A). The rumen fluid used
in the present study presumably contained a large number
of the indigenous rumen microflora and the inoculated
propionibacteria were almost certainly exposed to a dynamic
fermentation system. It is likely therefore that competition
and inhibition between microorganisms may have had a
significant differential influence on the proliferation and
metabolism of the Propionibacterium strains.
The total molar quantities of maximum acetic and
propionic acid produced were equal to 52 to 87% of the lactic
acid reduced in the rumen samples between different strains.
Thus the majority of the lactic acid added to the rumen
sample was converted to propionic and acetic acid when the
Propionibacterium were introduced. This conversion of lactic
acid to acetic and propionic acid represents the fundamental
rationale for the application of dairy propionibacteria in the
treatment of ruminal acidosis. As opposed to neutralizing
excess lactic acid with bicarbonates or using antibiotics
to suppress the lactate producing bacteria, the current
approach has demonstrated a successful conversion of
harmful lactic acid to the useful energy sources of acetic
and propionic acid in a rumen environment.
In several related in vivo studies however, observed
results were not entirely consistent with those reported
here. Stein et al. (2006) investigated the application of
Propionibacterium strain P169 to fistulated dairy Holstein
cows for 25 weeks starting from two weeks before
parturition. A significantly higher (17%) molar percentage
of the rumen propionate was exhibited in a high dosage
treatment group compared with a control group, however,
the molar percentage of the rumen acetate was not affected
by the treatment. Similarly, Weiss et al. (2008) reported
higher percentages of propionate in the treatment group than
the control in Holstein cows when Propionibacterium strain
P169 was fed at both early postpartum and established
lactation time points. However, concomitant decreases in
acetate were also recorded. For Angus×Hereford steers,
1574 Jianbiao Luo et al. Journal of Integrative Agriculture 2017, 16(7): 1566–1575
Lehloenya et al. (2008) reported that the molar proportion of
propionate in the rumen was increased 9.7% in the treatment
group fed P169 within the diet. The molar proportion of
acetate tended to decrease, along with the acetate:propionate
ratio. Ghorbani et al. (2002) investigated the application of
Propionibacteria P15 in cannulated steers with and without
coincident application of Enterococcus faecium EF212, but
found no significant influence on the rumen propionate and
acetate concentrations, while Raeth-Knight et al. (2007) found
no effect on acetate and propionate in the rumen by applying
P. freudenreichii strain PF 24, with or without L. acidophilus
strain LA 747, to Holstein dairy cows in mid-lactation.
The apparent inconsistent efficacy of the application of
dairy propionibacteria in live cattle with regard to acetate and
propionate concentrations in the rumen, may be a reflection
of rapid absorption of these components through the ruminal
wall. Acetic and propionic acid are energy sources for cattle
and are readily absorbed through the ruminal wall in normal
healthy animals, and this absorption of acetic and propionic
acid is related to its availability in the rumen (Huntington
et al. 1983; Peters et al. 1990). Therefore, the more acetic
and propionic acid produced in the rumen the greater the
absorption, thus the concentration of these acids tends to
remain at a relatively constant level in the rumen. Hence,
increased concentrations of acetate and propionate as
a result of the application of propionibacteria would not
necessarily be evident in in vivo studies, where simultaneous
absorption from the rumen is occurring. However, more
verified trials are needed in animal production.
In the present dosage study, lower acetic and propionic
acid concentrations were exhibited in treatments which
were inoculated with strain PJ 702 but were not fortified with
lactic acid, compared with the treatments which received
both. The comparison of the acid profiles in preparations
with or without the addition of lactic acid demonstrated that
the additional lactic acid was the key to higher production
of acetic acid and propionic acid. When lactate was
available, the metabolism of the propionibacteria was active,
otherwise the production of acetic and propionic acid was
markedly reduced. In the normal functioning rumen, added
propionibacteria would likely remain inert and not disturb
the fermentation process. However, once the fermentation
balance is lost and excess lactic acid is produced, the
metabolism of the added propionibacteria would result in
excessive lactic acid being converted to acetic and propionic
acid, and potentially absorbed as an energy source for the
animal. Hence, propionibacteria may potentially be effective
in reducing lactic acid accumulation in the rumen due to
acute or subacute ruminal acidosis.
Dosage effects When rumen samples were not fortified
with additional lactic acid, no significant effect on pH or
the concentration of acetic and lactic acid was observed,
regardless of dosage. The exception was that propionic
acid concentration levels increased during incubation in
the high dose preparations. Stein et al. (2006) also re-
ported that varying dose levels of Propionibacterium strain
169 inoculated into Holstein cows increased the rumen
propionate concentrations. In their study, milk fat and milk
lactose levels were also influenced by the application of
propionibacteria and its dosage. These findings and those
of the present study both indicate that the dosage level is
important for the potential application of particular strains
of Propionibacterium as direct-fed microbes in ruminants.
In the lactic acid fortified groups, dosage level was clearly
a significant factor. In the higher dose preparation lactic acid
concentration declined more rapidly and reached undetect-
able levels 24 h earlier than in the lower dose treatment,
while acetic and propionic acid levels were both markedly
higher. For the control of lactic acidosis in cattle, the rate at
which lactic acid levels can be reduced is obviously critical.
These findings indicate that a relatively high dose inoculation
may be required for potential reduction of lactic acid with
propionibacteria in the affected rumen.
5. Conclusion
All seven strains of Propionibacterium were able to remain
metabolically active in rumen fluid in vitro. Excessive lactic
acid levels in the simulated rumen content were reduced
following the introduction of propionibacteria and significant
differences were apparent between the strains in terms of
their apparent metabolic activity in the rumen samples.
Furthermore, the efficacy of dairy propionibacteria in re-
ducing lactic acid levels was seemingly dosage dependent.
Apparently, high dosage inoculation of propionibacteria can
result in efficient conversion of lactic acid to acetic and pro-
pionic acid. However, careful selection of Propionibacterium
strains would appear crucial to potentially successful appli-
cation in reducing lactic acid levels. Further investigations
are also necessary to elucidate their efficiency in reducing
lactic acid levels in the ruminal acidosis conditions in vivo.
Acknowledgements
The University of Newcastle, Australia for financial support.
References
Adams M C, Luo J, Rayward D, King S, Gibson R, Moghaddam
G H. 2008. Selection of a novel direct-fed microbial to
enhance weight gain in intensively reared calves. Animal
Feed Science and Technology, 145, 41–52.
Cousin F J, Mater D D, Foligné B, Jan G. 2011. Dairy
propionibacteria as human probiotics: A review of recent
 Jianbiao Luo et al. Journal of Integrative Agriculture 2017, 16(7): 1566–1575
evidence. Dairy Science & Technology, 91, 1–26.
Dawson K A, Rasmussen M A, Allison M J. 1997. Digestive
disorders and nutritional toxicity. In: Hobson P J, Stewart
C S, eds., The Rumen Microbial Ecosystem. Blackie
Academic & Professional, London. pp. 633–660.
Dirksen G. 1985. The rumen acidosis complex - recent
knowledge and experiences (1). A review. Tierarztliche
Praxis, 13, 501–512.
Elghandour M M, Salem A Z, Castañeda J S M, Camacho L M,
Kholif A E, Chagoyán J C V. 2015. Direct-fed microbes: A
tool for improving the utilization of low quality roughages in
ruminants. Journal of Integrative Agriculture, 14, 526–533.
Enemark J M D. 2008. The monitoring, prevention and treatment
of sub-acute ruminal acidosis (SARA): A review. Veterinary
Journal, 176, 32–43.
Enemark J M, Jorgensen R J, Kristensen N B. 2004. An
evaluation of parameters for the detection of subclinical
rumen acidosis in dairy herds. Veterinary Research
Communications, 28, 687–709.
Francisco C C, Chamberlain C S, Waldner D N, Wettemann R P,
Spicer L J. 2002. Propionibacteria fed to dairy cows: Effects
on energy balance, plasma metabolites and hormones, and
reproduction. Journal of Dairy Science, 85, 1738–1751.
Ghorbani G R, Morgavi D P, Beauchemin K A, Leedle J A.
2002. Effects of bacterial direct-fed microbials on ruminal
fermentation, blood variables, and the microbial populations
of feedlot cattle. Journal of Animal Science, 80, 1977–1985.
Huang Y, Adams M C. 2004. In vitro assessment of the upper
gastrointestinal tolerance of potential probiotic dairy
propionibacteria. International Journal of Food Microbiology,
91, 253–260.
Huntington G B, Reynolds P J, Tyrrell H F. 1983. Net absorption
and ruminal concentrations of metabolites in nonpregnant
dry Holstein cows before and after intraruminal acetic acid
infusion. Journal of Dairy Science, 66, 1901–1908.
Hutton P G, Durmic Z, Ghisalberti E L, Flematti G R, Duncan
R M, Carson C F, Riley T V. 2012. Inhibition of ruminal
bacteria involved in lactic acid metabolism by extracts from
Australian plants. Animal Feed Science and Technology,
176, 170–177.
Kung L, Hession A O. 1995. Preventing in vitro lactate
accumulation in ruminal fermentations by inoculation with
Megasphaeraelsdenii. Journal of Animal Science, 73,
250–256.
Koniarova I. 1993. Biochemical and physiologic properties of
strains of Propionibacterium acnes isolated from the rumen
of calves and lambs. Veterinarni Medicina, 38, 43–52.
Lehloenya K V, Krehbiel C R, Mertz K J, Rehberger T G, Spicer
L J. 2008. Effects of propionibacteria and yeast culture fed
to steers on nutrient intake and site and extent of digestion.
Journal of Dairy Science, 91, 653–662.
Mackie R I, White B A. 1990. Recent advances in rumen
microbial ecology and metabolism: Potential impact on
nutrient output. Journal of Dairy Science, 73, 2971–2995.
Martin S A, Streeter M N. 1995. Effect of malate on in vitro
1575
mixed ruminal microorganism fermentation. Journal of
Animal Science, 73, 2141–2145.
Nagaraja T G, Titgemeyer E C. 2007. Ruminal acidosis in beef
cattle: The current microbiological and nutritional outlook.
Journal of Dairy Science, 90, E17–E38.
Narvaez N, Alazzeh A Y, Wang Y, McAllister T A. 2014. Effect
of Propionibacterium acidipropionici P169 on growth
performance and rumen metabolism of beef cattle fed a
corn- and corn dried distillers’ grains with solubles-based
finishing diet. Canadian Journal Animal Science, 94,
363–369.
Nocek J E, Kautz W P, Leedle J A, Allman J G. 2002. Ruminal
supplementation of direct-fed microbials on diurnal pH
variation and in situ digestion in dairy cattle. Journal of Dairy
Science, 85, 429–433.
Owens F N, Secrist D S, Hill W J, Gill D R. 1998. Acidosis in
cattle: A review. Journal of Animal Science, 76, 275–286.
Parrot T D, Rehberger T G, Owens F N. 2001. Selection of
Propionibacterium strains capable of utilizing lactic acid from
in vitro models. Journal of Animal Science, 1(Suppl.), 79–80.
Pascal P. 1999. Metabolism of lactate and sugars by dairy
propionibacteria: A review. LAIT, 79, 23–41.
Peters J P, Shen R Y, Robinson J A, Chester S T. 1990.
Disappearance and passage of propionic acid from the
rumen of the beef steer. Journal of Animal Science, 68,
3337–3349.
Plaizier J C, Krause D O, Gozho G N, McBride B W. 2008.
Subacute ruminal acidosis in dairy cows: the physiological
causes, incidence and consequences. Veterinary Journal,
176, 21–31.
Ranadheera C S, Evans C A, Adams M C, Baines S K. 2012. In
vitro analysis of gastrointestinal tolerance and intestinal cell
adhesion of probiotics in goat’s milk ice cream and yogurt.
Food Research International, 49, 619–625.
Ranadheera C S, Evans C A, Adams M C, Baines S K.
2014. Effect of dairy probiotic combinations on in vitro
gastrointestinal tolerance, intestinal epithelial cell adhesion
and cytokine secretion. Journal of Functional Foods, 8,
18–25.
Raeth-Knight M L, Linn J G, Jung H G. 2007. Effect of direct-
fed microbials on performance, diet digestibility, and rumen
characteristics of Holstein dairy cows. Journal of Dairy
Science, 90, 1802–1809.
Seo J K, Kim S W, Kim M H, Upadhaya S D, Kam D K, Ha J
K. 2010. Direct-fed microbials for ruminant animals. Asian-
Australasian Journal of Animal Sciences, 23, 1657–1667.
Stein D R, Allen D T, Perry E B, Bruner J C, Gates K W,
Rehberger T G, Mertz K, Jones D, Spicer L J. 2006. Effects
of feeding propionibacteria to dairy cows on milk yield, milk
components, and reproduction. Journal of Dairy Science,
89, 111–125.
Weiss W P, Wyatt D J, McKelvey T R. 2008. Effect of feeding
propionibacteria on milk production by early lactation dairy
cows. Journal of Dairy Science, 91, 646–652.
(Managing editor ZHANG Juan)