FEMS Microbiology Ecology 32 (2000) 121^128
www.fems-microbiology.org
Mathematical estimations of hyper-ammonia producing ruminal
bacteria and evidence for bacterial antagonism that decreases
ruminal ammonia production1
Jennifer L. Rychlik a , James B. Russell
b
b;
*
a
Section of Microbiology, Cornell University, Wing Hall, Ithaca, NY 14853, USA
Agricultural Research Service, U.S. Department of Agriculture, Ithaca, NY 14853, USA
Received 21 September 1999 ; received in revised form 6 March 2000; accepted 6 March 2000
Abstract
Mixed ruminal bacteria (MRB) from cattle fed hay produced ammonia from protein hydrolysate twice as fast as MRB from cattle fed
mostly grain, and a mathematical model indicated that cattle fed hay had approximately four-fold more hyper ammonia-producing ruminal
bacteria (HAB). HAB had a high maximum velocity of ammonia production (Vmax ) and low substrate affinity (high Km ), but simulations
indicated that only large changes in Vmax or Km would cause a large deviation in HAB numbers. Some carbohydrate-fermenting ruminal
bacteria produced ammonia at a slow rate (CB-LA), but many of the isolates had almost no activity (CB-NA). The model indicated that the
ratio of CB-LA to CB-NA had little impact on HAB numbers. Validations based on predicted ratios of HAB, CB-LA and CB-NA overpredicted the specific activity of ammonia production by MRB, but co-culture incubations indicated that washed MRB from cattle fed grain
could inhibit HAB. Because autoclaved MRB had virtually no effect on HAB and the incubations were always carried out at pH 7.0, the
inhibition was not simply a chemical effect (e.g. low pH). Published by Elsevier Science B.V. All rights reserved.
Keywords : Ruminal bacterium; Ammonia production; Hyper-ammonia producing bacterium; Protein
1. Introduction
In ruminant animals, as much as 50% of the dietary
protein can be converted to ammonia by microorganisms
[1]. Some ammonia is utilized as a bacterial nitrogen
source ; however, rates of ammonia production often exceed rates of ammonia utilization [2]. Excess ammonia is
absorbed into blood, converted to urea by the liver, and
excreted. Because ruminant nitrogen excretion is a process
that increases the cost of production and causes groundwater contamination [3], nutritionists have sought ways of
decreasing wasteful ruminal ammonia production [4].
Insoluble proteins are not easily degraded by rumen
bacteria, but the use of `bypass' or `escape' proteins is
* Corresponding author. Tel. : +1 (607) 255-4508;
Fax: +1 (607) 255-3904; E-mail: [email protected]
1
Mandatory disclaimer : `Proprietary or brand names are necessary to
report factually on available data; however, the USDA neither guarantees
nor warrants the standard of the product, and the use of the name by the
USDA implies no approval of the product, and exclusion of others that
may be suitable.'
constrained by the increased cost of these supplements.
Ruminal ammonia utilization can be enhanced by adding
starch to the ration, but starch can cause ruminal acidosis,
acute indigestion and even death of the animal [5]. The
ruminal additive, monensin, can decrease ruminal ammonia in vitro and in vivo [6], but recent work indicated that
monensin was not able to decrease ruminal ammonia
when alfalfa hay was the dominant forage [7].
Some carbohydrate-fermenting bacteria produce ammonia, but these bacteria generally have low speci¢c activities
of amino acid deamination and cannot utilize amino acids
as a sole energy source for growth [8]. Hyper-ammonia
producing ruminal bacteria (HAB), identi¢ed as Peptostreptococcus anaerobius, Clostridium sticklandii, and Clostridium aminophilum [9], generate ammonia at a high rate
[10], and recent work by New Zealand researchers indicated that the rumen had additional species of HAB [11].
The following experiments were designed to: (1) compare the ammonia production of mixed ruminal bacteria
(MRB) with the deamination rates of HAB and carbohydrate-fermenting ruminal bacteria, (2) develop a mathematical model of ammonia production by various classes
of ammonia-producing ruminal bacteria, and (3) estimate
0168-6496 / 00 / $20.00 Published by Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 8 - 6 4 9 6 ( 0 0 ) 0 0 0 2 1 - 0
FEMSEC 1122 10-5-00
122
J.L. Rychlik, J.B. Russell / FEMS Microbiology Ecology 32 (2000) 121^128
the numbers of HAB in the rumen of cattle fed hay or
grain-based diets.
2. Materials and methods
2.1. Basal medium
Ruminal bacteria were grown anaerobically at 39³C in
a basal medium containing (per liter) : 292 mg
K2 HPO4 3H2 O, 240 mg KH2 PO4 , 480 mg (NH4 )SO4 ,
480 mg NaCl, 100 MgSO4 7H2 O, 64 mg CaCl2 2H2 O,
0.5 Yeast Extract (Difco Laboratories, Detroit, MI,
USA), 0.6 g cysteine, 1.0 g Trypticase (BBL Microbiology
Systems, Cockeysville, MD, USA), 0.5 g yeast extract, 4 g
Na2 CO3 , and fatty acids [12]. Previously isolated ruminal
HAB were grown anaerobically in medium containing (per
liter): 292 mg K2 HPO4 3H2 O, 240 mg KH2 PO4 , 480 mg
Na2 SO4 , 480 mg NaCl, 100 MgSO4 7H2 O, 64 mg
CaCl2 2H2 O, 0.6 g cysteine, 4 g Na2 CO3 , vitamins and
minerals [12]. P. anaerobius C was grown in medium containing 7.5 mg ml31 each of Casamino acids and Trypticase, C. sticklandii SR was grown on 15 mg ml31 Trypticase, and C. aminophilum F was grown on 15 mg ml31
Casamino acids.
b
b
b
b
b
b
2.2. Isolation of carbohydrate fermenting bacteria
Samples of digest were obtained from two 600-kg ruminally ¢stulated cows that were fed either 100% timothy
hay (14% crude protein, 40% neutral detergent ¢ber) or
90% grain mixture (89% rolled corn and 11% soybean
meal) and 10% hay. The cows were on the diets for 4
weeks prior to sampling. Ruminal contents were squeezed
through four layers of cheesecloth, placed in Erlenmeyer
£asks at 39³C and left undisturbed for 30 min. The feed
particles, which were buoyed up by gas production, rose to
the top of the £ask and protozoa settled on the bottom of
the £ask. MRB from the center of the £ask were serially
diluted and streaked onto basal medium plates containing
(per liter) 1.7 g starch, 0.7 g cellobiose, 0.4 g sucrose, 0.5 g
xylose, 0.5 g arabinose, and 0.2 g pectin and incubated in
an anaerobic glove box (Coy Laboratory Products, Ann
Arbor, MI, USA).
Forty-¢ve colonies of di¡ering morphologies were isolated from each diet, checked microscopically for purity,
and tested for the ability to produce ammonia. None of
these isolates grew in basal medium when the carbohydrate was deleted. Washed cell suspensions of each isolate
were incubated (39³C) with Trypticase in nitrogen free
bu¡er (per liter): 292 mg K2 HPO4 3H2 O, 240 mg
KH2 PO4 , 480 mg Na2 SO4 , 480 mg NaCl, 100
MgSO4 7H2 O, 64 mg CaCl2 2H2 O, 0.6 g cysteine, vitamins and minerals [12]) to determine rates of ammonia
production. Ammonia production was determined by the
colorimetric method of Chaney and Marbach [13]. The
isolates fell into two groups: those that had a speci¢c
activity less than 1 nmol NH3 mg protein31 min31 and
those with speci¢c activities ranging from 2 ^100 nmol
NH3 mg protein31 min31 . The speci¢c activity of the ¢rst
group was so low that the kinetics of ammonia production
could not be determined and these bacteria were designated as no activity, carbohydrate-fermenting bacteria
with essentially no capacity to produce ammonia (CBNA). Based on their ability to produce small amounts of
ammonia, the second group of bacteria were designated
low activity, carbohydrate-fermenting bacteria with low
rates of ammonia production (CB-LA).
2.3. Ammonia production and kinetics
To determine the kinetics of ammonia production, cultures were centrifuged (4000Ug at 22³C), washed with
nitrogen-free bu¡er, and incubated at 39³C in nitrogen
free bu¡er with varying amounts of Trypticase. HAB
were an equal part mix (based on optical density) of
P. anaerobius C, C. sticklandii SR, C. aminophilum F.
CB-LA consisted of four isolates with speci¢c activities
ranging from 10 to 100 nmol NH3 mg protein31 min31 .
MRB were spun at 1000Ug to remove feed particles and
protozoa. The supernatants containing MRB were centrifuged (6000Ug at 22³C), and washed with nitrogen free
bu¡er. The cell densities of MRB were typically 2.0 (1 cm
cuvette, 600 nm, Gilford 260 spectrophotometer). Cultures
were harvested by centrifugation (10 000Ug, 5³C, 10 min),
washed and re-suspended in 0.9% NaCl. Protein of
washed cells was measured by the method of Lowry et
al. [14].
2.4. Validation of the model
To test the model, HAB were mixed with CB-LA and
CB-NA. The HAB percentages were the average values
that were predicted by the model. The ratio of CB-LA
to CB-NA was 1 to 1. The total bacterial optical density
was 1.0. Ammonia production was determined after 2 h of
incubation at pH 7.0 with either 15 or 1.25 mg ml31
Trypticase.
2.5. Bacterial antagonisms
HAB were incubated alone or with washed MRB from
cattle fed hay or grain at pH 7.0. The optical density of
HAB was 0.75 and optical density of the MRB was 2.0.
Washed autoclaved MRB were used as a control. Ammonia production was determined after 2 h.
b
b
2.6. Statistics
b
All measurements were performed in duplicate and the
experimental variation was generally less than 10%. When
the coe¤cient of variation was greater than 10% and the
FEMSEC 1122 10-5-00
J.L. Rychlik, J.B. Russell / FEMS Microbiology Ecology 32 (2000) 121^128
123
standard deviations were large, statistical signi¢cance was
estimated by paired t-tests (P 6 0.05).
3. Results
3.1. Kinetics of MRB
When washed cell suspensions of MRB from cattle fed
hay were incubated in vitro with various concentrations of
Trypticase (0^15 mg ml31 ) at pH 7.0, the initial rates of
ammonia production were substrate-dependent and Lineweaver^Burk plots indicated that the maximal rate (Vmax )
and a¤nity constant (Km ) were 40 nmol NH3 mg
protein31 min31 and 2.8 mg ml31 , respectively (Table 1).
MRB from cows that were fed grain produced ammonia
at a slower rate even though the in vitro pH was 7.0. The
Vmax of MRB from cattle fed grain was only 20 nmol NH3
mg protein31 min31 , and the Km was 4.6 mg ml31 .
3.2. Kinetics of carbohydrate-fermenting ruminal bacteria
When washed cell suspensions of each carbohydrate isolate were incubated with Trypticase to determine rates of
ammonia production, the isolates fell into two groups: 1)
those that had a speci¢c activity less than 1 nmol NH3 mg
protein31 min31 (essentially no activity, CB-NA), and 2)
those with speci¢c activities ranging from 10 to 100 nmol
NH3 mg protein31 min31 CB-LA. The ammonia production of the CB-NA was so low that kinetics could not be
determined, but a mixture of CB-LA isolates had Vmax
and Km values of 34 nmol NH3 mg protein31 min31
and 9.3 mg ml31 , respectively (Table 1).
3.3. Kinetics of HAB
When washed cell suspensions of HAB were incubated
with Trypticase at pH 7.0, the optical density increased,
and the ammonia production was so rapid that it was
possible to estimate initial rates after only 2 h. Lineweaver^Burk plots demonstrated that the HAB had a
Table 1
Michaelis constants of ammonia production by various groups of ruminal bacteria
Bacteriaa
Dietb
MRB
MRB
HAB
CB-LA
Hay
40 þ 2
Grain 20 þ 3
500 þ 13
34 þ 2
a
Vmax
c
Km
d
2.8 þ 0.3
4.6 þ 0.4
7.0 þ 0.5
9.3 þ 0.3
Km Vmax 31
e
0.069
0.23
0.014
0.27
Bacteria are de¢ned as MRB, HAB, and CB-LA.
b
Diet consisted of either 100% hay or 90% grain and 10% hay.
c
Vmax of ammonia production (nmol NH3 mg protein31 min31 ).
d
Substrate a¤nity (Km ) for Trypticase (mg ml31 ) as measured by ammonia production.
e
Ratio of Km to Vmax ((nmol NH3 mg protein31 min31 ) (mg ml31 )31 ).
Fig. 1. Predicted numbers of HAB at varying Trypticase concentrations
for cattle fed either hay (light shading) or grain (dark shading) when
the Trypticase concentration was varied from 0 to 15 mg ml31 . Ranges
were estimated from the kinetic model of ammonia production (see
text).
high Vmax (500 nmol NH3 mg protein31 min31 ), but the
Km was high (7.0 mg ml31 ) (Table 1).
3.4. Kinetic model of ammonia-production
The kinetic model of ammonia production was constructed using four assumptions. The ¢rst assumption
was that the rumen micro£ora could be divided into three
bacterial groups according to their ability to produce
ammonia from amino nitrogen sources. MRB were de¢ned as comprising of CB-NA, CB-LA and HAB bacteria.
The second assumption was that ammonia £ux through
the total population was equal to the sum of ammonia
£uxes (nmol NH3 l31 min31 ) through all three bacterial
groups (£uxCBÿNA +£uxCBÿLA +£uxHAB = £uxMRB ). The
third assumption was that the £ux of ammonia
(nmol l31 min31 ) through a bacterium, bacterial group
or the total population was equal to the product of
bacterial mass (mg protein l31 ) and speci¢c activity
(nmol NH3 mg protein31 min31 ). The fourth assumption
of the model was that the sum of the CB-NA, CB-LA and
HAB had to be equivalent to the mass of MRB
(massCBÿNA +massCBÿLA +massHAB = massMRB ) expressed
as either mg protein l31 or percent of the total.
3.5. Estimation of HAB
Unique solutions were confounded by the fact that
the model had three unknowns (mass of CB-NA, CB-LA
and HAB) and only two equations that could be
solved ( massCBÿNA + massCBÿLA + massHAB = massMRB )
FEMSEC 1122 10-5-00
124
J.L. Rychlik, J.B. Russell / FEMS Microbiology Ecology 32 (2000) 121^128
and (massCBÿNA U SACBÿNA ) + (massCBÿLA U SACBÿLA ) +
(massHAB USAHAB ) = (massMRB USAMRB ). However, it
was possible to set mathematical limits for each equation
and to de¢ne general trends for HAB. The maximal ammonia £ux of MRB from cattle fed hay was so high that
CB-LA could never explain all of the ammonia production, and the HAB were always needed. Because HAB had
an ammonia £ux that was so much greater than the £ux of
the MRB, HAB could never account for a majority of the
MRB. The Vmax of MRB from cattle fed grain was only
20 nmol NH3 mg protein31 min31 , CB-LA could explain
all of the ammonia production, and the HAB were not
essential. However, if CB-NA were present, the activity
of the CB-LA was not su¤cient, and the HAB were still
needed.
By using the Michaelis constants to predict total ammonia £ux at di¡erent Trypticase concentrations, it was possible to assess the impact of amino N availability on the
numbers of HAB in cattle fed either hay or grain. When
only hay was fed, the MRB had a lower Km Vmax 31 than
CB-LA (Table 1), and the CB-LA alone could not account
for ammonia £ux at low Trypticase concentrations, and
HAB accounted for as much as 17% of the population
(Fig. 1). When the Trypticase concentration was increased,
the contribution of Km to ammonia £ux diminished, CBLA could account for a larger fraction of the ammonia
production, and the numbers of HAB declined. The MRB
from cattle fed grain had a higher Km Vmax 31 (Table 1)
than MRB from cattle fed hay, and CB-LA could explain
more of the ammonia production at low Trypticase concentrations ; therefore, HAB constituted never more than
6% of the population (Fig. 1). Again, HAB declined when
Fig. 3. The e¡ect theoretical variations in Vmax (a) and Km (b) on the
predicted numbers of HAB for cattle fed either hay (light shading) or
grain (dark shading). The Trypticase concentration was set at 1.25 mg
ml31 . Vmax in mmol NH3 mg protein31 min31 and Km is mg ml31 .
the Trypticase increased, but the decrease was not as dramatic as the decline observed with the HAB from cattle
fed hay. An increase in CB-NA (relative to CB-LA) increased the need for HAB, but this e¡ect was relatively
small ( 6 3% e¡ect on HAB) regardless of diet (data not
shown).
3.6. Contribution of HAB to total ammonia production
Fig. 2. Predicted minimum values of HAB for cattle fed either hay (solid line) or grain (dotted) when the when the Trypticase concentration
was varied from 0 to 15 mg ml31 . Values were estimated from the kinetic model of ammonia production (see text).
By using the estimates of percent HAB, it was possible
to estimate the minimum contribution of HAB to the
total ammonia production. This analysis was based on
FEMSEC 1122 10-5-00
J.L. Rychlik, J.B. Russell / FEMS Microbiology Ecology 32 (2000) 121^128
125
a comparison of the kinetics of ammonia production
by MRB and the predicted percentage of HAB at various
concentrations of Trypticase [S], vMRB = (Vmax MRB US)6
(Km MRB +S)) and (vHAB = (Vmax HAB US)6(Km HAB +S)U
HAB (%), respectively where vMRB and vHAB are the velocities at di¡erent substrate concentrations. Results indicated that HAB would account for at least 40% of the
ammonia £ux in cattle fed hay but as little as 0% of the
ammonia £ux in cattle fed grain (Fig. 2).
3.7. Vmax and Km sensitivity analyses
When the Vmax of HAB was varied, there was an inverse
relationship between Vmax and the percentage of HAB, but
the greatest change was observed at Vmax values less than
500 nmol NH3 mg protein31 min31 (Fig. 3a). Increasing
Vmax from 500 to 1000 had little impact on the predicted
HAB value for either diet. Theoretical changes in Km had
a smaller impact on the percent of HAB than Vmax , and
the response was linear (Fig. 3b). Changes in Vmax and Km
did not contradict the idea that cattle fed hay should have
more HAB than cattle fed grain.
3.8. Experimental validation
The model was then validated at high (15 mg ml31 ) and
low (1.2 mg ml31 ) concentrations of Trypticase to see if
the predicted percentages of HAB gave ammonia £uxes
Fig. 5. The e¡ect of washed MRB from cattle fed grain or hay on the
speci¢c activity of HAB. The e¡ect of autoclaved MRB is also shown.
The ammonia production of MRB incubated without HAB was subtracted to determine the speci¢c activity HAB in co-culture.
that were similar to the respective MRB. Because the ratio
of CB-NA to CB-LA had little impact on the prediction of
HAB% (data not shown), the ratio of CB-NA to CB-LA
was always set at 1.0. De¢ned mixed cultures that were
intended to mimic MRB from cattle fed either grain or
hay produced more ammonia than actual MRB, but the
over-prediction was greater for MRB from cattle fed grain
(Fig. 4).
3.9. Bacterial antagonism
Fig. 4. The speci¢c activity of de¢ned mixed cultures of ruminal bacteria (predicted by the kinetic model) versus corresponding MRB. De¢ned
mixed cultures that mimicked MRB from cattle fed hay are shown by
squares and those that mimicked cattle-fed grain are shown as circles.
Open symbols are 1.25 mg ml31 Trypticase and closed symbols are 15
mg ml31 Trypticase.
HAB that were incubated for 2 h with a high concentration of Trypticase (15 mg ml31 ) at pH 7.0 had a speci¢c
activity of approximately 500 nmol NH3 mg protein31
min31 (Fig. 5). If HAB were mixed with washed MRB
cells from cattle fed grain at pH 7.0, and the speci¢c activity of MRB was subtracted, the HAB had a speci¢c
activity of only 280 nmol NH3 mg protein31 min31 . If
the MRB from cattle fed grain were autoclaved prior to
the incubation, the inhibition was no longer observed.
Washed MRB from cattle fed hay had little impact on
HAB.
4. Discussion
Protein catabolism by ruminal bacteria is a multi-step
FEMSEC 1122 10-5-00
126
J.L. Rychlik, J.B. Russell / FEMS Microbiology Ecology 32 (2000) 121^128
process that employs proteinases, peptidases, amino nitrogen transport systems, and deaminases. Only a few species
of ruminal bacteria are actively proteolytic [15^18], but
soluble proteins are degraded at a rapid rate by MRB
[19]. Based on the observation that a variety proteinase
inhibitors decreased the proteolytic activity of MRB, it
appears that ruminal bacteria have a wide variety of proteinases [20]. Peptides arising from proteolysis can be utilized by ruminal bacteria [21,22], but peptides are also
degraded extracellularly by peptidases [23]. Wallace and
McKain [24] concluded that dipeptidyl peptidase activity
from Prevotella (Bacteroides) ruminicola was the most important peptidase in ruminal £uid, but it should be noted
that a variety of ruminal bacteria have peptidase activities.
Within the rumen, the pool of non-ammonia, non-protein nitrogen (peptides and amino acids) ranges from approximately 0.2 mg ml31 (before feeding) to approximately 1.2 mg ml31 (soon after feeding), and HCl
hydrolysis indicated that the average peptide size ranged
from nine amino acids (before feeding) to three amino
acids (soon after feeding) [25,26]. Casein is a non-glycolsylated protein that has been used as a model for protein
degradation in the rumen [1,27], but kinetic studies with
MRB and casein hydrolysate (Trypticase) indicated that
the Km of Trypticase was relatively high [28]. Trypticase
is a heterogeneous substrate. Many of the peptides from
Trypticase are readily utilized, but some are degraded so
slowly that they can pass undegraded from the rumen [27].
Fractionations [28] and studies with synthetic peptides [29]
indicated that the slowly degraded pool has an abundance
of proline containing peptides.
In the 1930s Stickland noted that clostridia deaminated
oxidized and reduced amino acids in coupled reactions
[30], but there has been little indication that Stickland
reactions are important in ruminal bacteria [31]. Within
the rumen, methanogenesis is an alternative mechanism
of reducing equivalent disposal, but methane inhibition
had relatively little impact on the production of ammonia
from protein hydrolysate [31]. Bladen et al. [8] concluded
that P. ruminicola was probably the most important ammonia-producing ruminal bacterium, but even the best
strain (B1 4) produced ammonia at a relatively slow rate.
When carbohydrates were available, virtually all of the
amino nitrogen was incorporated into microbial protein
and the deamination rate was even slower [32]. Carbohydrate-fermenting ruminal bacteria did not have deamination rates that could explain the ability of MRB to produce ammonia ; however, HAB have been shown to
generate ammonia 20-fold faster [33,34].
The enumeration of HAB has been confounded by the
observation that some HAB did not grow well in the laboratory [9]. Recent work indicated that MRB from cattle
fed grain produced ammonia 50% slower than MRB from
cattle fed hay, but HAB were not enumerated [35]. Some
HAB are relatively sensitive to low pH and the ruminal
pH of cattle fed grain can be lower than cattle fed hay.
However, MRB from cattle fed grain still produced ammonia 50% slower than MRB from cattle fed grain even if
the in vitro pH was near neutral, and this result indicated
that grain-fed cattle might have fewer HAB [35]. HAB
have a much higher Vmax than other ruminal bacteria,
but Km values for HAB had not been measured.
The mathematical model of ammonia production in the
rumen indicated that grain-fed cattle would have very low
numbers of HAB, particularly at high concentrations of
Trypticase, but HAB appeared to play a more signi¢cant
role in hay-fed cattle. Because the HAB had a relatively
high Vmax , we had originally thought that they might be
more important at high concentrations of Trypticase, but
the model indicated the reverse. The increase in HAB at
low Trypticase concentrations could be explained by the
observation that HAB had a lower Km than CB-LA. MRB
had a lower Km than either HAB or CB-LA, but ammonia
production is a function of Vmax as well as Km . MRB had
a Km Vmax 31 that falls between CB-LA and HAB, and Km
Vmax 31 is the index of relative £ux. Because CB-NA had
virtually no capacity to produce ammonia, we had initially
believed that an increase in CB-NA and a decrease in CBLA would have a large impact on the relative numbers of
HAB. However, the model indicated the ratio of CB-NA
to CB-LA had little e¡ect. By using data generated by the
kinetic model, it was possible to estimate the relative contribution of HAB to total ammonia production. These
calculations indicated that HAB produced at least 40%
of the ammonia in cattle fed hay, but contribution of
HAB in cattle fed grain was at least four-fold lower.
The ability of HAB to grow in the rumen is constrained
by their ability to derive ATP from amino acid fermentation [36]. P. anaerobius [37] had a yield of approximately
8 mg protein mmol NH31
3 , but the yields of C. sticklandii
and C. aminophilum were nearly 3-fold lower [33]. The
question then arose, would there be enough substrate for
HAB to achieve high numbers in the rumen? If a cow
consumed 1.5 kg day31 , the nitrogen content of the protein was 16%, the ruminal volume was 70 l, and 80% of
the protein was degraded to ammonia, total ammonia £ux
would be 200 mmol l31 . If HAB had a yield of 8 mg
protein mmol NH31
3 and they consumed 40% of the amino
acid nitrogen (cattle fed hay), the mass of HAB in the
rumen would be 640 mg protein l31 . If the total bacteria
protein in the rumen was 4000 mg protein l31 [38], HAB
could account for 16% of the MRB. If HAB only produced 10% of the ammonia (cattle fed grain), then HAB
would account for only 4% of the MRB. Both of these
estimates are within the range predicted by the mathematical model (Fig. 1).
The model was based on a mixture of HAB previously
isolated in our laboratory (P. anaerobius C, C. sticklandii
SR, C. aminophilum F) [33,34], but Attwood et al. recently
isolated HAB with even higher speci¢c activities of ammonia production (as high as 950 nmol mg protein31 min31 )
[11]. When the Vmax of the model was varied, the predicted
FEMSEC 1122 10-5-00
J.L. Rychlik, J.B. Russell / FEMS Microbiology Ecology 32 (2000) 121^128
numbers of HAB changed, but this trend was most apparent at low Vmax values. If the Vmax was increased from 500
to 1000 nmol mg protein31 min31 , the predicted HAB for
cattle fed hay only decreased from 18 to 12%. For cattle
fed grain, the change was even less. HAB had a relatively
high Km , but changes in Km had even less e¡ect on the
predicted number of HAB than changes in Vmax .
Previous work indicated that some HAB were de¢cient
in peptidase activity [33,37], but model validations with
predicted ratios of HAB, CB-LA and CB-NA indicated
that the model tended to over-predict ammonia production. If peptidase activity was restricting HAB, one would
have expected lower (not higher) ammonia production
rates. The model indicated that cattle fed grain would
have fewer HAB than cattle fed hay, but the mechanism
of this e¡ect was not initially apparent. Co-culture experiments with HAB that washed MRB from cattle fed grain
were able to decrease ammonia production from HAB,
but MRB from cattle fed hay did not a¡ect HAB.
Further research is needed to de¢ne the antagonism
between MRB from cattle fed grain and HAB, but it
should be noted that autoclaved MRB did not cause an
inhibition and the incubations were always carried out at
pH 7.0. This latter observation indicates that that e¡ect on
HAB was not simply a chemical inhibition (e.g. low pH).
Some ruminal bacteria produce bacteriocins [39] and many
bacteriocins are inactivated by heat [40]. Another possibility are ruminal bacteriophages. The rumen has a relatively
high concentration phage particles [41], but the e¡ect of
MRB on HAB was transitory. If HAB and MRB from
cattle fed grain were transferred successively in basal medium with Trypticase, the MRB died out, but the HAB
persisted.
Acknowledgements
J.B.R. is a member of the U.S. Dairy Forage Research
Center, Madison, WI.
References
[1] Nolan, J.V. (l975) Quantitative models of nitrogen metabolism in
sheep. In: Digestion and Metabolism in the Ruminant (MacDonald,
I.W., Warner, A.C.I., Eds.), pp. 416^431. Univ. New England Publishing Unit, Arimdale.
[2] Annison, E.F. (1956) Nitrogen metabolism in the sheep. Biochem. J.
64, 705^714.
[3] Russelle, M.R. (1996) Pasture nitrogen £ow. In: Informational Conference with Dairy and Forage Industries. pp. 73^80. U.S. Dairy
Forage Research Center.
[4] Ferguson, K.A. (1975) The protection of dietary proteins and amino
acids against microbial fermentation in the rumen. In: Digestion and
Metabolism in the Ruminant (MacDonald, I.W., Warner, A.C.I.,
Eds.), pp. 448^464. Univ. New England Publishing Unit, Arimdale.
[5] Slyter, L.L. (1976) In£uence of acidosis on rumen function. J. Anim.
Sci. 43, 910^929.
127
[6] Yang, C.M.J. and Russell, J.B. (1993) The e¡ect monensin supplementation on ruminal ammonia accumulation in vivo and the
numbers of amino-acid fermenting bacteria. J. Anim. Sci. 71, 3470^
3476.
[7] Lana, R.P. and Russell, J.B. (1997) E¡ect of forage quality and
monensin on the ruminal fermentation of ¢stulated cows fed continuously at a constant intake. J. Anim. Sci. 75, 224^229.
[8] Bladen, H.A., Bryant, M.P. and Doetsch, R.N. (1961) A study of
bacterial species from the rumen which produce ammonia from protein hydrolyzate. Appl. Microbiol. 9, 175^180.
[9] Paster, B., Russell, J.B., Yang, C.M.J., Chow, J.M., Woese, C.R. and
Tanner, R. (1993) Phylogeny of ammonia-producing ruminal bacteria, Peptostreptococcus anaerobius, Clostridium sticklandii and Clostridium aminophilum sp. nov.. Int. J. Syst. Bacteriol. 43, 107^110.
[10] Krause, D.O. and Russell, J.B. (1996) An rRNA approach for assessing the role of obligate amino acid-fermenting bacteria in ruminal
amino acid degradation. Appl. Environ. I 62, 815^821.
[11] Attwood, G.T., Klieve, A.V., Ouwerkerkand, D. and Patel, B.K.
(1998) Ammonia-hyperproducing bacteria from New Zealand ruminants. Appl. Environ. Microbiol. 64, 1794^1804.
[12] Russell, J.B., Sni¡en, C.J. and Van Soest, P.J. (1983) E¡ect of carbohydrate limitation on degradation and utilization of casein by
mixed rumen bacteria. J. Dairy Sci. 66, 763^775.
[13] Chaney, A.L. and Marbach, E.P. (1962) Modi¢ed reagents for determination of urea and ammonia. Clin. Chem. 8, 130^132.
[14] Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951)
Protein measurement with the Folin phenol reagent. J. Biol. Chem.
193, 265^275.
[15] Blackburn, T.H. and Hobson, P.N. (1960) Isolation of proteolytic
bacteria from the sheep rumen. J. Gen. Microbiol. 22, 282^289.
[16] Blackburn, T.H. (1968) The protease liberated from Bacteroides amylophilus strain H18 by mechanical disintegration. J. Gen. Microbiol.
53, 37^51.
[17] Cotta, M. and Hespell, R. (1986) Proteolytic activity of the ruminal
bacterium Butyrivibrio ¢brisolvens. Appl. Environ. Microbiol. 52, 51^
58.
[18] Russell, J.B., Bottje, W.G. and Cotta, M.A. (1981) Degradation of
protein by mixed cultures of rumen bacteria : Identi¢cation of Streptococcus bovis as an actively proteolytic rumen bacterium. J. Anim.
Sci. 53, 242^252.
[19] Mahadevan, S., Er£e, J.D. and Sauer, F.D. (1980) Degradation of
soluble and insoluble proteins by Bacteroides amylophilus protease
and by rumen microorganisms. J. Anim. Sci. 50, 723^728.
[20] Brock, F.M., Forsberg, C.W. and Buchanan-Smith, J.G. (1982) Proteolytic activity of rumen microorganisms and e¡ects of proteinase
inhibitors. Appl. Environ. Microbiol. 44, 561^569.
[21] Wright, D.E. (1967) Metabolism of peptides by rumen micro-organisms. Appl. Microbiol. 15, 547^550.
[22] Wright, D.E. and Hungate, R.E. (1967) Amino acid concentrations in
rumen £uid. Appl. Microbiol. 15, 148^151.
[23] Wallace, R.J, Onodera, R. and Cotta, M.A. (1997) Metabolism of
nitrogen-containing compounds. In: The Rumen Microbial Ecosystem (Hobson, P.N. and Stewart, C.S., Eds.), pp. 283^328. Blackie
Academic and Professional, London.
[24] Wallace, R.J. and McKain, N. (1991) A survey of peptidase activity
in rumen bacteria. J. Gen. Microbiol. 137, 2259^2264.
[25] Chen, G., Russell, J.B. and Sni¡en, C.J. (1987) A procedure for measuring peptides in rumen £uid and data suggesting that peptide uptake is the rate-limiting step in ruminal protein degradation. J. Dairy
Sci. 70, 1211^1219.
[26] Chen, G., Sni¡en, C.J. and Russell, J.B. (1987) Concentration and
estimated £ow of peptides from the rumen of dairy cattle : E¡ects of
protein quantity, protein solubility, and feeding frequency. J. Dairy
Sci. 70, 983^992.
[27] Mangan, J.L. (1972) Quantitative studies on nitrogen metabolism in
the bovine rumen. Br. J. Nutr. 27, 261^283.
[28] Chen, G., Strobel, H.J., Russell, J.B. and Sni¡en, C.J. (1987) The
FEMSEC 1122 10-5-00
128
[29]
[30]
[31]
[32]
[33]
[34]
[35]
J.L. Rychlik, J.B. Russell / FEMS Microbiology Ecology 32 (2000) 121^128
e¡ect of hydrophobicity on the uptake and deamination of peptides
by ruminal bacteria in vitro. Appl. Environ. Microbiol. 53, 2021^
2025.
Yang, C.M.J. and Russell, J.B. (1992) The resistance of proline-containing peptides to ruminal degradation in vitro. Appl. Environ. Microbiol. 58, 3954^3958.
Nisman, B. (1954) The Stickland reaction. Bacteriol. Rev. 18, 16^42.
Hino, T. and Russell, J.B. (1985) The e¡ect of reducing equivalent
disposal and NADH/NAD on the deamination of amino acids by
intact and cell-free extracts of rumen microorganisms. Appl. Environ.
Microbiol. 50, 1368^1374.
Russell, J.B. (1983) Fermentation of peptides by Bacteroides ruminicola B1 4. Appl. Environ. Microbiol. 45, 1566^1574.
Chen, G. and Russell, J.B. (1989) More monensin-sensitive, ammonia
producing bacteria from the rumen. Appl. Environ. Microbiol. 55,
1052^1057.
Russell, J.B., Strobel, H.J. and Chen, G. (1988) The enrichment and
isolation of a ruminal bacterium with a very high speci¢c activity of
ammonia production. Appl. Environ. Microbiol. 54, 872^877.
Lana, R.P., Russell, J.B. and Amburgh, M.E.V. (1998) The role of
[36]
[37]
[38]
[39]
[40]
[41]
pH in regulating ruminal methane and ammonia production. J. Anim.
Sci. 76, 2190^2196.
Russell, J.B., Onodera, R., and Hino, T. (1991) Ruminal protein
fermentation: new perspectives on previous contradictions. In: Physiological Aspects of Digestion and Metabolism in Ruminants, Proceedings of the Seventh International Symposium on Ruminant Physiology (Tsuda, T., Sasaki, Y., and Kawashima, R., Eds.), pp. 681^
700. Academic Press, San Diego, CA.
Chen, G. and Russell, J.B. (1988) Fermentation of peptides and amino acids by a monensin-senstive ruminal peptostreptococcus. Appl.
Environ. Microbiol. 54, 2742^2749.
Czerkawski, J.W. (1986) Rumen Studies, 1st edn., 236 pp. Pergamon
Press, Oxford.
Kalmoko¡, M.L., Bartlett, F. and Teather, R.M. (1996) Are ruminal
bacteria armed with bacteriocins ? J. Dairy Sci. 79, 2297^2308.
Jack, R.W., Tagg, J.R. and Ray, B. (1995) Bacteriocins of Grampositive bacteria. Microbiol. Rev. 59, 171^200.
Klieve, A.V. and Swain, R.A. (1993) Estimation of ruminal bacteriophage numbers by pulsed-¢eld gel electrophoresis and laser densitometry. Appl. Environ. Microbiol. 59, 2299^2303.
FEMSEC 1122 10-5-00