![]() |
|
|
Animal Science Department, University of California at Davis, Davis, CA 95616 and * The University of Reading, Department of Agriculture, Earley Gate, Reading RG6 6AT, UK
| ABSTRACT |
|---|
|
|
|---|
KEY WORDS: recycling channeling protein fractional synthesis rate protein turnover rodents
| INTRODUCTION |
|---|
|
|
|---|
The choice of precursor pool used to calculate fractional synthesis
rates is based on previous knowledge of the protein synthesis process.
In a two-pool model of protein synthesis, the free amino acid
(precursor) pool is considered to be homogeneous. All of the amino
acids in the extracellular, intracellular and plasma pools are
available for protein synthesis. If only a portion of the free amino
acids are available for protein synthesis, differences among the
specific radioactivities of the aminoacylated-tRNA pool and the
intracellular, plasma or extracellular pools could have a significant
effect on estimates of FSR (Matthews and Cobelli 1991
). Channeling, the
transportation of amino acids for protein synthesis from the
extracellular pool to the aminoacyl-tRNA pool without mixing with the
intracellular pool, implies that the extracellular pool specific
radioactivity can be used as the precursor pool to calculate FSR.
However, if amino acids from protein degradation are being reused for
protein synthesis without mixing with the general pool of amino acids
(recycling), the aminoacyl-tRNA pool specific radioactivity will be
diluted, resulting in underestimates of protein synthesis rate when the
specific radioactivity of the extracellular pool is used to calculate
FSR. The specific radioactivity of aminoacyl tRNA will be different
from both the intracellular and extracellular pool specific
radioactivities. Therefore, estimates of the percentages of channeling
and recycling will help define which pool specific radioactivity best
approximates aminoacyl-tRNA specific radioactivity.
When the synthesis rate is measured for a general population of
proteins, each being synthesized and degraded at different rates, the
observed rate could depend on the length of time over which the
measurement was made (Obled et al. 1991
). Rapidly turning over
proteins may incorporate and release the radiolabeled amino acid many
times during an experiment, whereas slow turnover proteins may
incorporate very little radiolabeled amino acid. The effect on the FSR
of multiple protein pools turning over at different rates has never
been examined quantitatively (Matthews and Cobelli 1991
).
In the companion paper (Johnson et al. 1999
), a
mechanistic model of protein turnover was described and evaluated. The
model predicts specific radioactivity changes in protein and free
leucine pools using the flooding dose, pulse dose and continuous
infusion methods. The objective of this work was to use the model to
evaluate the effect of recycling, channeling and multiple protein pools
turning over at different rates on changes in the specific
radioactivities of the pools. Model sensitivity to changes in each flux
is evaluated to determine which fluxes have the greatest influence on
specific radioactivity estimates for each method independent of changes
in percentages of recycling and channeling. Then the model is used to
simulate the effects of high and low channeling and recycling on
predictions of specific radioactivity for each method. Finally,
quantitative effects of multiple protein pools, recycling and
channeling on the estimation of FSR are evaluated.
| SENSITIVITY ANALYSIS |
|---|
|
|
|---|
The whole-body rodent model used in this paper was described previously
(Johnson et al. 1999
). Protein synthesis rate was
assumed to equal protein degradation rate and intake
(FOE) was assumed to equal oxidation
(FIO). The only pools that remained constant in
size were the aminoacyl-tRNA pool (QT) and the protein
pools (QS, QM, QF). Therefore the
flow of amino acids from aminoacyl tRNA to the intracellular pool
(FTI) was a balance equation to keep
QT in steady state. The percentage of recycling (PR)
determined how much amino acid from protein degradation flowed to the
aminoacyl-tRNA pool (FFT,
FMT, FST) and how much
flowed to the intracellular amino acid pool (1-PR;
FFI, FMI,
FSI). The percentage of channeling (PC) defined
how much amino acid was supplied by the extracellular pool
(FET) and how much was supplied by the
intracellular pool (1-PC; FIT).
Each flux and pool size was individually increased and decreased by
25% (Tables 13)
. Therefore one row in a table represents one model
simulation. The percentages of change in specific radioactivity for
each pool as a result of an increase or decrease of 25% in the model
fluxes and pools sizes were computed for each method. Because the same
flux equations and pool sizes were used for each method, comparison of
the percentages of change in each pool specific radioactivity indicates
which fluxes or pool sizes have the greatest influence on model
estimates of specific radioactivity when each method is used. The
influence of measurement error and error associated with estimating
fluxes on specific radioactivity estimates independent of changes in
the percentages of recycling and channeling can also be examined.
Twenty-five percent is a substantial change in model rates and pool
sizes. If experimental error is considered to be within 15%, changes
in specific radioactivities of <15% would probably not be significant
experimentally. The percentage of recycling was set at 75% for the
fast pool only; the percentage of channeling was also set at 75%.
Values for an increase in the flux from the intracellular pool to the
extracellular pool (FIE) could not be determined
because a small increase in this flux resulted in negative specific
radioactivities and pool sizes. Therefore a weakness of the model is in
the estimate of FIE because it is impossible to
measure separately from FEI and therefore can
only be estimated by fitting data to both FEI
and FIE.
|
|
|
Only the protein pool specific radioactivity changed more than 15%
(Table 1
)with changes in the slow protein turnover rate (3040% change), the
medium protein turnover rate (1520% change) and the extracellular
pool size (17%). The other pool specific radioactivities were most
sensitive to the extracellular pool size (12%), recycling (79%) and
intake (FOE; 67%).
Continuous infusion sensitivity.
The continuous infusion method was very sensitive to intake
(Table 2
)with changes in specific radioactivities of all pools of 3550%.
Intake is easy to measure and so should not be a weakness in the model
estimates of specific radioactivity. The specific radioactivities of
the protein, extracellular, intracellular and aminoacyl-tRNA pools were
also sensitive to the turnover rate of the slow protein turnover pool
(590%) and medium protein turnover pool (550%). The rate of
recycling caused changes between 10 and 15% in the specific
radioactivities of the protein and aminoacyl-tRNA pools.
Pulse dose sensitivity.
The greatest changes in specific radioactivities of the total protein
pool (Table 3
)were from changes in the slow and medium protein turnover rates
(20120% and 1030%, respectively). Changes in protein pool size
also caused large changes in the specific radioactivities of the
intracellular, aminoacyl-tRNA and extracellular pools (2040%).
Changes in intake affected the specific radioactivities of the
intracellular pool (18%) and protein pool (13%).
Summary of sensitivities (all methods).
All three methods were very sensitive (>25% change in specific
radioactivity) to the rates of protein turnover in the slow and medium
protein turnover pools. The rates used in the model were based on
extensive data from the literature on estimates of FSR (Johnson et al. 1999
). The values used are probably good estimates of
protein turnover rates. The continuous infusion method was also very
sensitive to intake (FOE), and the pulse dose
method was sensitive to protein pool size (QP). The
flooding dose method was the least sensitive to changes in fluxes and
pools, followed by the continuous infusion and pulse dose methods.
| BEHAVIORAL ANALYSIS OF MODEL IN RESPONSE TO CHANGES IN RECYCLING AND CHANNELING |
|---|
|
|
|---|
The time course of the specific radioactivities of four of the six pools of the model was examined to compare methods of estimating FSR with different rates of channeling and recycling. Recycling is the preferential use of amino acids from protein degradation for protein synthesis without mixing with the extracellular or intracellular free amino acid pools. Because protein pools contain large amounts of unlabeled amino acid, it would be expected that the greater the rate of recycling, the more similar the specific radioactivities of the precursor pool and the protein pools would become. In our simulations, recycling was allowed only in the fast turnover protein pool and set at either 100% recycling (100R) or 0% recycling (0R). Recycling was only in the fast protein turnover pool because it is the most likely to affect specific radioactivity measurements over short experimental time periods and as the smallest pool, will be a conservative estimate of the influence of recycling. Channeling is the flow of amino acids from extracellular sources to protein synthesis without mixing with the intracellular pool of amino acids. As the rate of channeling increases, the more similar the specific radioactivity of the precursor pool will be to the extracellular pool. Channeling was set at 100% (100C) or 0% (0C).
Methods.
Three experimental methods for determining FSR were examined. The
experimental protocol according to Bernier and Calvert (1987)
for a 30-g mouse was used for the flooding dose method
(111 MBq 14C Leu/30 µmol Leu). The experimental protocols
of Pomposelli et al. (1985)
and Peters and Peters (1972)
were used for the continuous infusion and pulse dose
methods, respectively. Values for the pulse dose (111 MBq
14C Leu/0.0091 µmol Leu) and continuous infusion (37 MBq
14C Leu/0.02 µmol Leu for 180 min) specific
radioactivities had to be adjusted to a 30-g mouse.
Flooding dose.
In the flooding dose method (Fig. 1
),the specific radioactivities of the intracellular and extracellular
pools remained close after a 5-min equilibration period. However,
aminoacyl-tRNA specific radioactivities at 100% recycling and
channeling and 100% recycling and 0% channeling were much lower.
Therefore, at high levels of recycling, the specific radioactivity of
the aminoacyl-tRNA pool cannot be estimated by either extracellular or
intracellular pool specific radioactivities. Although aminoacyl-tRNA
specific radioactivity was dependent on the percentage of recycling,
the protein specific radioactivity was dependent on the rates of
channeling and recycling. Protein specific radioactivity was high with
0% recycling and even higher with high channeling. Therefore, even
though the intracellular or extracellular specific radioactivity may
approximate the aminoacyl-tRNA specific radioactivity, the rates of
recycling and channeling (even in a small protein pool) could cause an
overestimate of the specific radioactivity of the entire protein pool.
|
|
|
| SIMULATION OF CHANGES IN PROTEIN TURNOVER RATES, RECYCLING AND CHANNELING BY THE MODEL |
|---|
|
|
|---|
Heterogeneous protein pool.
FSR was estimated at different time points using the pulse dose and
flooding dose methods (Table 4
).For the first run, fractional rates of protein synthesis and
degradation for the fast protein pool were set to 104%/d. In the
second run, fractional rates of protein synthesis and degradation for
the fast protein pool were doubled (210%/d). In the flooding dose
method, FSR was relatively constant from 5 to 60 min. However, after 60
min, the estimate of FSR decreased. The specific radioactivity of the
amino acid pool (sA) decreased and the specific
radioactivity of the protein pool (sP) increased (data not
shown). Therefore it appears that the flooding dose stabilized the
specific radioactivity of the precursor pool, causing the specific
radioactivity of the protein pool to increase slightly. In the pulse
dose method, FSR increased from 5 min to 24 h. After 24 h,
amounts of labeled amino acid in both the amino acid and protein pools
were very small and sP had decreased less than sA(data not shown). Therefore an increase in the length of time
over which FSR was measured decreased the estimate of FSR by the
flooding dose method but increased the estimate of FSR by the pulse
dose method.
|
Recycling.
The effect of recycling on fits to the data of Bernier and Calvert (1987)
and Obled et al. (1991)
was
examined next. Fluxes (FIO,
FOE, FIE), percentages of
protein synthesized (KSF,
KSM, KSS), percentage
channeling (PC) and percentage recycling (PR) were fitted using
Simusolv (Dow Chemical 1990
) to have the model produce
specific radioactivities as close to the data as possible. FSR
represented the total fractional synthesis rate predicted by the model
based on KSS, KSM and
KSF using the flooding dose method. Using the
fluxes determined previously (Johnson et al. 1999
), the
percentage of recycling was then forced to fit the Bernier and Obled
data sets. Only the fast protein turnover pool was allowed to recycle.
The results are listed in Table 5
.KRF was the percentage of leucine from protein
degradation in the fast turnover protein pool that recycled to
QT. KSF was the FSR of the fast
protein turnover pool estimated by fitting the data. Recycling did not
improve the fit of the data from Obled et al. (1991)
and
Bernier and Calvert (1987)
. The percentage of error for
the specific radioactivity predictions for free leucine and protein
from the Bernier and Obled data stayed approximately the same.
Recycling decreased intake (FOE) from 0.843 to
0.741 for the Bernier data set. The percentage of protein synthesized
per minute in the fast protein turnover pool and the total (moles
protein synthesized per minute) increased for both data sets. Overall,
however, fractional synthesis rates were unchanged. Therefore,
recycling enabled more protein to be synthesized without increasing the
FSR calculated experimentally.
|
|
The changes in FSR calculated by the model due to channeling
(Fig. 5
)were much greater for the pulse dose (530%/d) than the flooding dose
method (3540%/d). In the previous figures (Figs.13)
, the pulse
dose resulted in a greater difference in protein specific radioactivity
than the other two methods at the end of the simulation (30 min).
Therefore larger differences in estimates of FSR would be expected.
However, in the flooding dose and continuous infusion figures (Figs. 1 and 3)
, there were large differences in specific radioactivities among
the possible precursor pools for protein synthesis, QE,
QI and QT. Therefore the choice of precursor
pool for the measurement of FSR would necessarily result in very
different estimates.
|
|
The effect of different sources of leucine for protein synthesis on
fits to the data of Bernier and Obled was examined next. Fluxes
(FIO, FOE,
FIE), percentage of protein synthesized
(KSF, KSM,
KSS), percentage of channeling (PC) and
percentage of recycling (PR) were fitted using Simusolv (Dow
Chemical 1990
) to have the model produce specific
radioactivities as close to the data as possible (Johnson et al. 1999
). Zero percentage channeling (FET =
0; PC = 0) was then forced to fit the data sets of Bernier and
Obled. The results are presented in Table 7
.Errors of prediction of specific radioactivities of the free amino acid
pools were approximately the same as in the previous solutions
(Johnson et al. 1999
) as were the fluxes, excepting
those for FET and FIT.
|
| DISCUSSION |
|---|
|
|
|---|
The two-pool whole-body model of protein synthesis proposed by
Waterlow et al. (1978)
assumed that the product pool
contained a mix of proteins that turned over at the same rate. Although
it was known that proteins are synthesized and degraded at different
rates, they assumed that over a short experimental time period, the
difference in turnover rates would not affect estimates of FSR.
Heterogeneous protein pools could affect FSR by decreasing the specific
radioactivity of the amino acid pool more rapidly than a single
homogeneous protein turnover pool. In the model, when the FSR of the
fast turnover pool was doubled (from 104 to 210 %/d), the whole-body
FSR estimate using the flooding dose technique decreased more (by
55%/d). Therefore extrapolating a 15-min measurement (in flooding
dose) or a 3-h measurement (in continuous infusion) to a daily
measurement of FSR would only increase the error associated with the
estimate of FSR. In addition, all of the methods of estimating FSR were
most sensitive to the rate of turnover of the fast and medium protein
pools. Therefore differences in rates of protein turnover do change FSR
estimates, and a uniform time of measurement within methods is critical
for consistent estimates of FSR that may be incorrect in either case.
Channeling.
Another assumption made when measuring FSR is that the amino acid pool is homogeneous or that channeling does not occur. At high rates of channeling, the specific radioactivity of the extracellular pool approximates the specific radioactivity of the aminoacyl-tRNA pool. Similarly, when the intracellular pool is the main source of amino acid for charging, the aminoacyl-tRNA specific radioactivity becomes approximately equal to the specific radioactivity of the intracellular pool. Therefore, when channeling is high, pulse and flooding dose estimates are equivalent, especially after 5 min. Because the measurement of the specific radioactivity of the aminoacyl-tRNA pool is difficult, in some cases the intracellular or extracellular pool specific radioactivities can be used to approximate the aminoacyl-tRNA pool specific radioactivity. If there is recycling, however, the specific radioactivity of the aminoacyl-tRNA pool will be intermediate or below the extracellular and intracellular specific radioactivities. Therefore, for all methods, the rate of recycling and channeling must be known in order to determine whether the extracellular, intracellular or aminoacyl tRNA pool specific radioactivity should be used as the precursor pool to estimate FSR.
The flux and specific radioactivity changes predicted by the model from
the data sets of Bernier and Obled indicate that 100% of the leucine
for tRNA charging (protein synthesis) is from the extracellular pool
(Table 7)
. The added dilution by unlabeled leucine from the flooding
dose in the initial QI appears to be great enough to
prevent the specific radioactivity of the protein pool from increasing
fast enough to match the observed values.
The model prediction of a high channeling rate may not be true with all
methods or for all physiologic states. Because the flooding dose was
used, leucine was not in short supply and the use of the intracellular
pool (if it is thought of as a "buffer") was probably not
necessary. The high rate of channeling observed with the flooding dose
method may not be consistently true when other methods are used. The
estimate of channeling is dependent on a high specific radioactivity of
the protein pools and a higher specific radioactivity of the source of
amino acid for protein synthesis. The estimates of the specific
radioactivities of the intracellular and extracellular pools by the
model are dependent on the fluxes estimated from the flooding dose
data. Because the flooding dose method uses concentrations of leucine
that are higher than physiologic levels, the resulting predictions of
fluxes may not be the same between methods. For instance, using the
Bernier data, the model predicted twice the level of intake for a mouse
and four times the rate of oxidation compared with values from the 26-g
reference mouse (Johnson et al. 1999
). The high levels
were necessary to dilute the specific radioactivity of the free amino
acid pools to match the levels in the Bernier data. Therefore oxidation
could not be equal to intake in the flooding dose experiment. The
radiolabeled leucine may also be used preferentially by the tRNA
synthetase enzymes for protein synthesis, causing higher specific
radioactivities in the protein pools (Hatch et al. 1995
). The conflicting predictions of the changes in
intracellular pool size from the data sets of Obled and Bernier could
also be a result of perturbations due to the use of a flooding dose.
Data from the continuous infusion and pulse dose methods in amino acid
deficient and balanced states are needed to confirm the high rate of
channeling.
Recycling.
The third assumption is that recycling does not occur or is not
sufficient to affect estimates of protein synthesis during short times
of measurement. At high rates of recycling, the specific radioactivity
of the aminoacyl-tRNA pool is lowered, which leads to an underestimate
of FSR. Similarly, at low rates of recycling, the specific
radioactivity of the aminoacyl-tRNA pool is closer to the extracellular
or intracellular pool, depending on how much channeling is occurring.
In Figures 13
, recycling was simulated as occurring only in the fast
turnover protein pool, and large differences were observed in the
specific radioactivities of the aminoacyl tRNA and protein. In
addition, the specific radioactivities of the aminoacyl tRNA and
protein were relatively sensitive to the percentage of recycling.
Therefore the amount of recycling has to be determined to assure
accurate estimates of FSR.
Data from the Obled and Bernier experiments were not adequate to
determine whether recycling was occurring. If recycling was forced into
the solution, only 11% was predicted to be occurring; this would be
difficult to separate from experimental error. The fact that the high
level of channeling was still maintained in the solution could have
caused the recycling prediction to be artificially low because the
flooding dose method cannot predict the rate of channeling if recycling
is high. Recycling did cause the predicted level of intake to decrease
to be closer to actual values and also caused the estimated FSR to
increase. Therefore more recycling would improve the fit of the
solution to the Bernier data. In addition, because the turnover rates
of the protein pools were based on estimates of FSR using the flooding
dose for the individual protein pools, they would underestimate the
actual synthesis rate if significant recycling was occurring. The data
of Bernier and Obled did not distinguish between fast, medium and slow
turnover pools; thus, all of the differences between actual and
predicted turnover rates were included in the slow turnover pool
estimate of FSR. The slow turnover pool was also the largest.
Therefore, small changes in the turnover rate of the slow pool would
affect the specific radioactivity much more than the other protein
pools (Tables 13)
. The predictions of synthesis rates from the data
of Obled and Bernier were even lower than the average synthesis rates
from the literature; therefore it would appear that more recycling may
have been possible.
Accurate estimates of FSR are dependent on the specific radioactivity of the pool that is the source of amino acid for tRNA charging and the amount of amino acids that are recycled to protein synthesis without mixing with the amino acid in the intracellular pool. From flooding dose data fit to the model, it appears that channeling is high and recycling is low. If recycling is occurring, it is not great enough to be distinguished from experimental error. The large dose of amino acid appears to perturb the fluxes so that recycling is relatively low compared with channeling or may not be necessary because of a large supply of leucine.
The model used to explore the implications of recycling, channeling and multiple protein pools turning over at different rates is a unique model that was built to represent protein turnover. Then data of Bernier and Obled were fit to the model to determine if it could represent specific radioactivity changes over time and predict FSR. In the first paper, the model was able to reproduce changes in specific radioactivities, estimate measured FSR and predict true FSR in rodents. In this paper, the model was used to predict the influence of recycling, channeling and multiple protein pools on changes in specific radioactivity and estimates of FSR. Because rates of recycling and channeling vary among tissues and with the amino acid used as a tracer, it is imperative that limitations associated with each of the methods are known for individual tissues and whole-body estimates.
| FOOTNOTES |
|---|
1 The costs of publication of this article were
defrayed in part by the payment of page charges. This article must
therefore be hereby marked "advertisement" in accordance with 18
USC section 1734 solely to indicate this fact ![]()
3 To whom reprint requests should be addressed. ![]()
4 Abbreviations used: FET
is flux of leucine from extracellular pool to aminoacyl-tRNA pool
(channeling), FEI is flux of leucine from
extracellular pool to intracellular pool, FFT is
flux of leucine from protein degradation in fast protein turnover pool
to aminoacyl-tRNA pool (recycling), FIE is flux
of leucine from intracellular pool to extracellular pool,
FIO is flux of leucine oxidized from
intracellular pool, FIT is flux of leucine from
intracellular pool to aminoacyl tRNA pool, FOE
is intake flux of leucine to extracellular pool,
FTI is flux of leucine from aminoacyl-tRNA pool
to intracellular pool, FTF,
FTM, FTS are fluxes of
leucine from aminoacyl-tRNA pool to fast protein turnover pool, medium
protein turnover pool and slow protein turnover pool; FSR is whole-body
protein fractional synthesis rate for a rodent; FSR EXP is whole-body
protein fractional synthesis rate calculated using the average combined
specific radioactivities of the intracellular, extracellular and
aminoacyl-tRNA pools as the precursor pool specific radioactivity;
KRF is the percentage of leucine from protein
degradation in the fast turnover protein pool which recycled to
QT; KSF is the protein synthesis
rate in the fast protein turnover pool (%/d);
KSM is the protein synthesis rate in the medium
protein turnover pool (%/d) ; KSS is the
protein synthesis rate in the slow turnover pool (%/d); PC is the
percentage of channeling; PR is the percentage of recycling;
QE is leucine in extracellular pool; QI is
leucine in intracellular pool; QP is leucine in protein
pool; QT is leucine in aminoacyl-tRNA pool. True FSR is the
whole-body protein fractional synthesis rate determined from the model
fluxes as (FSS + FSM +
FSF) 100 1440/QP; 100R is 100%
recycling, 0R is 0% recycling; 100C is 100% channeling, 0C is 0%
channeling. ![]()
Manuscript received March 26, 1998. Initial review completed August 7, 1998. Revision accepted December 1, 1998.
| REFERENCES |
|---|
|
|
|---|
1. Bernier J. F., Calvert C. C. Effect of a major gene for growth on protein synthesis in mice. J. Anim. Sci. 1987;65:982-995
2. Dow Chemical Co. (1990) Simusolv Modeling and Simulation Software. Dow Chemical Co., Midland, MI.
3. Hatch K. A., Sacksteder K. A., Treichel I. W., Cook M. E., Porter W. P. Early detection of catabolic state via change in 13C/12C ratios of blood proteins. Biochem. Biophys. Res. Commun. 1995;212:719-726[Medline]
4.
Johnson H. A., Baldwin R. L., France J., and Calvert C. C. Model of whole-body protein turnover based on leucine kinetics in rodents. J. Nutr. 1999;129:728-739
5.
Matthews D. E., Cobelli C. Leucine metabolism in man: lessons from modeling. J. Parent. Enteral Nutr. 1991;15:86S-89S
6. Mitchell and Gauthier Assoc. Inc ACSL: Advanced Continuous Simulation Language 1996 MGA Inc Concord, MA.
7. Obled C., Barre F., Arnal M. Flooding-dose of various amino acids for measurement of whole-body protein synthesis in the rat. Amino Acids 1991;1:17-27
8.
Peters T., Peters J. C. The biosynthesis of rat serum albumin. J. Biol. Chem. 1972;247:3858-3863
9. Pomposelli J. J., Palombo J. D., Hamawy K. J., Bistrian B. R., Blackburn G. L., Moldawer L. L. Comparison of different techniques for estimating rates of protein synthesis in vivo in healthy and bacteraemic rats. Biochem. J. 1985;226:37-42[Medline]
10. Waterlow J. C., Garlick P. J., Millward D. J. Protein Turnover in Mammalian Tissues and in the Whole Body 1978:119 North-Holland Amsterdam, The Netherlands.
This article has been cited by other articles:
![]() |
H. A. Johnson, R. L. Baldwin, K. C. Klasing, J. France, and C. C. Calvert A Rodent Model of Protein Turnover Used to Design an Experiment for Measuring the Rates of Channeling, Recycling and Protein Synthesis J. Nutr., December 1, 2000; 130(12): 3097 - 3102. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||