A Net Carbohydrate and Protein System for
Evaluating Cattle Diets: I. Ruminal Fermentation
J. B. RusselI*J, J. D. O’ConnortJ, D. G. Fox+,
P. J. Van Soestt, and C. J. Sniffentvz
*U.S. Dairy Forage Research Center, ARS, USDA, Madison, WI 53706 and
U.S. Plant, Soil, and Nutrition Laboratory, Ithaca, NY 14853 and
+Department of Animal Science, Cornell University, Ithaca, NY 14853
ABSTRACT: The Cornell Net Carbohydrate and
Protein System (CNCPSI has a kinetic submodel
that predicts ruminal fermentation. The ruminal
microbial population is divided into bacteria that
ferment structural carbohydrate (SC) and those
that ferment nonstructural carbohydrate (NSC).
Protozoa are accommodated by a decrease in the
theoretical maximum growth yield (.50vs .40 g of
cells per gram of carbohydrate fermented), and the
yields are adjusted for maintenance requirements
LO5 vs .150 g of cell dry weight per gram of
carbohydrate fermented per hour for SC and NSC
bacteria, respectively). Bacterial yield is decreased
when forage NDF is < 20% (2.5% for every 1 YO
decrease in NDF). The SC bacteria utilize only
ammonia as a N source, but the NSC bacteria can
utilize either ammonia or peptides. The yield of
NSC bacteria is enhanced by as much as 18.7%
when proteins or peptides are available. The NSC
bacteria produce less ammonia when the carbohydrate fermentation (growth) rate is rapid, but 34%
of the ammonia production is insensitive to the
rate of carbohydrate fermentation. Ammonia
production rates are moderated by the rate of
peptide and amino acid uptake (.07 g of peptide
per gram of cells per hour), and peptides and
amino acids can pass out of the rumen if the rate
of proteolysis is faster than the rate of peptide
utilization. The protein-sparing effect of ionophores is accommodated by decreasing the rate of
peptide uptake by 34%. Validation with published
data of microbial flow from the rumen gave a
regression with a slope of .94 and a n r2 of .88.
Key Words: Ruminants, Rumen, Fermentation, Microbial Proteins, Ammonia,
Computer Simulation
J. Anim. Sci. 1002. 70:3551-3561
Introduction
In ruminants, feedstuffs are fermented in the
rumen before gastric and intestinal digestion, and
this fermentation has confounded the prediction of
animal performance from dietary ingredients.
Over the years, there has been considerable
improvement in the feeding of ruminants, but this
progress has been based on an empirical approach that has largely treated the rumen as a
“black box.” The experience of other scientific
disciplines has shown that a mechanistic understanding is usually needed for sustained develop
ment. If ruminant nutrition is to progress past the
point at which diets are continually tested in
virtually infinite combinations, the details of the
fermentation must be considered. The Cornell Net
Carbohydrate and Protein System (CNCPS) has a
mechanistic submodel that provides quantitative
estimates of fermentation end products (the ME
from VFA production, microbial protein and ammonia) and materials that escape ruminal degradation (carbohydrates, protein, and undegraded
peptides). The CNCPS can serve as a research tool
or a guide for practical ration balancing.
Ruminal Ecosystem
‘Present address: P. 0. Box 2077, Corvallis, OR 97339.
2Present address: W. H. Miner Agric. Res. Inst., Chazy, NY
12921.
Received May 24, 1991.
Accepted July 13, 1992.
The quality and quantity of ruminal fermentation products are dependent on the types and
activities of microorganisms in the rumen, and the
ruminal microbial ecosystem is very diverse (Hun-
3551
3552
RUSSELL ET AL.
gate, 1966; Bauchop, 1979; Coleman, 1980; Citron et
al., 1987). Our understanding of ruminal microbial
ecology is also complicated by the fact that there
are numerous interrelationships among the various ruminal microbes (Russell and Hespell, 1981).
The complexity of the ruminal ecosystem has led
many nutritionists to conclude that the ruminal
ecosystem is so complex that it cannot be understood or described in quantitative terms. More
optimistic workers have attempted to model certain aspects of ruminal fermentation (Reichel and
Baldwin, 1976; Baldwin et al., 19771, but the degree
of “aggregation” has differed greatly. The persistent question has always been, What level of
aggregation and detail will allow a n adequate
representation?
Reichel and Baldwin (19761 presented a linear
programming model for evaluating ruminal
microbial function that used eight groups of
ruminal microorganisms. They found that “during
several solutions of the model, considerable simplification of the rumen microflora occurred. This
implies that current data and concepts, and the
hypothesis regarding relative metabolic rate, as
represented in the model, do not accommodate
adequately competition among several rumen
microbial species and, thus, that additional data
and concepts regarding rumen microbial interactions are required.” In subsequent years, Baldwin
and his colleagues discontinued the multimicrobe
approach and adopted a one-“bug”model in which
microbial diversity was not accommodated Baldwin et al., 1977).
The CNCPS divides the ruminal microbial
ecosystem into two microbial groups, microbes
that ferment nonstructural carbohydrate (NSC)
and those that ferment structural carbohydrate
(SC). This segregation reflects differences in N
utilization and growth efficiency as well as an
almost exclusive partition of energy source utilization. The SC bacteria ferment only cell wall
carbohydrate, use only ammonia as a N source,
and do not ferment peptides or amino acids. The
NSC bacteria ferment nonstructural carbohy
drates (starch pectin, sugars, etc.), use either
ammonia or peptides and amino acids as a N
source, and can produce ammonia. Certain strains
of Butyrivibrio fibrisohens can ferment both starch
and cellulose and produce ammonia, but they
degrade cellulose a t a much slower rate than do
other cellulolytic bacteria (Bryant, 1973). In the
CNCPS, B. fibrisolvem would be classified as an
NSC fermenter.
An Ideal Ruminal Fermentation
Any economic approach to ration balancing
should seek to maximize the beneficial aspects of
ruminal fermentation while minimizing fermenta-
tion losses. Because mammals do not produce
enzymes that can hydrolyze cellulose or hemicellulose, rapid rates of ruminal fiber digestion are
desirable. In many dietary situations, much of the
amino acid N reaching the intestines is of
microbial origin, and this dependency on
microbial protein means that the efficiency of
microbial growth can have a pronounced effect on
the ruminant. Ruminal microorganisms can utilize
ammonia, but in many cases the ammonia production rate exceeds the utilization rate. Excessive
ammonia production and absorption increases N
excretion and the energy cost of urea synthesis.
The 15N studies of Nolan (1975) indicated that >
25% of the protein N may be lost in this manner.
Volatile fatty acids arising from ruminal fermentation are the primary energy source for the
animal. Acetate and butyrate are used efficiently
by fattening animals, but they cannot make a net
contribution to the glucose supply. Propionate can
be used for gluconeogenesis, but a high ratio of
propionate to acetate in the rumen is associated
with a reduction in milk fat. If the fermentation
rate in the rumen is rapid, lactic acid sometimes
accumulates. Lactate can be converted to blood
glucose, but it is a much stronger acid than the
VFA. Lactate accumulation leads to ruminal acidosis, a decline in fiber digestion, decreased feed
intake, founder, and, in extreme cases, even death
(Slyter, 1976). In vitro studies indicated that the
efficiency of microbial protein synthesis can
decline significantly at pH values < 6.0 (Strobel
and Russell, 1986). Methanogenesis provides a n
alternative means of reducing equivalent disposal,
which allows the microflora to increase ATP
production, but methane can represent a significant loss of feed energy. The CNCPS provides
estimates of VFA energy and ammonia and
microbial protein production.
Rates vs Amounts
Fiations for ruminants traditionally have been
balanced according to the amounts of specific feed
components (crude fiber, ether extract, nitrogenfree extract, and crude protein). Recent work,
however, has indicated that the rate of feedstuff
degradation in the rumen can have a profound
effect on fermentation end products and on animal
performance (Nocek and Russell, 1988): 1) if the
rate of protein degradation exceeds the rate of
carbohydrate fermentation, large quantities of N
can be lost as ammonia; 21 if the rate of carbohydrate fermentation exceeds the protein degradation rate, microbial protein production can
decrease; 3) if feedstuffs are degraded very slowly,
rumen fill will decrease intake; and 4) if the
CARBOHYDRATE AND PROTEIN SYSTEM FOR CATTLE
degradation rate is slow, some of the feed may
escape ruminal fermentation and pass directly to
the lower gut.
The procedures of Goering and Van Soest (1970)
provided a basis for estimating amounts of available fiber, and in vitro and in situ studies have
provided data on the rates of fiber digestion.
Protein degradation rates can be estimated by
enzymatic procedures (Krishnamoorthy et al.,
19821, but starch degradation is still difficult to
estimate. The fermentation rate of starch can vary
greatly, and this rate is influenced by feed treatment (e.g., pelleting), the method of storage (e.g.,
dried shell corn vs high-moisture corn), and the
type of cereal grain that is fed (e.g.,barley vs corn;
0rskov, 19761. Until recently, the best estimates of
starch degradation have come from in situ studies
in which cereal grain was placed in a nylon bag
and suspended in the rumen (Nocek, 1988). Better
methods are needed to estimate the rate at which
starch will be degraded in vivo. A companion
paper (Sniffen et al., 1992) describes the composition and fermentation rates of various feedstuffs.
Passage vs Fermentation
Much of the feed that enters the rumen is
fermented, but some feed escapes ruminal degradation. The fate of the feed is ultimately determined by the relative rates of fermentation and
passage (Waldo and Smith, 1972). Fermentation
rate is an inherent property of the feed, and in the
CNCPS fermentation rates are described by firstorder kinetics (substrate-limited, enzyme excess)
and specific rate constants
Passage rates
are regulated by feed intake, processing
(chopping, grinding, and other means of particle
size reduction), and the type of feed that is
consumed (e.g., forage vs cereal grain). In the
CNCPS, passage rates can be altered by the user;
the guidelines for selecting passage rates are
given in a companion paper (Sniffen et al., 1092).
For many years it was assumed that proteolysis
was the rate-limiting step in ruminal protein
degradation and that proteolysis was synonymous
with utilization. This assumption was supported
by the observation that amino acids were usually
present a t low concentrations in ruminal fluid
(Wright and Hungate, 10671, but peptides were not
measured. Winters et al. (1964) noted that ruminal
fluid contained large amounts of nonammonia N
that could not be precipitated by trichloroacetic
acid, and in vitro experiments showed that p e p
tides sometimes accumulated (Mangan, 1972; Flussell et al., 1983). These peptide accumulations
indicated that certain types of peptides were not
readily utilized by ruminal bacteria; the CNCPS
uses a rate constant of .07 g of peptide per gram of
3553
microorganism per hour to reflect this limitation
(Hino and Russell, 1985). Peptide flow from the
rumen usually accounts for a small percentage of
the dietary N (Chen et al., 1987a,b), but the peptide
uptake rate can have a significant effect on the
rate of ammonia production (see Protein Fermentation and Ammonia Accumulation below).
The passage of undegraded feed from the rumen
can affect nutrient availability. If the passed feed
can be digested in the small intestine (e.g., protein
and starch), there may be a decrease in fermentation losses (ammonia and methane) and an increase in nutrient retention. However, if the feed is
digested poorly postruminally, digestibility will
decline. A reduction in digestibility is not always
detrimental. If an increase in feed intake counteracts the decrease in digestibility, the rate of
nutrient absorption may increase. The optimal
feeding strategy and degree of passage is dependent on the cost of the feed and the value of
animal production.
Passage can have a profound effect on the
balance of ruminal fermentation products. If carbohydrates are not digested in the rumen, there
will be a decrease in microbial growth and
ammonia utilization. Although the passage of
protein from the rumen may be advantageous
from the standpoint of ammonia accumulation, a
reduction in ruminally degraded protein may
cause a decrease in the efficiency of microbial
protein synthesis. The concept that ruminal microbes can only utilize ruminally available feedstuffs is obvious; the use of TDN or total digestibility to predict ruminal microbial yield in many
instances will cause errors of prediction.
NRC Requirements
The National Research Council (NRC) of the
National Academy of Sciences publishes estimates
of the nutrient requirements of domestic animals,
including ruminants. These guidelines are used by
researchers, extension agents, nutritional consultants, and, ultimately, by animal producers. The
1985 NRC publication about N usage by ruminants
discusses various aspects of microbial growth in
the rumen, and it provides equations that attempt
to relate various aspects of ruminal function to
animal performance. These empirical recommendations have a variety of limitations: 1) microbial
growth in the rumen is driven by TDN or total
digestibility rather than by an estimate of ruminally available carbohydrate; 2) the dairy cattle
equation relating TDN to microbial yield has a
negative intercept, which may underestimate
microbial protein production at low TDN intakes;
31 microbial growth efficiency is constant; 4) the
relationship between microbial yield and
RUSSELL ET AL.
3554
23)
Y
0
.2
.4
.6
.8
CARBOHYDRATE FERMENTATION or
GROWTH RATE (% per h)
Figure 1. The effect of carbohydrate fermentation
rate (growth rate) on the yield of ruminal bacteria that
ferment structural carbohydrate (SC) and nonstructural
carbohydrate (NSC). The theoretical maximum growth
yield is .4 g of cells per gram of carbohydrate, and the
maintenance energy requirements for SC and NSC
bacteria are .05 and .150 g of carbohydrate per gram of
bacteria per hour, respectively. Growth rates in the
rumen usually range from .05 to .2 h-l.
microbial maintenance energy requirement is ignored; 5) the microbial population is not partitioned according to metabolite activity and N
requirements; 6) rates of carbohydrate fermentation are not integrated with rates of protein
degradation; and 7) feed degradations are fixed
and thus not sensitive to variations in feed intake
and passage. The CNCPS rumen submodel has
allowances for each of these factors and provides
a more quantitative analysis of ruminal fermentation and nutrient availability.
Microbial Growth
Ruminal microorganisms derive most of their
energy from the fermentation of carbohydrates,
and ruminal bacteria may be categorized in a
general way according to the type of carbohydrate
that they ferment (Russell, 1984). In the CNCPS,
the ruminal microorganisms are partitioned into
those that ferment SC and those that ferment
NSC. Butyrivibrio fibrisolvens has the potential to
occupy both of these niches, but most species (e.g.,
cellulolytics vs amylolytics) can be separated by
this arbitrary classification.
Microorganisms that ferment cellulose and
hemicellulose (SC) grow slowly and utilize ammonia as a N source for microbial protein synthesis.
Microorganisms that ferment starch, pectin, and
sugars (NSC) grow more rapidly than those that
ferment SC and utilize either ammonia or amino
acids as a N source. In the CNCPS, the growth
rate of both groups is directly proportional to the
rate of carbohydrate digestion, so long a s a
suitable N source is available (Hungate, 1966;
Bryant, 1973; Hespell and Bryant, 1979; Russell
and Baldwin, 1981).
Empirical models of microbial growth in the
rumen have often assumed that ruminal microbial
growth yields are a constant function of DMI or
OM digestion (Nocek and Russell, 19881, and the
NRC (1985, 1989) used a static efficiency of 26.12 g
of microbial N per kilogram of TDN. The use of
TDN to determine microbial yield ignores the fact
that most ruminal bacteria are unable to utilize
protein, fat, lipid, or ash as an energy source, and
that carbohydrate is the primary energy source for
growth (Nocek and Russell, 1988).A newly isolated
group of ammonia-producing bacteria can utilize
peptides and amino acids as a n energy source
(Russell et al., 1988; Chen and Russell, 1989).These
bacteria have a significant effect on ammonia
production (see Protein Fermentation and Ammonia Accumulation below), but they are present at
low numbers in vivo and only account for a small
percentage of the microbial protein production.
The use of a static efficiency by the NRC also
ignores the fact that ruminal microorganisms have
maintenance energy requirements. When bacteria
grow slowly, a large proportion of the energy is
used to maintain the cells, and in this regard
maintenance energy is analogous to the fixed
overhead of a business. One can only make a
profit (growth) after the overhead is met [maintenance). If cash flow is large (rapid rates of
energy utilization), overhead (maintenance) will
make up a small proportion of the total budget.
Because the growth rate of microorganisms in the
rumen is sometimes very slow, maintenance
energy can have a significant effect on microbial
growth efficiency (Russell and Wallace, 1988).
In the derivation of Pirt (19851, bacterial maintenance is defined as a time-dependent function
that is directly proportional to cell mass (m =
grams of carbohydrate per gram of bacteria per
hour), and the theoretical maximum yield (YG=
grams of bacteria per gram of carbohydrate) is
defined as the yield that would be obtained if there
were no maintenance. The CNCPS uses the Pirt
derivation to adjust microbial yield as a function
of growth rate, but there are few data concerning
the maintenance energy or theoretical growth
yields of ruminal microorganisms.
Isaacson et al. (1975) reported that mixed ruminal microorganisms had a theoretical maximum
AND PROTEIN SYSTEM FOR CATTLE
3555
yield .5 g of cell dry weight per gram of carbohy20
drate. These in vitro experiments were, however,
conducted in the absence of protozoa. Protozoal
predation leads to a turnover of bacteria in the
15
rumen (Coleman, 1980). The CNCPS makes a n
allowance for protozoal predation by decreasing
the theoretical maximum growth yield from 50 to
40%.
10
Pure culture experiments indicated that ruminal
bacterial maintenance energy was not a constant,
and the range was .022 to .187 g of carbohydrate
5
per gram of bacteria per hour (Russell and
Baldwin, 1981). Based on these studies, NSC and
SC bacteria are assigned maintenance values of
0
.150 and .05 g of carbohydrate per gram of bacteria
0
2
4
6
8
10 12 14
per hour, respectively. The effect of bacterial
maintenance on the efficiency of microbial growth
Ratio of Amino Acids to
is depicted in Figure 1. Over the range of specific
Total Organic Mattet%
growth rates possible in the rumen, yield can vary
by more than threefold.
Figure 2 . Effect of amino acid nitrogen on the yield of
In vitro studies showed that ruminal bacteria
ruminal
microbial protein in vitro. Redrawn from
respond in a positive fashion to the provision of
Russell
and
Sniffen (1984).
peptides and amino acids [Russell and Sniffen,
18841, but it should be realized that SC bacteria
are unable to utilize amino N (Bryant, 1973).In the
CNCPS, the yield of NSC bacteria is increased as
growth. In vitro studies (Russell et al., 1983)
much as 18.7%as the ratio of peptides to NSC plus
peptides in ruminal fluid increases from 0 to 14%. indicated that microorganisms that ferment NSC
Above 14% peptide, there is no further improve- derived 66% of their N from peptides or amino
acids and 34% of their N from ammonia, and that
ment in yield (Figure 2). If ammonia becomes
this
proportion was not influenced by the growth
limiting, yield should theoretically decline, but the
of the microorganisms. When peptides and
rate
CNCPS does not make any provision for this type
amino
acids are no longer available, all the N
of limitation. Because ammonia is one of the
must
be
derived from ammonia. Because the SC
outputs of the model, a n ammonia deficiency can
bacteria
do
not utilize peptides or amino acids, all
be remedied by adding NPN to the ruminant diet.
Ruminal cellulolytic bacteria require branched- their N must come from ammonia (Bryant, 1973).
The rumen is well buffered by salivary secrechain VFA for growth (Bryant, 1973). The current
rumen submodel has a provision for branched- tions, but if the amount of dietary fiber is reschain VFA, but we are developing a modification tricted and the rate of carbohydrate fermentation
is rapid, pH can decline. The NRC (1985)indicated
to accommodate this aspect of ruminal fermentathat
diets containing c 40% forage (20% NDF)
tion. Because branched-chain VFA are derived
have
lower microbial growth yields; mixed ruminal
from the fermentation of branched-chain amino
bacteria that were incubated in vitro produced
acids (Russell and Sniffen, 19841, it should be
50% less protein at pH 5.7 than at 6.7 (Strobel and
possible to estimate branched-chain VFA from the
Russell, 1986). The model uses this latter relationfermentation of peptide and amino acids. The
ship to adjust yield, and pH is predicted from the
submodel calculates total ammonia production
from true protein and NPN separately, and there- NDF content of the ration. When forage NDF is 2
20% of the DM, microbial yield is reduced 2.5% for
fore this parameter should be easy to estimate.
every 1% decrease in NDF. This adjustment will
Branched-chain VFA requirements, in turn, could
probably
be most acceptable when the forage is
be estimated from the amount of branched-chain
amino acids in SC microbial protein Purser and
coarsely chopped and the feeding management
Buechler, 1966). A branched-chain VFA deficiency
allows a uniform DMI. If the NDF in the ration is
could be encountered if high-forage diets are low
finely chopped the user can discount yield by 3.0%
in true protein and NPN is used as a supplement.
per unit of NDF. Additional experiments are
Nitrogen utilization (ammonia vs amino N) is
needed to relate ruminal pH changes with the pool
dependent on peptide availability and the type of
of NSC, the carbohydrate fermentation rate, and
carbohydrate that has the primary influence on
the buffering capacity of saliva and fiber.
CARBOHYDRATE
3556
RUSSELL ET AL.
Protein Fermentation and Ammonia
Accumulation
Most ruminal bacteria are able to use ammonia
as a N source for microbial protein synthesis, but
ruminal protein fermentation often creates more
ammonia than the microorganisms can utilize. In
many cases, 2 25% of the protein can be lost as
ammonia (Nolan, 1975). Because protein is the
most expensive ingredient in many ruminant
rations, there has been considerable interest in
reducing ruminal protein fermentation.
Proteins are degraded by extracellular enzymes,
and these proteinases must come in contact with
proteins through a n interaction involving water. If
the protein goes into solution quickly, the rate of
enzymatic degradation is often increased (Tamminga, 1979). Many of the proteins found in
forages and soybeans are very soluble and are
degraded rapidly by ruminal bacteria. Heat treatments that can denature the protein can decrease
solubility and the rate of protein degradation.
Some feeds contain proteins that are naturally
insoluble brewers grains, distiller byproducts, fish
meal, etc.). The submodel uses enzymatic data to
predict the rate at which feedstuff proteins will be
degraded (Krishnamoorthy et al., 1982).
Carbohydrate has little effect on the rate of
protein degradation by extracellular proteinases,
but it can greatly affect the end product of amino
acid metabolism (Russell et al., 1983). In the
CNCPS, NSC microorganisms take up peptides at
a rate of .07 g of peptide per gram of microorganism per hour, and this N is used for microbial
protein synthesis or ammonia production (Russell
and Martin, 1984; Hino and Russell, 1985). The
diversion of peptides to microbial protein or
ammonia is regulated by the availability of carbohydrate. When carbohydrate availability allows
growth, 6 8 % of the NSC microbial protein comes
from peptides and 34% comes from ammonia
(Russell et al., 1983). In the absence of carbohydrate, all the peptide N is converted to ammonia.
Bladen et al. (1961) examined the capacity of
various ruminal bacteria to ferment protein
hydrolyzate and produce ammonia. They c o n
cluded that Bacteroides ruminicola was the most
important amino acid fermenting bacterium in the
rumen of cattle, but subsequent experiments indicated that this species could not account for
ammonia production in vivo. B. ruminicola B14,
one of the most active strains, had a specific
activity of 13.5 nmol of ammonia per milligram of
protein per minute (Russell, 19831, but mixed
ruminal bacteria produced ammonia at a rate of
31 nmol per milligram of protein per minute (Hino
and Russell, 1985). This comparison introduced a
paradox. How could the best strains have an
activity that was less than the average?
When mixed ruminal bacteria were incubated
with an excess of casein and with amounts of
mixed carbohydrates that were inadequate to
support maximum growth rates, there was a linear
decline in ammonia production a s the fermentation rate (growth rate1 increased, but 14N labeling
studies indicated that 34% of the ammonia was
not influenced by carbohydrate availability (Russell et al., 19831. Because B. ruminicola produces
little ammonia when carbohydrate is available
(Russell, 19831, this carbohydrate-insensitive ammonia production could not be explained.
Recent ruminal enrichments, however, revealed
three new species of ruminal bacteria (Russell et
al., 1988; Chen and Russell, 1989). These strains
that do not utilize carbohydrate grew on peptides
and amino acids and produced ammonia
20-fold faster than other ruminal bacteria. Because
their rates of protein synthesis are 10- to
25-fold lower than their rates of ammonia production and their numbers in vivo are not high, these
bacteria contribute little to microbial protein
production and are in most cases detrimental.
Ionophores decreased ammonia production in
vivo (Dinius et al., 1976) and in vitro (Van Neve1
and Demeyer, 1977; Russell and Martin, 19841, but
they have little effect on proteolysis. Whetstone et
al. (1981) noted that monensin caused an increase
in nonammonia, nonprotein nitrogen in vitro.
Previously isolated, ammonia-producing bacteria
were resistant to monensin (Chen and Wolin, 1979;
Dennis et al., 19811, but the newly isolated ammonia-producing bacteria are sensitive to ionophores
(Russell et al., 1988; Chen and Russell, 1989).
The new isolates are included in the NSC
bacteria even though they are unable to utilize
carbohydrate. Their contribution arises from the
relationship between peptide transport and ammonia production. If only 64% of the peptides that are
taken up can ever be incorporated into microbial
protein, the remainder must end up as ammonia.
The influence of ionophores on ruminal ammonia
production can be accommodated by a 34%
reduction in the peptide uptake rate constant.
Recycled Nitrogen
Recycled nitrogen may be a significant ruminal
input when ammonia is deficient. The NRC (1985,
1989) assumed that recycled nitrogen can be as
much a s 70% of the protein intake if the dietary
protein intake is low (5% CP). When the protein
intake is high (20% CPI the contribution of
recycled nitrogen decreases ( 1 1% of intake CP).
The CNCPS uses the NRC (1985) equation for
recycled nitrogen: Y = 121.7 - 12.01X + .3235X2,
where Y = urea N recycled (percentage of N
CARBOHYDRATE
AND PROTEIN SYSTEM FOR CATTLE
intake) and X = intake of crude protein, as a
percentage of diet DM.
If recycled N makes up a large proportion of the
total ruminal N, the long-term protein needs of the
animal may be underestimated. When protein
intake increases to support maximal rates of milk
production or gain, more amino acids may be used
for gluconeogenesis (for lactose and fetal growth),
the efficiency of amino acid N utilization may
decrease, and there can be a n increase in recycled
N. Further work (possibly a n additional submodell
is needed to assess the value of recycled N as a N
source for ruminal microbial growth. Recycled N
3557
can supply ruminal ammonia, but not ruminal
peptides.
Composition of Ruminal Bacteria
Bacterial composition can be influenced by
factors such as changes in growth rate, growth
phase, and growth media. In this model, bacteria
are assumed to be 62.5% CP, 21.1% carbohydrate,
12% fat, and 4 . 4 % ash on a DM basis (Hespell and
Bryant, 19791. Only 50 to 70% of microbial N is
available protein, and the remainder is bound in
Table 1. Glossary of terms
Y
urea nitrogen recycled, 96 of nitrogen intake
intake crude protein, % of dry matter
percentage of NDF dry matter that is forage NDF
yield efficiency of SC bacteria from the available fiber fraction of the jth feedstuff, g of SC bacterialg of
SC digested
yield efficiency of NSC bacteria from the sugar fraction of the jth feedstuff, g of NSC bacteria/g of NSC
digested
yield efficiency of NSC bacteria from the starch fraction of the jth feedstuff, g of NSC bacteria/g of NSC
digested
maintenance rate of the structural carbohydrate bacteria, g of SC.g of bacteria-'.h-I
maintenance rate of the nonstructural carbohydrate bacteria, g of NSC/g
theoretical maximum yield of the structural carbohydrate bacteria, g of bacterialg
theoretical maximum yield of the nonstructural carbohydrate bacteria, g
ratio of peptides to peptide plus NSC in the jth feedstuff
peptides in the j* feedstuff
g of NSC in the A (sugar)fraction of the jth feedstuff
g of NSC in the B (starch and pectins) fraction of the
feedstuff
g of SC in the B2 (available cell wall) fraction in the jt feedstuff
g of ruminally degraded true protein
g of ruminally degraded true protein nitrogen
X
RFNDF
y11
YZl
y31
KM1
KM2
YGl
YGZ
Ratioj
RDPEPj
RDCAj
RDCB1j
RDCB2j
RDPA
RDPAN
th
Kd4j
Kd5j
Kd6j
IMPj
SCBACTj
NSCBACTj
BACTj
BACTNj
SCBACTNj
NSCBACTNj
K1
PEPUPj
PEPUPNj
PEPUN&
NPN
NSCAMMNR
SCAMMNR
KUPI
KUP
RAN
REBTPj
REBCWj
REBNAj
REBCHOj
REBFATi
REBASHj
=
=
=
=
--
I
=
I
growth rate of the sugar-fermenting bacteria, h-'
growth rate of the starch-fermenting bacteria, h-'
growth rate of the SC-fermenting bacteria, h-'
percentage improvement in bacterial yield, 46, due to the ratio of peptides to peptides plus NSC in jth feedstuff, maximum = 18% (see Figure 21
yield of bacteria fermenting SC, from the jth feedstuff, g/d
yield of bacteria fermenting NSC from the jth feedstuff, g/d
yield of bacteria, from the jth feedstuff, g/d
bacterial nitrogen, g/d
bacterial nitrogen of bacteria fermenting SC, g/d
bacterial nitrogen of bacteria fermenting NSC, g/d
liquid passage rate, % / h
bacterial peptide from the jth feedstuff, g/d
bacterial peptide nitrogen from the jth feedstuff, g/d
bacterial peptide nitrogen retained from the jth feedstuff, g/d
nonprotein nitrogen intake, g/d
NSC bacteria ammonia nitrogen retained from the jth feedstuff g/d
SC bacteria ammonia nitrogen retained from the jth feedstuff, g/d
peptide uptake rate by the NSC bacteria when ionophores are fed, .07 g
peptide uptake rate by the NSC bacteria, % / g of NSC bacteria/h
Ruminal ammonia nitrogen, g
bacterial true protein escaping the rumen by the jth feedstuff, g/d
bacterial cell wall protein escaping the rumen by the jth feedstuff, g/d
bacterial nucleic acid nitrogen escaping the rumen by the jth feedstuff, g/d
bacterial carbohydrate escaping the rumen by the jth feedstuff, g/d
bacterial fat escaping the m e n by the jth feedstuff, g/d
bacterial ash escaping the rumen by the jth feedstuff, g/d
3558
RUSSELL ET AL.
cell wall structures and nucleic acids (Ling and
Buttery, 1978; Van Soest, 1982). The CNCPS assumes that nucleic acid N is 15% of the total
microbial N (Purser and Buechler, 1966). Cell wall
protein is assumed to be 25% of the microbial
protein (Bergen et al., 1967). The remainder is
available true protein. Ruminal organisms also
contain significant quantities of lipid, storage
carbohydrate, and minerals (Van Soest, 1987).
Ruminal Fermentation Model
The CNCPS uses the inputs of feed intake, feed
composition, and the inherent rate of carbohy
drate and protein degradation to calculate the
lIU"lina1 ammonia and peptide production, the
amounts of carbohydrate, protein, and peptide
escaping ruminal degradation, and the amount of
metabolizable energy that will be available to the
animal; these calculations are summarized in a
companion paper (Sniffen et al., 1992). See Table 1
for a glossary of terms used in the CNCPS model.
In the-CNCPS, bacterial yield is dependent on the
rate of carbohydrate fermentation a d , hours-'), an
input, theoretical maximum growth yield CYG,
grams of bacteria per gram of carbohydrate
fermented), the maintenance energy coefficient
(KM, grams of carbohydrate per gram of bacteria
per hour), and the input, forage NDF (RFNDF,
grams):
if RFNDF e 20 then YG1 = YG1-1
- ((20 - RFNDF)..025))
if RFNDFe20then YG2 = YG2.1
- ((20 - RFNDF)*.O25))
I/Ylj = (KMl/Kdsj) + (I/YG1)
1/Y2j = (KM2/Kd4jl + (1/YG21
I/Y3j = (KM2/Kd5jl + (I/YG2)
Then the yield of NSC bacteria is adjusted
according to the availability of ruminal peptides
PEP) and the input of ruminally degraded carbohydrate (RDC):
RATIOj
=
IMPj
=
RDPEPj/(RDCAj + RDCBIj
+ RDPEPj)
EXP(.4O4* LnRATIOj 100)
+
1.942)
SCBACTj = Y'j'RDCB2j
Y2j = Yzj'(1 + IMPj*.O1)
Y3j = Y3j.(l + IMPj*.O11
NSCBACTj = Yzj'RDCAj + Y3j'RDCBlj
The total flow of bacteria (BACT) is the s u m of
NSC and SC bacteria:
BACTj
NSCBACTj
+
NSCBACTNj
.LO.NSCBACTj
=
1
.O * SCBACTj
SCBACTNj
In the CNCPS, the input RDPA (ruminally degradable protein) is converted to peptides, RDPEPj, a t a
rate that is the inherent property of the protein
(see Sniffen et al., 1992). RDPEPj can be taken up
(KUP)by NSC bacteria or passed out of the rumen
a t the fluid dilution rate (Kl), an input of the model
(see Sniffen et al., 1992):
PEPUPj
=
((KUP NSCBACT)/
((KUP NSCBACT) + K1)) RDPEPj)
4
If ionophores are fed the peptide uptake rate can
be decreased by 34olO:
KUPI = KUPq.66
peptides taken up are 160,,o
PEPUPN
=
N:
PEPUPj/6.25
The CNCPS then calculates ruminal N balance.
Because peptides can only provide 66% of the N to
NSC bacteria, then
PEPUPNR = .66 * NSCBACT
when PEPUPN is =eater
(.66 NSCBACTN); otherwise,
PEPUPNR = PEPUPN
than
and
NSCAMMNR = NSCBACTN - PEPUPNR
SCAMMNR = SCBACTN
ammonia
and
(RAN) is
RAN = (Y + RDPA + NPN) - (PEPUPNR +
NSCAMMNR + SCAMMNR)
Bacteria (both SC and NSC) escaping the rumen
(REB) are 62.5% CP (of which 25% is cell wall
nitrogen [CWI and 15% is nucleic acid nitrogen
[NAD,2 1% carbohydrate (CHOI, 12% fat (FAT) and
4.4% ash CASH):
REBTPj =
REBCWj =
REBNAj =
REBCHOj
REBFATj =
REBASHj =
.6O. .625 BACTj
.25 * .625 * BACTj
.15 .625 BACTj
.21*BACTj
.12*BACTj
.044 BACTj
SCBACTj
The bacteria (BACTI are 10% N:
Undegraded carbohydrate and protein, peptides,
and ruminal bacteria for each individual feedstuff
CARBOHYDRATE AND PROTEIN SYSTEM FOR CATTLE
n
\
V
ply needed by growing calves. Similar results were
4
0
observed
in field studies with beef and dairy cattle 0
300
200
100
0
p!
n
3559
0
100
200
300
400
OBSERVED BACTERIAL N (g/dey)
Figure 3. The relationship between observed flows of
microbial nitrogen from the rumen and those predicted
by the Cornell Net Carbohydrate and Protein System.
Data were taken from Robinson and Sniffen (1985),
Garret et al. (19871, Glenn et al. (1989),McCarthy et al.
(1989), and Song and Kennelly (1989). The regression
line (not shown) had an r2 .88, a slope of .94, and an
intercept of -12 g of N/d. The high N flows were
observed in lactating dairy cows, whereas the low N
flows were from trials with steers.
-
pass out of the rumen. Their absorption is discussed in companion papers (Fox et al., 1992;
Sniffen et al., 1992).
Evaluation
The literature contains an abundance of performance data, but few studies have 11 described the
diets so that the rates of carbohydrate fermentation and protein degradation can be estimated
with accuracy and 2) reported the flow of microbial
protein and undegraded feed to the lower gut.
Thus, only certain studies could be used to
validate the model (Robinson and Sniffen, 1985;
Garrett et al., 1987; Glenn et al., 1989; McCarthy et
al., 1989; Song and Kennelly, 1989). These studies
used Holstein cows and steers that were fed corn,
barley, and silage-based diets at various intakes.
The regression coefficient of the observed and
predicted BACTN was .94 k2 = .881 and the slope
was .94. The slightly negative intercept (-12 g of N/
d) indicates that the model is biased toward a n
underprediction of BACTN, but this bias is minor
(Figure 3).
We have used this submodel to design feeding
experiments and evaluate the results (Fox et al.,
1990b). The model correctly predicted metabolizable protein (microbial plus undegraded feed) sup-
(Fox et al., 1990a). We are encouraged by these
results, and feel that the CNCPS can account for
the effect of variations in feed carbohydrate and
protein fractions on microbial yield, feed protein
escaping ruminal degradation, and carbohydrate
utilization in the rumen.
Any scheme of ration formulation depends on
an accurate and meaningful description of dietary
ingredients. Representative feeds are listed in a
companion paper (Sniffen et al., 18921, but the user
may encounter feeding situations in which an
appropriate comparison is not readily available.
Because rate values are not yet an aspect of most
commercial analyses, the user may think that the
model is not applicable. Our experience, however,
indicates that the model can still be valuable. The
user can use present performance to estimate and
validate the inputs. After the inputs have been
validated, the user then evaluates the proposed
changes to see whether the potential allocation of
diet is beneficial or detrimental.
Implications
Traditional schemes of ration balancing for
ruminants have been based on equations that
have attempted to predict nutrient availability
and animal productivity. In many cases these
empirical equations have not provided a n accurate representation of ruminal fermentation and
nutrient availability. The Cornel1 Net Carbohydrate and Protein System has a fermentation
submodel model that compares the rate of carbohydrate fermentation with the rate of protein and
predicts the ruminally digestible organic matter,
microbial protein synthesis, ammonia production,
and the flow of undigested feed to the lower gut.
Based on the evaluations, it seems that this
submodel gives a n accurate assessment of
microbial growth and fermentation end products.
Literature Cited
Baldwin, R. L., J. J. Koong, and M. J. Ulyatt. 1977. A dynamic
model of ruminant digestion for evaluation of factors affecting nutritive value. Agric. Systems 2:255.
Bauchop, T. 1979. Rumen anaerobic b g i of cattle and sheep.
Appl. Environ. Microbiol. 38: 148.
Bergen, W. G., D. B. Purser, and J. H. Cline. 1967. Enzymatic
determination of the protein quality of individual rumen
bacteria. J. Nutr. 92:357.
Bird, S.H., and R. A. Leng. 1978. The effects of defaunation of
the rumen on the growth of cattle on low-protein high
energy diets. Br. J. Nutr. 40:163.
Bladen, H. A., M. P. Bryant, and R. N. Doetsch. 1961. A study of
bacterial species from the rumen which produce ammonia
from protein hydrolyzate. Appl. Microbiol. 9:175.
3560
RUSSELL ET AL.
Bryant, M. P. 1973. Nutritional requirements of the predominant
rumen cellulolytic bacteria. Fed. Proc. 32:1809.
Chen, G.. and J. B. Russell. 1988. Fermentation of peptides and
amino acids by a monensin-sensitive ruminal peptostrep
tococcus. Appl. Environ. Microbiol. 54:2742.
Chen, G., and J. B. Russell. 1989. More monensin-sensitive,
ammonia producing bacteria from the rumen. Appl. Environ. Microbiol. 55:1052.
Chen, G., J. B. Russell, and C. J. Sniffen. 1987a. A procedure for
measuring peptides in rumen fluid and evidence that p e p
tide uptake can be a rate-limiting step ruminal protein
degradation. J. Dairy Sci. 70:1211.
Chen, G., C. J. Sniffen, and J. B. Russell. 1987b. Concentration
and estimated flow of peptides from the rumen of dairy
cattle Effects of protein quantity, protein solubility and
feeding frequency. J. Dairy Sci. 70:983.
Chen, M., and M. J. Wolin. 1979. Effect of monensin and
lasalocid-sodium on the growth of methanogenic and Nmen saccharolytic bacteria. Appl. Environ. Microbiol. 38~72.
Citron, A., A. Breton, and G. Fonty. 1987. Rumen anaerobic
fungi. Bull. Inst. Pasteur 85:329.
Coleman, G. S.1980. Rumen ciliate protozoa. Adv. Parisitol. 18:
121.
Dennis, S. M., T. G. Nagaraja, and E. E. Bartley. 1981. Effects of
lasalocid or monensin on lactate-producing or using bacteria. J. Anim. Sci. 52:418.
Dinius, D. A., M. E. Simpson, and P. B. Marsh. 1976. Effect of
monensin fed with forage on digestion and the ruminal
ecosystem of steers. J. h i m . Sci. 42:229.
Fox, D. G., C. J. Sniffen, J. D. O’Connor, J. B. Russell, and P. J.
Van Soest. 1992. A net carbohydrate and protein system for
evaluating cattle diets: 111. Cattle requirements and diet
adequacy. J. Anim. Sci. 703578.
Fox, D. G., C. J. Sniffen, J. D. OConnor, J. B. Russell, P. J. Van
Soest, and W. Chalupa. 199Oa. Formulating cattle rations
for optimum carbohydrate, ammonia, peptides and amino
acid concentrations, using The Cornell Net Carbohydrate
and Protein System. Roc. le90 Annu. Mtg. of the California-Nevada Chapter of the Am. Registry of Prof. Anim.
Scientists.
Fox, D. G., C. J. Sniffen, J. D. O’Connor, J. B. Russell, P. J. Van
Soest, and W. Chalupa. 199Ob. Using the Cornell Net Carbohydrate and Protein System to predict the effects of
metabolic modifiers on the metabolizable energy and protein requirement of growing cattle. Proc. Cornell Nutr.
Cod. p 28.
Garrett, J. E., R. D. Goodrich, J. C. Meiske, and M. D. Stem.
1987. Influence of supplemental nitrogen source on digestion of nitrogen, dry matter, and organic matter and on in
vivo rate of ruminal protein degradation. J. Anim. Sci. 64:
1801.
Glenn, B. P., G. A. Varga, G. B. Huntington, and D. R. Waldo.
1989. Duodenal nutrient flow and digestibility in Holstein
steers fed formaldehyde- and formic acid-treated alfalfa or
orchardgrass silage at two intakes. J. Anim. Sci. 67613.
Goering, H. K., and P. J. Van Soest. 1970. Forage fiber analyses
(apparatus, reagents, procedures, and some applications).
Agric. Handbook 379. AFlS, USDA, Washington, DC.
Hespell, R. B., and M. P. Bryant. 1979. Emciency of rumen
microbial growth: Influence of some theoretical and e x
perimental factors on YAP. J. Anim. Sci. 49:1640.
Hino, T., and J. B. Russell. 1985. The effect of reducing equivalent disposal and NADH/NAD on the determination of
amino acids by intact and cell-free extracts of rumen
microorganisms. Appl. Environ. Microbiol. 50:1368.
Hungate, R. E. 1966. The Rumen and its Microbes. Academic
Press, New York.
Isaacson, H. R.,F. C. Hinds, M. P. Bryant, and F. N. Owens.
1975. Efficiency of energy utilization by mixed rumen bacteria in continuous culture. J. Dairy Sci. 58:1645.
Krishnamoorthy, U., T. V. Muscato, C. J. Sniffen, and P. J. Van
Soest. 1982. Nitrogenous fractions in feedstuffs. J. Dairy
Sci. 65:217.
Ling, J. R.,and P. J. Buttery. 1978. The simultaneous use of
ribonucleic acid, 358, 2,6 diaminopimelic acid and 2
aminoethylphosphoric acid as markers of microbial nitrogen at the duodenum of sheep. Br. J. Nutr. 39:165.
Mangan, J. L. 1972. Quantitative studies on nitrogen
metabolism in the bovine rumen. Br. J. Nutr. 27~281.
McCarthy, R.D., Jr., T. H. Klusmeyer, J. 3. Vicini, J. H. Clark,
and D. R. Nelson. 1989. Effects of source of protein and
carbohydrate on ruminal fermentation and passage of
nutrients to the small intestine of lactating cows. J. Dairy
Sci. 72~2002.
Nocek, J., and J. B. Russell. 1988. Protein and carbohydrate as
an integrated system. Relationship of ruminal availability
to microbial contribution and milk production. J. Dairy Sci.
71:2070.
Nocek, J. E. 1988. In situ and dther methods to estimate ruminal
protein and energy digestibility: A review. J. Dairy Sci. 71:
2051.
Nolan, J. V. 1975. Quantitative models of nitrogen metabolism
in sheep. In: I. W. MacDonald and A.C.I. Warner (Ed.)
Digestion and Metabolism in the Ruminant. p 416. Univ. of
New England Publishing Unit, Armidale, Australia.
NRC. 1985. Ruminant Nitrogen Usage. National Academy
Press, Washington, DC.
NRC. 1989. Nutrient Requirements of Dairy Cattle. National
Academy Press, Washington, DC.
0rskov, E. R. 1976. The effect of processing on digestion and
utilization of cereals by ruminants. Proc. Nutr. SOC.35:245.
Pirt, S.J. 1965. The maintenance energy of bacteria in growing
cultures. Proc. R. SOC.Lond. Ser. B Biol. Sci. 163:224.
Purser, D. B., and S.M. Buechler. 1966. Amino acid composition
of rumen organisms. J. Dairy Sci. 49:81.
Reichel, J. R.,and R.L. Baldwin. 1976. A rumen linear program
ming model for evaluation of concepts of rumen microbial
function. J. Dairy Sci. 59:439.
Robinson, P. H., and C. J. Sniffen. 1985. Forestomach and whole
tract digestibility for lactating dairy cows as influenced by
feeding frequency. J. Dairy Sci. 68:857.
Russell, J. B. 1983. Fermentation of peptides by Bacteroides
rurninicola B14. Appl. Environ. Microbiol. 45:1586.
Russell, J. B. 1984. Factors influencing competition and compositions of rumen bacterial flora. Proc. Symp. Herbivore
Nutrition in Sub-tropics and Tropics. p 313. The Science
Press, Craighall, South Africa.
Russell, J. B., and R.L. Baldwin. 1981. Microbial rumen fermentation. J. Dairy Sci. 64:1153.
Russell, J. B., and R. B. Hespell. 1985. Regulation of lactate
production in Streptococcus bovis A spiraling effect that
contributes to rumen acidosis. J. Dairy Sci. 68:1712.
Russell, J. B., and S.A. Martin. 1984. Effects of various methane
inhibitors on the fermentation of amino acids by mixed
rumen microorganisms in vitro. J. Anim. Sci. 59:1329.
Russell, J. B., and C. J. Sniffen. 1984. Effect of carbon-4 and
carbon-5 volatile fatty acids on growth of mixed rumen
bacteria in vitro. J. Dairy Sci. 67987.
Russell, J. B., C. J. Sniffen, and P. J. Van Soest. 1983. Effect of
carbohydrate limitation on degradation and utilization of
casein by mixed rumen bacteria. J. Dairy Sci. 66:763.
Russell, 3.B., H. J. Strobel, and G. Chen. 1988. The enrichment
and isolation of a ruminal bacterium with a very high
specific activity of ammonia production. Appl. Environ.
Microbiol. 54:872.
Russell, J. B., and R. J. Wallace. 1988. Energy yielding and
consuming reactions, In: P. N. Hobson (Ed.)pp 185-216. The
Rumen Microbial Ecosystem. Elsevier Applied Science,
London.
Slyter, L. L. 1976. Influence of acidosis on rumen function. J.
CARBOHYDRATE AND PROTEIN SYSTEM FOR CATTLE
Anim. Sci. 43910.
Sniffen, C. J., J. D. O’Connor, P. J. Van Soest, D. G. Fox, and J.
B. Russell. 1992. A net carbohydrate and protein system for
evaluating cattle diets: 11. Carbohydrate and protein availability. J. Anim. Sci. 70:3562.
Song, M. K., and J. J. Kennelly. 1989. Effect of ammoniated
barley silage on ruminal fermentation, nitrogen supply to
small intestine, ruminal, and whole tract digestion, and
milk production of Holstein cows. J. Dairy Sci. 72:2981.
Strobel, H. J., and J. B. Russell. 1886. Effect of pH and energy
spilling on bacterial protein synthesis by carbohydrate
limited cultures of mixed rumen bacteria. J. Dairy Sci. 69:
2941.
Tamminga, S.1979. Protein degradation in the forestomachs of
ruminants. J. Anim. Sci. 49:1615.
3561
Van Nevel, C. J., and D. I. Demeyer. 1977. Effect of monensin on
rumen metabolism in vitro. Appl. Environ. Microbiol. 34:
251.
Van Soest, P. J. 1987. Nutritional Ecology of the Ruminant.
Cornell University Press, Ithaca, NY.
Waldo, D. R., and L. W. Smith. 1872. Model of cellulose disappearance from the rumen. J. Dairy Sci. 55:472.
Whetstone, H. D., C. L. Davis, and M. P. Bryant. 1981. Effect of
monensin on the breakdown of protein by ruminal microorganisms in vitro. J. Dairy Sci. 53:803.
Winters, K. A., R. R. Johnson, and B. A. Dehority. 1964.
Metabolism of urea nitrogen by mixed cultures of rumen
bacteria grown on cellulose. J. Dairy Sci. 47:793.
Wright, D. E., and R.E. Hungate. 1967. Amino acid concentrations in rumen fluid. Appl. Microbiol. 15:148.