Journal of Nutrition LabDiet, Your World of Nutritional Answers

Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Johnson, H. A.
Right arrow Articles by Calvert, C. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Johnson, H. A.
Right arrow Articles by Calvert, C. C.
(Journal of Nutrition. 2000;130:3097-3102.)
© 2000 The American Society for Nutritional Sciences

Articles

A Rodent Model of Protein Turnover Used to Design an Experiment for Measuring the Rates of Channeling, Recycling and Protein Synthesis

H. A. Johnson1, R. L. Baldwin, K. C. Klasing, J. France* and C. C. Calvert2

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

1To whom correspondence should be addressed. E-mail: >HAJohnson@UCDavis.edu" locator-type="email">locator-type="email">HAJohnson@UCDavis.edu locator="" locator-type="email">


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 THE WHOLE-BODY MODEL
 THE TISSUE MODELS (TM)
 EXPERIMENTAL DESIGN
 DISCUSSION
 REFERENCES
 
We described previously a mechanistic model of whole-body protein turnover in rodents. Channeling was defined as the flow of amino acids from the extracellular compartment to aminoacyl tRNA and protein synthesis. Recycling was defined as the flow of amino acids from protein degradation to aminoacyl tRNA (protein synthesis) without mixing with the intracellular pool of amino acids. In this paper, the model is applied to tissues and whole body and is used to develop an experimental protocol for estimating protein fractional synthesis rate, recycling and channeling. Channeling, recycling and protein synthesis must be estimated simultaneously because changes in specific radioactivities over time are highly dependent on the rate of protein synthesis. Injection-specific radioactivities, body weights and experimental variation were used with the model to generate data at different rates of recycling and channeling. The data generated were then used to determine the best time points and experimental method to estimate percentages of recycling, channeling and protein synthesis rate by the iterative Method of Maximum Likelihood. Specific radioactivity at each time point was based on simulated data from three rodents at each of six time points. Predicted protein synthesis rates were within 5%/d of observed rates for all methods. Predicted rates of recycling and channeling were generally within 15% of observed rates except recycling in muscle at high channeling and high recycling. Standard deviations of the predictions of percentages of channeling and recycling were between 0.148 and 44.5% for the pulse dose method, 0.0655 and 197% for the continuous infusion method and 0.351 and 962% for the flooding dose method. The experimental design that yields the best estimates of channeling, recycling and protein synthesis is the pulse dose. Changes in amino acid specific radioactivities in the extracellular, aminoacyl tRNA and protein pools were greatest and should be measured at 2, 6, 10, 40, 70 and 100 min in the pulse method.

KEY WORDS: • rodents • experimental design • mathematical model • protein turnover


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
THE WHOLE-BODY MODEL
 THE TISSUE MODELS (TM)
 EXPERIMENTAL DESIGN
 DISCUSSION
 REFERENCES
 
Computer models of biological systems are useful tools for understanding processes and quantifying input and output relationships. As a dynamic, quantitative representation of a system, a model can also be used to identify an appropriate experimental design. Estimates of experimental error and sample variation can be used to create data sets, which are then used with the model to determine numbers of samples, sampling times and standard deviations of estimated parameters. In this paper, previously described models of protein turnover are used to determine the best experimental design for estimating protein fractional synthesis rate (KS)3 and percentages of recycling (PR) and channeling (PC).

Fractional synthesis rate is the protein synthesis rate compared with the amount of protein. In general, KS is estimated by measuring the incorporation of a radiolabeled amino acid into protein, relative to the proportion of radiolabeled amino acid found in the precursor pool per unit of time. Three methods of estimating KS using a radiolabeled amino acid are pulse dose, continuous infusion and flooding dose. Lajtha et al. (1957)Citation determined KS by injecting a trace amount of uniformly labeled 14C lysine (pulse dose), estimating the gradient of the protein specific radioactivity curve and dividing by the difference between intracellular specific radioactivity and protein specific radioactivity at one time point. Garlick et al. (1973)Citation continuously infused uniformly labeled 14C glycine to produce a constant precursor specific radioactivity. KS was estimated from the plateau specific radioactivity of plasma glycine and 14C glycine in protein. The flooding dose method involves injecting a large dose of cold amino acid with a trace amount of a radiolabeled amino acid so that the specific radioactivity of the protein precursor pool (aminoacyl tRNA) is close to the plasma specific radioactivity. Using a flooding dose, Garlick et al. (1980)Citation calculated KS as the ratio of the 4-3H phenylalanine specific radioactivity in protein and the average specific radioactivity of phenylalanine in the precursor pool (estimated by intracellular, plasma or aminoacyl tRNA pool specific radioactivities). Each of the three methods of estimating KS is based on similar assumptions about the process of protein turnover in cells, tissues or the whole body. In this paper, dynamic, mechanistic models of protein turnover in whole body and tissues (brain, muscle and liver) were used to identify an experimental design that would allow estimation of the percentage of channeling, percentage of recycling and KS simultaneously. Channeling is the flow of amino acids from the extracellular pool of amino acids to aminoacyl tRNA for protein synthesis. Recycling is the flow of amino acids from protein degradation to aminoacyl tRNA (protein synthesis) without mixing with the intracellular pool. To determine the best method and sampling times, data were generated from the models for three rodents at each time point. Differences among rodents were defined by differences in body weights; 15% random variation representing experimental error was added to each rodent at each data point. The final data sets were used to determine how well the models could estimate protein synthesis, recycling and channeling rates by iterative Maximum Log Likelihood.


   THE WHOLE-BODY MODEL
TOP
 ABSTRACT
 INTRODUCTION
 THE WHOLE-BODY MODEL
THE TISSUE MODELS (TM)
 EXPERIMENTAL DESIGN
 DISCUSSION
 REFERENCES
 
The whole-body model (WBM) is a dynamic representation of leucine kinetics in a nongrowing rodent. The model traces the turnover of leucine from the free amino acid pools (QE, QI and QT) to oxidation (FIO) or the protein pool (QP) in Figure 1Citation . Leucine can be incorporated into protein (FTP), be released through protein degradation to be converted back into protein (FPT) or be released into the intracellular pool (FPI). In the intracellular pool, leucine can be oxidized, remain as a reserve or exchange with the extracellular pool (FEI and FIE). Unlabeled leucine intake and radiolabeled leucine are flooded, pulsed or continuously infused into the extracellular pool (FOE). Thus, amino acids for protein synthesis can arise from several sources, i.e., the intracellular pool (FIT), the extracellular pool (FET) or protein degradation (FPT). Because the rodent is nongrowing, protein is always in steady state; protein synthesis is always equal to protein degradation (Waterlow et al. 1978Citation ). Conservation of mass principles give the differential equations for the model for unlabeled and radiolabeled amino acid (Table 1Citation ).



View larger version (12K):
[in this window]
[in a new window]
  		Figure 1. The four-pool model of leucine turnover. This is the model structure for the whole-body model, the brain model, liver model and the muscle model. QE, QI and QT are leucine in the extracellular (plus plasma), intracellular and aminoacyl tRNA pools, respectively; QP is leucine in the protein pool; FEI is flux of leucine from extracellular pool to intracellular pool; FET is flux of leucine from extracellular pool to aminoacyl tRNA pool (channeling); 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; FPI is flux of leucine from protein to intracellular pool; FPO is flux of leucine from protein exported out of the tissue (liver model only); FPT is flux of leucine from protein degradation to aminoacyl tRNA pool (recycling); FTP is flux of leucine from aminoacyl tRNA pool to protein.

 

View this table:
[in this window]
[in a new window]
  		Table 1. Whole-body (WBM) and tissue model (TM) equations1

 

   THE TISSUE MODELS (TM)
TOP
 ABSTRACT
 INTRODUCTION
 THE WHOLE-BODY MODEL
 THE TISSUE MODELS (TM)
EXPERIMENTAL DESIGN
 DISCUSSION
 REFERENCES
 
The structure of the brain and muscle models is identical to the whole-body model structure. The liver model, however, includes protein export (FPO) as shown in Figure 1Citation . The primary differences between the whole-body and tissue models are the estimates of initial pool sizes and flux equations (Table 1)Citation . Initial pool sizes, protein synthesis rates and maximum percentage of recycling for whole-body (20-g mouse), brain, muscle and liver models were from previous data (Johnson et al. 1999aCitation ).

Comparison of methods to identify rates of recycling and channeling

Three experimental methods were examined, i.e., flooding dose, continuous infusion and pulse dose. For the TM, the injection specific radioactivity (111 MBq 14C Leu/30 µmol Leu for 30 min) and the experimental protocol according to Bernier and Calvert (1987)Citation were used for the flooding dose method. The injection specific radioactivity and experimental protocols according to Pomposelli et al. (1985)Citation (37 MBq 14C Leu/0.02 µmol Leu for 180 min) and Lajtha (1959)Citation (7.4 MBq 14C Leu/0.025 µmol Leu for 60 min) were used for the continuous infusion and pulse dose methods, respectively. Specific radioactivity doses for flooding dose, pulse dose and continuous infusion methods were scaled according to tissue and body volume to estimate the initial doses and rates of radiolabeled leucine available to the tissues. Due to the difficulty in separating extracellular and intracellular amino acids, specific radioactivity measurements were considered only for the aminoacyl tRNA, extracellular and protein pools. Extracellular specific radioactivity was assumed to be the same as that for plasma (Johnson et al. 1999a and 1999bCitation Citation ).

Channeling was set at 100% (100PC) or 0% (0PC). The maximum amount of recycling in the whole body, brain, liver and muscle was 12, 42, 53 and 55%, respectively, based on estimated rates of recycling by Smith and Sun 1995Citation . Recycling had to be limited to <100% of the total protein pool because at 100 PR, all of the amino acids for protein synthesis would be supplied by protein degradation and channeling could not occur. Because model predictions based on the data of Bernier and Calvert (1987)Citation and Obled et al. (1991)Citation indicated that channeling was occurring in the WBM (Johnson et al. 1999bCitation ), it was assumed the rate of recycling must be <100% of the total protein pool.


   EXPERIMENTAL DESIGN
TOP
 ABSTRACT
 INTRODUCTION
 THE WHOLE-BODY MODEL
 THE TISSUE MODELS (TM)
 EXPERIMENTAL DESIGN
DISCUSSION
 REFERENCES
 
To evaluate the model’s ability to estimate recycling and channeling using the pulse dose method in WBM and TM, the following four data sets were generated using each TM and the WBM: 1) 0% recycling of amino acids from the protein pool and 0% channeling; 2) 0% recycling and 100% channeling; 3) 0% channeling and maximum recycling; and 4) maximum channeling and maximum recycling. At each time point, body weights from 16 to 24 g were generated randomly on the basis of a normal distribution of body weights for mice at 100 d (Calvert et al. 1985Citation ) and used as inputs into the models to represent variation in initial pool sizes among mice. Three different mice were used at each time point to generate differences in specific radioactivities for each data set. In addition, -15% to + 15% numbers were generated by a random number generator on the basis of a normal distribution and were added to each data point to simulate experimental errors of measurements. Thus, variance was introduced into the data sets in two ways, i.e., specifying differences among mice (± 4g body weight) for each time point in each data set and introducing up to 15% measurement error at each mouse-time point (Bernier and Calvert 1987Citation ). Data were fit to each model using generalized reduced gradient within ACSL Optimize (Aegis Technologies 2000Citation ). ACSL Optimize uses an iterative Method of Maximum Likelihood to estimate parameters (PR, PC and KS) on the basis of data input into a model. Three measurements were considered possible, i.e., the specific radioactivity of plasma, which was used to estimate the specific radioactivity of the extracellular pool, the specific radioactivity of aminoacyl tRNA and the specific radioactivity of the protein pool. The data sets were fit to the model to determine which and how many measurements were required to predict KS, PR and PC. Time point data were systematically eliminated to determine the best combination of sampling times for each method and model. The best time points were 2, 6, 10, 40, 70 and 100 min for the continuous infusion and pulse dose methods and 2, 6, 8, 10, 40 and 70 min for the flooding dose method.

Brain model estimates of PR, PC, KS.

In Table 2Citation , observed (Obs) model settings for KS, PR and PC are compared with those predicted (Pred) from fitting data generated by the brain model for each method. Estimates of PR, PC and KS were very close to observed data for each method. Most standard deviations were within 15% of the error of generated data. However, standard deviations of predictions of PC using flooding dose method were very high. Therefore continuous infusion and pulse dose methods were best.


View this table:
[in this window]
[in a new window]
  		Table 2. Brain model predicted rates of protein synthesis, recycling and channeling from fitting model-generated specific radioactivity data with experimental variation using each method (continuous infusion, flooding dose and pulse dose methods) to the brain rodent model

 
Liver model estimates of PR, PC, PE and KS.

Data from the liver model were difficult to fit partly because the percentage of protein synthesized that is exported from the tissue (PE) must also be estimated. PE was estimated because protein export is a large proportion of liver protein synthesis and will affect specific radioactivities of extracellular, intracellular, leucyl tRNA and protein pools. PE may also vary among animals and different physiologic states. Table 3Citation lists results from a comparison of observed to predicted rates. Estimates of PR, PC, PE and KS were very close to those observed (set) in the model except for PR in the flooding dose method at high recycling and high channeling. Standard deviations of predictions of PC were very high at high rates of channeling for all methods and at high recycling and low channeling for the flooding dose method. Standard deviations were lowest for the pulse dose method.


View this table:
[in this window]
[in a new window]
  		Table 3. Liver model predicted rates of protein synthesis, recycling and channeling from fitting model-generated specific radioactivity data with experimental variation using each method (continuous infusion, flooding dose and pulse dose methods) to the liver rodent model1

 
Muscle model estimates of PR, PC and KS.

Predictions of PR, PC and KS were within 2–3% of observed values except PC at high PC and PR, which was 85% PC in the pulse dose method. Standard deviations of predictions for PC were very high for the continuous infusion and flooding dose methods. Pulse dose estimates and standard deviations were much lower (Table 4Citation ).


View this table:
[in this window]
[in a new window]
  		Table 4. Muscle model predicted rates of protein synthesis, recycling and channeling from fitting model-generated specific radioactivity data with experimental variation using each method (continuous infusion, flooding dose and pulse dose methods) to the muscle rodent model1

 
WBM estimates of PR, PC, KS.

Observed, predicted and standard deviations of predictions for PR, PC and KS are given in Table 5Citation . Estimates of PR, PC and KS were very close to observed values for all methods. Standard deviations of predictions were also <16% except for flooding dose PC at high PC and PR (28.1%).


View this table:
[in this window]
[in a new window]
  		Table 5. Whole-body model predicted rates of protein synthesis, recycling and channeling from fitting model-generated specific radioactivity data with experimental variation using each method (continuous infusion, flooding dose and pulse dose methods) to the whole-body rodent model1

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 THE WHOLE-BODY MODEL
 THE TISSUE MODELS (TM)
 EXPERIMENTAL DESIGN
 DISCUSSION
REFERENCES
 
Dynamic models that quantify biological processes can be useful tools with which to analyze different experimental methods. A model is only a representation of a physiologic process and is limited by our knowledge of the process. Therefore, it is important to first test a model with data and perform sensitivity analyses to identify potential weaknesses in model predictions before designing an experiment. Then, once strengths and potential weaknesses have been identified, the model can be used to predict an experimental design.

The model’s structure was evaluated in Johnson et al. (1999a and 1999b)Citation Citation . The changes in specific radioactivities were well reproduced by the model, which was sensitive to PR, PC and KS and insensitive to changes in pool sizes. Potential weaknesses were lack of data for rates of exchange of leucine between the extracellular and intracellular pools. Therefore in the WBM and TM, it was assumed that FIE was due primarily to diffusion and FEI was due primarily to leucine transport by the L system. Diffusion constants and Km and Vmax were set accordingly (Miller et al. 1985Citation , Oxender and Christensen 1963Citation , Stevens et al. 1984Citation ). FEI and FIE were dependent on the concentration of leucine in the extracellular and intracellular pools. Previous solutions for the WBM implied that FEI and FIE were due to mass action and were constant. However, the solutions were based on short-term (30 min) flooding dose data, which may not reflect physiologic levels of leucine.

Previously, the pulse dose method showed the most promise for predicting PR, PC and KS in the WBM (Johnson et al. 1999bCitation ). Therefore when variation was added to the specific radioactivity data and body weights, it was not surprising that the pulse dose method best distinguished between different rates of PR, PC and KS. Estimates of PR, PC, KS and PE (liver only) using the pulse dose method were closest to observed (set) values used to generate data (Tables 2Citation 3Citation 4Citation 5)Citation . In addition, predicted specific radioactivities were close to generated values (within 15%). The exception was liver, which had high standard deviations on estimates of PC. Due to high PE and KS by the liver, it may be difficult to estimate PC using the model. For estimates of KS, standard deviations of predictions were much less than the percentage of error that was added randomly to each data point (experimental error and error due to differences in rodent body weight). Thus, estimating KS by resolving time-course data using a model rather than by traditional calculations can decrease the final errors of prediction and lead to more accurate estimates.

Accurate estimates of KS are dependent on 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. Because the amount of recycling and channeling can vary among tissues and with the amino acid used as a tracer, it is imperative that limitations associated with each of the methods be known for individual tissues and whole-body estimates. According to the model, PR, PC and KS can be determined simultaneously using a pulse dose with measurements of the specific radioactivities of the extracellular, leucyl tRNA and protein pools at 2, 6, 10, 40, 70 and 100 min with three mice per time point.


   FOOTNOTES
 
2 To whom reprint requests should be addressed. Back

3 Abbreviations used: FEI, flux of leucine from extracellular pool to intracellular pool (µmol/min); FET, flux of leucine from extracellular pool to aminoacyl tRNA pool (channeling) (µmol/min); FPT, flux of leucine from protein degradation to aminoacyl tRNA pool (recycling) (µmol/min); FIE, flux of leucine from intracellular pool to extracellular pool (µmol/min); FIO, flux of leucine oxidized from intracellular pool (µmol/min); FIT, flux of leucine from intracellular pool to aminoacyl tRNA pool (µmol/min); FOE, intake flux of leucine to extracellular pool (µmol/min); FPI, the flux of leucine from protein to the intracellular pool (µmol/min); FPO, flux of leucine from protein exported out of the tissue (liver model only) (µmol/min); FPT, the flux of leucine from protein to aminoacyl tRNA and represents recycling (µmol/min); FTP, flux of leucine from aminoacyl tRNA pool to protein pool (µmol/min); KS, protein synthesis rate (%/d); PC, the percentage of channeling; PE, the percentage of protein synthesized that is exported from the tissue (liver); PR, the percentage of recycling; QE, leucine in extracellular pool (µmol); QI, leucine in intracellular pool (µmol); QP, leucine in protein pool (µmol); QT, leucine in aminoacyl tRNA pool (µmol); TM, the tissue models for protein turnover in rodent including a brain model, liver model and muscle model; WBM, whole-body protein turnover model for a rodent. Back

Manuscript received March 20, 2000. Initial review completed May 25, 2000. Revision accepted September 13, 2000.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 THE WHOLE-BODY MODEL
 THE TISSUE MODELS (TM)
 EXPERIMENTAL DESIGN
 DISCUSSION
 REFERENCES
 

1. Aegis Technologies Optimize, ACSL: Advanced Continuous Simulation Language 2000 Aegis Technologies Huntsville, AL.

2. 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

3. Calvert C. C., Famula T. R., Bernier J. F., Bradford G. E. Serial composition during growth in mice with a major gene for rapid postweaning growth. Growth 1985;49:246-257[Medline]

4. Garlick P. J., McNurlan M. A., Preedy V. R. A rapid and convenient technique for measuring the rate of protein synthesis in tissues by injection of [3H] phenylalanine. Biochem. J. 1980;192:719-723[Medline]

5. Garlick P. J., Millward D. J., James W.P.T. The diurnal response of muscle and liver protein synthesis in vivo in meal-fed rats. Biochem. J. 1973;136:935-945[Medline]

6. John A. M., Bell J. M. Amino acid requirements of the growing mouse. J. Nutr. 1976;106:1361-1367

7. Johnson H. A., Baldwin R. L., France J., Calvert C. C. Development and evaluation of a model of whole body protein turnover based on leucine kinetics in rodents. J. Nutr. 1999a;129:728-739[Abstract/Free Full Text]

8. Johnson H. A., Baldwin R. L., France J., Calvert C. C. Recycling, channeling and heterogeneous protein turnover based on leucine kinetics in rodents. J. Nutr. 1999b;129:740-750[Abstract/Free Full Text]

9. Lajtha A. Amino acid and protein metabolism of the brain-V. Turnover of leucine in mouse tissues. J. Neurochem. 1959;3:358-365[Medline]

10. Lajtha A., Furst S., Gerstein A., Waelsch H. Amino acid and protein metabolism of the brain-I. J. Neurochem. 1957;1:289-300[Medline]

11. Miller L. P., Pardridge W. M., Braun L. D., Oldendorf W. H. Kinetic constants for blood-brain barrier amino acid transport in conscious rats. J. Neurochem. 1985;45:1427-1432[Medline]

12. 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

13. Oxender D. L., Christensen H. N. Distinct mediating systems for the transport of neutral amino acids by the Erlich cell. J. Biol. Chem. 1963;238:3686-3699[Free Full Text]

14. 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]

15. Smith C. B., Sun Y. Influence of valine flooding on channeling of valine into tissue pools and on protein synthesis. Am. J. Physiol. 1995;268:E735-E744[Abstract/Free Full Text]

16. Stevens B. R., Kaunitz J. D., Wright E. M. Intestinal transport of amino acids and sugars: advances using membrane vesicles. Annu. Rev. Physiol. 1984;46:417-433[Medline]

17. Waterlow J. C., Garlick P. J., Millward D. J. Protein Turnover in Mammalian Tissues and in the Whole Body 1978 North-Holland Amsterdam, The Netherlands.




This article has been cited by other articles:


Home page
 Am. J. Physiol. Regul. Integr. Comp. Physiol.Home page
M. R. Deschenes, J. A. Giles, R. W. McCoy, J. S. Volek, A. L. Gomez, and W. J. Kraemer
Neural factors account for strength decrements observed after short-term muscle unloading
Am J Physiol Regulatory Integrative Comp Physiol, February 1, 2002; 282(2): R578 - R583.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Johnson, H. A.
Right arrow Articles by Calvert, C. C.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Johnson, H. A.
Right arrow Articles by Calvert, C. C.

Home Help [Feedback] [For Subscribers] [Archive] [Search] [Contents]
Copyright © 2000 by American Society for Nutrition