The Journal of Nutrition Vol. 128 No. 9 September 1998, pp. 1562-1569

A New Stable Isotope Method Enables the Simultaneous Measurement of Nucleic Acid and Protein Synthesis In Vivo in Mice^1,2

Jose F. Perez³ and Peter J. Reeds⁴

USDA/ARS Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX 77030

	ABSTRACT

Abstract Introduction Methods Results Discussion References

We developed a method based on the incorporation of ¹³C₂-units derived from [U-¹³C]glycine that allows the simultaneous quantification of tissue protein and RNA synthesis in vivo. Two groups of 26 mice were fed diets containing a high (HF diet) or a low quantity of fiber (LF diet). After 6 d, [U¹³C]glycine was added to the diet and groups of four mice were killed after 2, 4, 6, 8, 12 and 24 h. Hepatic and intestinal mucosal free and RNA-bound purine nucleosides were extracted and enzymically degraded to allantoin. Allantoin was degraded to glyoxylate, which was then reductively aminated to glycine, which contains the two ¹³C-atoms incorporated via de novo synthesis. Ingestion of the HF diet was associated with significantly (P < 0.05) higher rates of total RNA synthesis in both the liver ( HF diet, 29%/d; LF diet, 21%/d) and mucosa (HF diet, 27%/d; LF diet, 17 %/d). The mean rates of RNA synthesis in each tissue were significantly (P < 0.01) lower than the respective rates of protein synthesis (liver, 67%/d; mucosa, 74%/d). The isotopic enrichment of the free purine nucleotide pool increased rapidly and exponentially, but the steady-state value was substantially (P < 0.001) lower than that of the RNA-bound purines. The results suggest that the free nucleotide pool consists of two kinetically distinct compartments, one of which is small and has a rapid rate of turnover. This, we propose, acts as the RNA precursor pool. The other is large, has a low rate of turnover and, we believe, is the pool of adenosine triphosphate involved in cellular energetics.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Although there is a large body of mechanistic information on the regulation of protein and RNA synthesis in cells, very little quantitative information is available on any aspect of nucleic acid turnover in vivo. Such information would be invaluable in studies of the turnover of ribosomal RNA, as it relates to the control of cell and tissue protein synthetic capacity (Eichler and Craig 1994, Morley et al. 1994, Reeds et al. 1993) and the regulation of mRNA formation and turnover (Shikama and Brack, 1996).

The lack of information is largely a reflection of methodological difficulties in making the basic measurement, including the choice of the most appropriate tracer and the kinetic complications that arise from compartmentation within the free nucleotide pools (Schroder and Rapaport 1984). Nucleotides incorporated into nucleic acids can derive either from de novo synthesis or from the salvage of bases and ribose released by the breakdown of cellular nucleic acids. In addition, within the splanchnic tissues, and especially in the intestinal mucosa, there is the additional possibility of incorporation of preformed nucleosides derived either from the diet or from microbes resident in the lumen. Evidence for the simultaneous operation of all three pathways can be found in the literature (e.g., Berthold et al. 1995, Leleiko et al. 1979 and 1983, Savaiano et al. 1980).

The relative activities of the de novo and salvage pathways in different tissues have been estimated by assessing the activities of the enzymes involved in these pathways in vitro (Allsop and Watts 1980). However, although such data demonstrate the potential for base salvage, there is no certainty that measurements of enzyme activity under saturating substrate conditions in vitro necessarily reflect the rates at which metabolites flow along the individual pathways in vivo. Recently, we (Berthold et al. 1995, Boza et al. 1996) used mass isotopomer distribution analysis of ribonucleosides isolated from mice that had consumed diets containing uniformly ¹³C-labeled amino acids or nucleotides to study the relative contribution of the de novo and salvage pathways to RNA synthesis in mucosa and liver. From these studies, we concluded that >90% of the purine bases incorporated into the nucleic acids had arisen from de novo synthesis. These results suggested that the use of carbon-labeled amino acid precursors should provide particularly effective tracers for the measurement of nucleic acid synthesis in vivo. This paper reports the development of such an approach and its application to the determination of RNA synthesis in the intestinal mucosa and liver of mice. In an attempt to examine changes in RNA turnover associated with increased mucosal cell turnover (Ecknauer et al. 1981, Vahouny and Cassidy 1984), we quantified this process in groups of animals that ingested diets that contained different amounts of cellulose.

	MATERIAL AND METHODS

Abstract Introduction Methods Results Discussion References

Principle of the method.  The synthesis of the purine ring involves the initial synthesis of inosine monophosphate in which each carbon and nitrogen atom derives from specific precursors (Fig. 1). Three of the nitrogen atoms derive from aspartate (N1) and the amide group of glutamine (N3, N9); C2 and C8 derive from single carbon units incorporated via N₁₀-methyl-tetrahydrofolate and N₅,N₁₀-methenyl-tetrahydrofolate, respectively. C6 is derived from bicarbonate. From the perspective of the method that we have developed, it is critical that C4, C5 and N7 derive from a single molecule of glycine. That being so, we reasoned that if a method for the isolation and isotopic characterization of the moiety C4-C5-N7 could be devised, then [¹³C₂]glycine, which would label both carbon atoms 4 and 5 as a single [¹³C₂]unit, would be a particularly useful way of approaching the problem of quantifying the synthesis of RNA in vivo.

View larger version (13K): [in this window] [in a new window]   Fig 1. Biosynthetic origin of the atoms of the purine ring. The circled atoms are those that are derived from a single glycine molecule.

The method is therefore based on the use of [U-¹³C]glycine as tracer followed by the isolation and targeted degradation of RNA-bound purines to yield carbons 4 and 5 of the purine molecule as glyoxylic acid. Because glyoxylate is unstable, the final stage of the method is its conversion by reductive amination (Robins and Reeds 1984) to glycine. The labeled glycine, so synthesized, is then used for the mass spectrometric determinations. A key feature of the method therefore is the use of selected ion monitoring mass spectrometry to measure the unique ¹³C₂-signal in two specific carbons of the purine ring, whose labeling kinetics can be used to calculate the rate of RNA synthesis.

Experimental.  The experiment received prior approval from the Animal Protocol Review Committee of Baylor College of Medicine. All animal housing and husbandry conformed to USDA guidelines. [U-¹³C]Glycine was purchased from Cambridge Isotopes (Woburn, MA). As determined by HPLC, it was >99% chemically pure; 93% of the molecules contained two¹³C atoms and 6.25% contained one.

Animals and diets.  Fifty-two female adult mice of the ICR strain (mean weight. 22 g) were purchased from Harlan Sprague Dawley (Indianapolis, IN). They were housed individually on sterile bedding in rooms maintained at 22°C with a 12-h light:dark cycle (lights off at 1800 h). On arrival, they were randomly divided into two groups of equal mean weight and were then offered either a high (HF) or a low (LF) fiber diet (Table 1). The powdered diets were presented at a rate of 5 g/d in containers that prevented spillage.

For the first 6 d of the experiment, the diets contained unlabeled free glycine (4.3 g/kg diet) and were offered in two equal meals at 0600 and 1800 h. The preliminary feeding period, which served to accustom the animals to the feeding regimen, also allowed us to ascertain that the animals consumed all of the feed that was offered. On d 7, unlabeled glycine was completely replaced by the same amount of [U-¹³C]glycine, and 1.65 g of the labeled diet was offered to all of the animals at 0600 h. Four mice from each dietary group were killed at 2, 4, 6 and 8 h after first exposure to the labeled diet. The remaining mice were offered a second meal (1.65 g) of labeled diet at 1400 h. Two groups of four mice from each diet group were then killed at 1800 h, i.e., 4 h after their second meal. These mice had therefore been exposed to [U-¹³C]glycine for 12 h. A third meal of labeled diet was then offered to the remaining mice at 2200 h. These mice were killed at 0600 the next day, after having been exposed to [U-¹³C]glycine for 24 h. Two additional mice in each group that had not received [U-¹³C]glycine were used to determine the baseline tracer:tracee ratios.

Sampling.  Each animal was deeply anesthetized by inhalation of isofluorane, the thoracic cavity was opened, and the whole liver and the small intestine (from the pylorus to the ileal cecal junction) were removed and immediately frozen in liquid nitrogen. The mice were then exsanguinated by cardiac puncture. All samples were stored for up to 1 mo at 70°C until processed. When taken for analysis, the intestine was thawed to 4°C; the luminal contents were removed under gravity and by gently flushing the lumen with an additional 2 mL of ice-cold NaCl solution. This freeze-thaw cycle disrupts the mucosal integrity so that the mucosa can be isolated merely by applying gentle pressure to the intestine. The sloughed mucosa were frozen at 70°C until analyzed.

Isolation of free amino acids and nucleotides.  Samples of liver and mucosa (100 mg) were homogenized in 1 mL perchloric acid (0.2 mol/L; 4°C) and centrifuged (3000 × g for 10 min at 4°C). The supernatant was neutralized with a minimum volume of KOH (4 mol/L) and the potassium perchlorate separated by centrifugation (3000 × g for 15 min at 4°C). The supernatant was taken for analysis of free glycine and purine nucleotides. The perchloric acid-insoluble precipitate was washed with two changes of perchloric acid (0.2 mol/L) and hydrolyzed in sealed tubes with 1 mL of hydrochloric acid (6 mol/L) at 110°C for 24 h. The hydrolysate was then diluted with 2 mL of deionized water. Samples of the diluted hydrolysate and the original neutralized perchloric acid supernatant were applied to a 2- mL bed volume column of Dowex AG-50 WX8 (Biorad, Richmond, CA) (H⁺ form) cation exchange resin. The column was washed with eight bed volumes of water to elute the free nucleotides, and the amino acids were released from the column by elution with 2 mL of NH₄OH (5 mol/L). The amino acid and nucleotide fractions were then dried under vacuum.

Isolation and enzymatic digestion of RNA to free nucleosides.  All enzymes were purchased from Sigma Chemical (St Louis, MO). High-molecular-weight RNA was isolated as described previously (Berthold et al. 1995). RNA pellets were dissolved in 100 µL of autoclaved diethyl pyrocarbonate-treated water and digested to nucleosides as described by Crain (1990) except that the amounts of nuclease P1, phosphodiesterase and alkaline phosphatase and the times of incubation were increased fourfold over those recommended by Crain. Free nucleotides obtained from cation chromatography were dissolved in 100 µL of deionized water and incubated for 3 h at 37°C with 2 U of alkaline phosphatase.

Nucleoside degradation to allantoin.  Purine nucleosides (adenosine and guanosine) were degraded to allantoin by a series of enzymatic reactions. The enzymes used were as follows: (a) adenosine deaminase (EC 3.5.4.4; 150-200 U/mg, from calf spleen); (b) purine nucleoside phosphorylase (EC 2.4.2.1; 20 U/mg, from calf spleen); (c) guanase (EC 3.5.4.3; 0.06-0.20 U/mg, from rabbit liver); (d) xanthine oxidase (EC 1.2.3.2; 1-2 U/mg, from buttermilk); (e) uricase (EC 1.7.3.3; 4-8 U/mg; from porcine liver).

The enzyme reactions were as follows:

The dried nucleosides (either from the acid-soluble or RNA-bound pool) were dissolved in 150 µL water and mixed with 100 µL potassium phosphate buffer (0.2 mol/L, pH 7.5). The reactions were carried out in three steps. Purine nucleoside phosphorylase (1.5 U), adenosine deaminase (480 mU) and guanase (5mU) were added and the solution was incubated for 8 h at 37°C. After this period, xanthine oxidase (48 mU) and 30 µL of an aqueous solution of L-histidine (4.3 mmol/L) were added, and the incubation continued for 2 h at 37°C. The histidine was added to minimize substrate inhibition of xanthine oxidase activity. At the end of this incubation, the pH of the solution was increased to 9-9.5 with KOH. Uricase (26 mU) was added and the incubation continued at 37°C for 2 h. The pH of the final solution was adjusted to pH 7, and the enzymes were precipitated by adding 1 mL methanol followed by incubation at 0°C for 30 min. After centrifugation (3000 × g for 10 min at 4°C), the supernatant was dried under reduced pressure in a centrifugal lyophilizer.

The dried allantoin was redissolved in 1 mL of acetic acid (0.5 mol/L) and applied to a 2-mL bed volume column of Dowex AG-50 WX8 (H⁺ form) cation exchange resin. At this pH, amino acids are completely retained and allantoin elutes in the combined sample front and a single (bed volume) water elution. The molar recovery (see Results) of an initial amount of purine nucleosides as allantoin was checked by quantification according to the method of Young and Conway (1942) against allantoin standards.

Conversion of allantoin into glycine.  Allantoin was degraded to glyoxylate with a two-step procedure modified from Young and Conway (1942). The dried allantoin was dissolved in 70 µL of KOH (1 mol/L); the tube was sealed and incubated for 7 min at 100°C. After chilling, the pH was reduced to 1-2 by adding 15 µL HCl (6 mol/L) and incubated for 2 min at 100°C. Samples were neutralized with KOH to reach a final volume of 100 µL. Glyoxylate was converted to glycine by reductive amination (Robins and Reeds 1984). A total of 45 µL of ammonium formate (8.5 mol/L) was added and the pH adjusted to 6.5. An aqueous solution (20 µL) of sodium cyanoborohydride (0.16 mol/L) was then added. The solution was heated in sealed tubes at 105°C for 5 h. Finally, the samples were dried and stored until final derivatization for gas chromatographic mass spectrometric analysis. Molar recovery of glycine (see Results), either from allantoin, from pure purine nucleosides or from standard RNA was measured by reversed-phase HPLC of glycine phenylisothiocyanate.

Isotopic analysis.  Purified dried amino acids, either from the free and bound protein pool or isolated from a particular nucleotide pool, were converted to the N-propyl ester, N-heptafluorobutyramide derivatives (Jahoor et al. 1994). Mass spectrometry was performed on a Hewlett-Packard 9890 gas chromatograph quadropole mass spectrometer (Hewlett-Packard, Palo Alto, CA). We used methane negative chemical ionization for the determination of glycine (mass/charge 293-295).

Calculations.  The crude ion yields were converted into excess tracer:tracee ratios by using the matrix approach of Brauman (1966) as used in previous papers (Berthold et al. 1995). The baseline ion spectrum used in the calculation was that of glycine isolated from mice that had not received labeled amino acids. The fractional synthesis rate of hepatic and mucosal protein was calculated on the assumption that the incorporation kinetics were monoexponential, and hence

(1)

in which k is the fractional rate of synthesis (per day), t is time (d) and A is the asymptotic value for the tracer:tracee ratio of the free glycine incorporated into protein. The curve fitting was carried out by using the numerical routine in SAAM-II (SAAM Institute, University of Washington, Seattle, WA). Data from all time points were used in the curve fitting, and it is important to note that the SAAM-II program uses the standard deviations of the data to weight the nonlinear regression analysis. The standard deviations of the final values of k and A therefore include both between-animal and analytical variation.

In a previous paper (Boza et al. 1996), we showed that after a prolonged period of labeling (5 d) with [U¹³C]amino acids, hepatic and mucosal protein glycine and the [M+2]isotopomer of RNA-purines approached a similar asymptotic value, suggesting that the same pool of glycine is used as precursor for both protein and purine synthesis. Therefore, in the equation used to calculate the fractional rate of RNA synthesis, the model-derived value of A was used as the denominator. To calculate RNA synthetic rate, we used a simplified equation that assumed quasilinear tracer incorporation over a 24-h period of labeling. Thus, the fractional rate of RNA synthesis (k) was calculated from the equation

(2)

in which A is the model-derived value for the steady-state isotopic enrichment as calculated with Equation 1 and B the baseline enrichment.

Data presentation and statistics.  Data are presented as mean values and between-animal SD. Comparison between tissues (liver vs. muscle) within a mouse was performed with a paired t test and between diets within a tissue by a grouped t test. A two-tailed P value of 0.05 was considered significant.

	RESULTS

Abstract Introduction Methods Results Discussion References

Analytical approach.  The efficiency of the method to yield glycine, and likely pitfalls during the process, were checked in initial studies to optimize the conditions of the reactions. The mean molar recovery of 1:1 mixtures of purified adenosine and guanosine as allantoin was 65 ± 5% (n = 5), which, given the number of enzymes involved, is compatible with an average efficiency in each step of ~92%. Allantoin conversion into glycine was essentially complete (97.6 ± 0.7%; n = 8). The yield of glycine, following the degradation of RNA standards, was 470 ± 40 pmol glycine/mg RNA (n = 7). Assuming a purine base content of 1.025 µmol purine/mg RNA, this yield implied an overall efficiency of ~48 ± 4%. Glycine derivatization to the N-propyl ester, N-heptafluorobutyramide derivative was not significantly affected by the presence in the vial of substantial amounts of ammonium formate (27 mg). However, preliminary results showed that potassium phosphate, carried over from the enzyme reaction steps, was a potential inhibitor, and we found that the final quantity of potassium phosphate used during the enzymatic degradations should be kept below 20 nmol per incubation.

The final sensitivity of the method is vulnerable to the introduction of glycine from the various reagents used. In initial experiments, we found that autolysis of the enzymes was capable of releasing 8 nmol glycine per incubation; it was for that reason that we introduced the cationic isolation of allantoin before its conversion to glyoxylate. Under the conditions of the assay, the ammonium formate used for the reductive amination contained 200 pmol glycine per incubation.

Isotopic data.  The tracer:tracee ratios (mol/100 mol) of free and protein-bound [U-¹³C]glycine in liver and mucosa are presented in Table 2. Tissue glycine labeling was rapid, but there was variation in the tracer:tracee ratio of glycine at different time points. The lowest enrichments were at 8 and 24 h at which times the glycine isotopic enrichment was between 50 and 60% of the mean value for the whole period of labeling. This variation, which increases the variance of the final estimates of protein and RNA labeling kinetics, presumably related to the feed intake pattern of the animals. Over the whole of the labeling period, the average tracer:tracee ratio (mol/100 mol) of free hepatic glycine was 23 ±7 for the HF group and 22 ± 5 for the LF group. Corresponding values for the mucosa were 31 ± 9 (HF) and 35 ± 8 (LF). The isotopic enrichment of [M+2]glycine in the bound protein pool of liver and mucosa increased throughout the 24-h period.

Fitting the protein isotopic data to a monoexponential model that included all time points predicted that the steady-state tracer:tracee (mol/100 mol) of the [U-¹³C]glycine incorporated into hepatic protein was 11.4 ± 3.8 for the HF group and 11.3 ± 1.6 for the LF group. These values were between 48 and 51% of the mean measured tracer:tracee ratios of hepatic free glycine. Corresponding estimates of the isotopic enrichment of the mucosal protein synthetic precursor pool were 21.3 ± 2 mol/100 mol for the HF group and 20.3 ± 2 mol/100 mol for the LF group, 68 and 58% of the respective tracer:tracee ratios of mucosal free glycine. The model-calculated rates of hepatic protein synthesis were 67.7 ± 15.4%/d (HF) and 67.9 ± 24.0%/d (LF), and were unaffected by the previous diet. The model estimates of mucosal protein synthetic rates were 66.7 ± 3.1%/d and 81.4 ± 14.4%/d for the LF and HF groups, respectively. The effect of diet was significant (P < 0.025).

The time course of labeling of [M+2]glycine synthesized from the free and RNA-bound purines is shown in Figure 2 (mucosal) and Figure 3 (hepatic). The labeling of the purine nucleotides extracted from RNA increased throughout the 24-h period. In the mucosa, there was a distinct lag phase before quasilinear incorporation was established, whereas in the liver, the rise was quasilinear for the whole labeling period. In both tissues, the free purine pool was labeled rapidly, but only to a low isotopic enrichment (~1.5 mol/100 mol in both tissues). Thus, from 8 h of labeling, the isotopic enrichment of the mixed free purine pool was lower than that of RNA-bound pool and clearly could not have been the immediate precursor of the purines incorporated into the RNA.

View larger version (19K): [in this window] [in a new window]   Fig 2. Labeling of the [M+2]glycine extracted from the RNA bound and free purines of the small intestine mucosa obtained from groups of adult mice that ingested a diet (Table 1) containing [U-13C]glycine (4.3 g/kg diet) and either a low (50 g/kg diet) or a high (150 g/kg diet) quantity of cellulose. Values are expressed as mean ratios of the free glycine tracer:tracee ratios ± SD, n = 4 per time point.

View larger version (20K): [in this window] [in a new window]   Fig 3. Labeling of the [M+2]glycine extracted from the RNA bound and free purines in the livers obtained from groups of adult mice that ingested a diet (Table 1) containing [U-13C]glycine (4.3 g/kg diet) and either a low (50 g/kg diet) or a high (150 g/kg diet) quantity of cellulose. Values are expressed as mean ratios of the free glycine tracer:tracee ratios ± SD, n = 4 per time point.

By using the model-derived values for the average tracer:tracee of [U-¹³C]glycine in the protein synthetic precursor pool (A in Equation 1) as the basis for the calculation, the fractional rate of synthesis of total mucosal RNA (mean ± SD) was 26.9 ± 2.1%/d in the HF group and 16.6 ± 4.8%/d in the LF group. These values were significantly (P < 0.01) different. In liver, the rates of total RNA synthesis (29.3 ± 4.29, HF and 21.1 ± 2.44%/d, LF) also differed with diet (P < 0.01). The fractional rates of RNA synthesis in each tissue were significantly (P < 0.001) lower than the fractional rates of tissue protein synthesis.

To investigate further whether the pool of glycine used for the synthesis of RNA-bound purines and protein was the same, we examined the ratios of the isotopic enrichments of the [M+2] and [M+1]glycine in the various isolates (Table 3). The latter isotopomer is derived from glycine that has been synthesized de novo by the mice. The [M+1]glycine/[M+2]glycine ratios of the free, bound protein and extracted RNA were significantly higher in liver than in the small intestinal mucosa (P < 0.001). Within a tissue, no significant differences were observed between the free and bound protein isotopomer ratios, but glycine extracted from the RNA showed a slightly higher ratio than that of the bound protein glycine. The difference (22%) was significant (P < 0.05) in the liver. Strikingly, the ratio of [M+1]glycine to [M+2]glycine in the free purines (1.06 ± 0.5 in liver and 0.35 ± 0.28 in mucosa) was higher (P < 0.05) and much more variable than in any of the other sampled pools. This result is another demonstration of the lack of a simple precursor product relationship between free and RNA-bound purines.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

Methodological considerations.  An important question regarding the general applicability of the method is its overall sensitivity, both in relation to isotopic enrichment and to the quantities of starting RNA that are needed for accurate mass spectrometry. Regarding the former, the use of the [M+2]isotopomer of a small amino acid requires that the purine C4-C5 unit be labeled to ~0.2 mol/100 mol (4 SD of the between-sample variation at baseline enrichments) when gas chromatography mass spectrometry is the method of quantification, although isotopic sensitivity could be increased substantially by the use of gas chromatography combustion isotope ratio mass spectrometry.

Regarding the minimum quantity of RNA that can be analyzed, the overall yield of the C4-C5 units was 46%, and negative chemical ionization gas chromatography mass spectrometry can readily quantify 500 pmol of glycine. Given that the average concentration of purines in RNA is ~1 nmol/µg RNA, it should be possible, other factors being equal, to measure the labeling of purines extracted from 2 µg of RNA. This would be the equivalent of 4, 1 and 0.6 mg of mucosa, muscle and liver, respectively. At this level of sensitivity, it should also be possible to measure the rate of synthesis of some of the more abundant mRNAs.

However, the sensitivity of the method is limited by potential contamination with unlabeled glycine that has entered from other sources. In our hands, this amounted to 100-200 pmol per preparation. Thus, the limit for the accurate quantification of isotopic enrichment is ~20 µg of RNA.

Precursors for RNA synthesis.  The estimation of the rate of synthesis of any macromolecule should in theory proceed from the identification of the isotopic enrichment of the true precursor. This problem has received considerable attention in the literature concerned with the measurement of protein synthesis, but the use of nutritionally essential amino acids as tracers for protein synthesis has resolved the problem to one of measuring the isotopic dilution of the precursor pool rather than identifying the source. However, because nucleotides are undoubtedly synthesized by most, if not all cells, nucleic acids can derive from at least three sources of nucleotides: 1) preformed nucleotides or bases transported from the extracellular space, and hence, possibly from the diet; 2) nucleotides and nucleosides recycled from cellular nucleic acid degradation; and 3) bases and ribose synthesized de novo within the cell. There is qualitative evidence that in vivo RNA labeling can be achieved from labeled nucleotides given either systematically (Savaiano et al. 1980, Simmons et al. 1973) or enterally (Berthold et al. 1995, Ho et al. 1979, Sonoda and Tatibana 1978) and from intravenous or enteral (Boza et al. 1996) amino acid precursors of nucleotide base synthesis.

Most previous efforts to measure nucleic acid synthesis and turnover have relied on radioactive tracers (Chikenjii et al. 1988, Edmonds and LePage 1955) with the subsequent measurement of tracer atom incorporation. Basing conclusions on the incorporation of isotopic atoms (e.g., ¹⁴C or ³H) rather than true precursor molecules adds a further difficulty of interpretation because isotopic atoms can be incorporated via multiple pathways. It is possible, for example, for ¹³C or ¹⁴C in a labeled nucleoside to have become incorporated into nucleic acid after the degradation of the purine nucleotide tracer, rather than by direct incorporation of the tracer molecules themselves. This is particularly likely because three of the five carbons in the purine ring derive independently from single-carbon atoms.

Selected ion monitoring mass spectrometric analysis allows the quantification of specific isotopologues (i.e., molecules bearing 1,2...X ¹³C atoms), thereby avoiding the pitfalls of inferences drawn from atomic incorporation. In principle, the method could be applied to radio tracers although in practice, the molar enrichment of radio tracers is orders of magnitude below the detection ability of selected ion monitoring mass spectrometric analysis. ¹³C-Labeled tracers, however, offer the possibility of measuring the incorporation of distinct isotopic species and thus the potential for choosing tracers that can become incorporated only via specific pathways. For example, by following the incorporation of intact [U¹³C]-labeled purine and pyrimidine nucleosides during the oral administration of [U¹³C]nucleic acids (Berthold et al. 1995) or nucleosides (Boza et al. 1996), our group has shown that virtually no dietary purines escape degradation in the murine enterocyte and that the majority of the ¹³C-incorporation into RNA from labeled nucleotides represents the recycling of labeled 1-carbon units derived from the degradation of the labeled nucleoside tracers.

The method that we report is a further refinement of the mass isotopomer approach. The basis of the method comprised the two following previous observations: 1) in the mouse, the large majority of the hepatic and mucosal RNA purines derive from de novo synthesis (Berthold et al. 1995, Boza et al. 1996) and 2) protein-bound glycine and the glycine incorporated into RNA via purine synthesis reach near isotopic equilibrium (Boza et al. 1996). The first observation suggested to us that the use of [U-¹³C]glycine, together with the isolation of the specific purine fragment derived from glycine, would avoid problems associated with label recycling, which, as shown in Table 3, do occur. The second observation suggested that the true fractional rate of RNA synthesis could be calculated by using the isotopic enrichment of the protein synthetic pool of glycine to define the purine precursor pool isotopic enrichment.

The method, of course, relies on the ability of given tissues to synthesize purines de novo. Previous authors have presented evidence that the intestinal mucosa either do not carry out purine synthesis (Mackinnon and Deller 1973) or do so only to a limited extent (Leleiko et al. 1983). It has also been suggested that purines in the mucosa derive from nucleosides exported by the liver (Konishi and Ichihara 1979). In this work, we used the relative isotopic enrichments of the two glycine ¹³C- isotopologues ([M+1]glycine and [M+2]glycine) to examine this possibility, our reasoning being that identity of the ratio of [M+1]glycine to [M+2]glycine in the two tissues would indicate a common source of purine nucleotides. The results showed that not only were the ratios different in the liver and mucosa but that the ratio was substantially lower in the mucosa, a result that suggests preferential incorporation of enteral glycine into mucosal purine synthesis.

The results also provided further evidence in favor of compartmentation in the purine nucleotide pools. First, although the free purines were rapidly labeled to isotopic equilibrium, the steady-state value was lower than the eventual isotopic enrichment of the RNA-bound purines. Thus, a direct precursor-product relationship did not exist. Second, the ratio of [M+1]glycine/[M+2]glycine in the free nucleotide pool was not only higher than that of tissue RNA but was highly variable. Similar kinetic evidence for compartmentation of the RNA precursors has been previously obtained from radiolabeling data. As in the present experiment, Plagemann (1971 and 1972), working with pyrimidine nucleotides, and Schroder and Rapaport (1984) with purine nucleotides, showed that the specific activity of the total ribonucleosides triphosphate precursor pool did not correlate with the rate of incorporation of the precursor into RNA. In both papers, the authors also observed that the rates of [³H]-uridine or [³H]-adenosine incorporation into RNA were unaffected by the expansion of the total cellular ATP or UTP pool. It appears therefore that the cellular acid-soluble free purine pool consists of a small, rapidly turning over pool of purine nucleotides destined for RNA synthesis and a much larger pool with a very low rate of turnover. The latter presumably represents the pools of ATP and GTP that are involved in cellular energy transactions.

Rates of protein and RNA synthesis.  An advantage of using [U¹³C]glycine as tracer is that it allows the simultaneous determination of protein and RNA synthesis within a single sample and with a single tracer. Nevertheless, it is important to emphasize that because we elected to use a lengthy period of "steady-state" labeling, the rates of protein synthesis represent largely those of the resident proteins within the mucosa and liver (Pain et al. 1978).

The first important observation was that the rates of synthesis of RNA in both liver and mucosa were only 38% (liver) and 29% (mucosa) of the respective rates of protein synthesis. There is not an extensive literature on this subject with which to compare our results. Sander et al. (1986), using the production of modified bases, reported a half-life of whole-body RNA of 3-5 d (17%/d). It is possible also to calculate from the results of Boza et al. (1996) that RNA synthesis in the liver and mucosa of pregnant mice were 27 and 41%/d, respectively. Finally, in both the perfused rat liver (Lardeux et al. 1987) and in the rat in vivo (Enwonwu and Munro 1970), it appears that in the presence of high levels of extracellular amino acids, total hepatic RNA was degraded at ~20-25%/d.

The marked difference in the relative rates of RNA and total protein synthesis implies that ribosomal RNA (and hence ribosome) turnover, which should dominate the kinetics of total RNA labeling, is much lower than the average for hepatic total protein. There is other evidence to support this conclusion. Ashford and Pain (1986) compared in vivo the rates of synthesis of total and ribosomal protein in rat liver and muscle. In both tissues, they found that the average fractional rate of ribosomal protein synthesis was much lower than that of total protein synthesis. Interestingly, the rate of ribosomal protein turnover that they obtained in the liver (22%/d) is very similar to the value for total hepatic RNA synthesis that we obtained.

A second interesting finding from the study was the remarkably low rate of RNA synthesis in the mucosa (16.6-26.9%/d). This was much lower than we had anticipated but is close to estimates of cell turnover of the villous enterocyte (25%/d; Reeds et al. 1993). On the basis of this result, it could be speculated that RNA synthesis and the assembly of polysomes are located mainly in the crypt and at the crypt-villus junction, and that the RNA population of the villous cells is very stable once differentiation has occurred.

If this is so, then any changes in the rates of mucosa protein and RNA synthesis, such as those observed with the dietary fiber intake, may largely reflect intestinal cytokinetics. Numerous studies have demonstrated a relationship between dietary fiber intake and morphological patterns in the small intestine (Cassidy et al. 1981, Schneeman 1982). In general, dietary fiber increases intestinal length and mass (Schneeman 1982) and increases the structural maturity of the villous enterocytes (Dirks and Freeman 1987). However, whether these morphological changes are truly associated with a modified cell turnover of the crypts of Lieberkuhns and/or changes in the functional and secretory activity must be assessed by isotopic studies. Vahouny and Cassidy (1984) and Ecknauer et al. (1981) used autoradiographic analysis of intestinal sections previously labeled with tracers of nucleic acid ([³H]thymidine) and secretory protein ([³H]glycine or Na₂³⁵SO₄) synthesis and confirmed that the administration of dietary fiber promoted an increased crypt cell turnover, a faster transit of cells up to the villous column and an enhanced production of glycoproteins by the goblet cells.

A final observation was that the rate of RNA synthesis in the liver was also higher in the HF than in the LF group, even though there were no significant differences in the rates of hepatic protein synthesis. We are not aware of previous evidence of any relationship between the level of dietary fiber or readily absorbable carbohydrates and the rates of hepatic RNA turnover and can only speculate as to the mechanistic significance.

First, it has been suggested that the purines incorporated into mucosal RNA can derive from nucleosides exported by the liver (Konishi and Ichihara 1979). It is possible therefore, that the hepatic RNA turnover in the liver of the HF group reflected this precursor role. However, as discussed above, this seems unlikely because the hepatic and mucosal RNA purines had apparently derived from different sources of glycine. Second, it should be emphasized that our dietary design inevitably confounded a change in cellulose with a change in the dietary content of readily absorbable monosaccharides. Thus, changes observed in the rates of hepatic RNA turnover could be related to a modified hormonal status, due to differences in the level or pattern of monosaccharide absorption, or associated with an increase in the hepatic disposal of products resulting from the hindgut fermentation of the cellulose. Further studies are clearly necessary to resolve this issue.

In summary, this paper proposes a method that we believe avoids some previous problems associated with the measurement of RNA synthesis in vivo. The method is certainly readily applied to the synthesis of ribosomal RNA, and because it is able to simultaneously determine the rate of tissue protein synthesis, it should provide a useful addition to methods applicable to the study of growth regulation, especially as it relates to the composition of the diet.

	FOOTNOTES

Manuscript received 29 January 1998. Initial reviews completed 3 April 1998. Revision accepted 21 May 1998.

	ACKNOWLEDGMENT

We are very grateful to Leslie Loddeke for her sound editorial comments.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Allsop J., Watts W. E. Activities of amidophosphoribosyl transferase (EC 2.4.2.14) and the purine phosphoribosyl transferases (EC 2.4.2.7 & 2.4.2.8) and the phosphoribosylpyrophosphate content of rat central nervous system at different stages of development. J. Neurol. Sci. 1980; 46:221-233[Medline]
• Ashford A. J., Pain V. M. Effect of diabetes on the rates of synthesis and degradation of ribosomes in rat muscle and liver in vivo. J. Biol. Chem. 1986; 261:4059-4065[Abstract/Free Full Text]
• Berthold H. K., Crain P. F., Gouni I., Reeds P. J., Klein P. D. Evidence for incorporation of intact dietary pyrimidine (but not purine) nucleosides into hepatic RNA. Proc. Natl. Acad. Sci. U.S.A. 1995; 92:10123-10127[Abstract/Free Full Text]
• Boza J. J., Jahoor F., Reeds P. J. Ribonucleic acid nucleotides in maternal and fetal tissues derive almost exclusively from synthesis de novo in pregnant mice. J. Nutr. 1996; 126:1749-1758
• Brauman J. I. Least squares analysis and simplification of multi-isotope mass spectra. Anal. Chem. 1966; 38:607-610
• Cassidy M. M., Lightfoot F. G., Grau L., Story J. A., Kritchevsky D., Vahouny G. V. Effect of chronic intake of dietary fibers on the ultrastructural topography of the rat jejunum and colon. Am. J. Clin. Nutr. 1981; 34:218-228[Abstract/Free Full Text]
• Chikenjii T., Kita K., Tatibana M Stimulation of de novo biosynthesis of purine and pyrimidine nucleotides in the liver of rats following burn injury. Metabolism 1988; 37:1114-1119[Medline]
• Crain P. F. Preparation and enzymatic hydrolysis of DNA and RNA for mass spectrometry. Methods Enzymol. 1990; 193:782-790[Medline]
• Dirks P., Freeman H. J. Effects of differing purified cellulose, pectin and hemicellulose fiber diets on mucosal morphology in the rat small and large intestine. Clin. Investig. Med. 1987; 10:32-38[Medline]
• Ecknauer R, Sirear B., Johnson L. R. Effect of dietary bulk on small intestinal morphology and cell renewal in the rat. Gastroenterology 1981; 81:781-786[Medline]
• Edmonds M. P., LePage G. A. The incorporation of gycine-2-¹⁴C into acid-soluble purine nucleotides. Cancer Res. 1955; 15:93-99
• Eichler D. C., Craig N. Processing of eukaryotic ribosomal RNA. Prog. Nucl. Acid Res. Mol. Biol. 1994; 49:197-239[Medline]
• Enwonwu C. O., Munro H. N. Rate of RNA turnover in rat liver in relation to intake of protein. Arch. Biochem. Biophys. 1970; 138:532-539[Medline]
• Ho C. Y., Miller K. V., Savaiano D. A., Crane R. T., Ericson K. A., Clifford A. J. Absorption and metabolism of orally administered purines in fed and fasted rats. J. Nutr. 1979; 109:1377-1382
• Jahoor F., Burrin D. G., Reeds P. J., Frazer M Measurement of plasma protein synthesis rate in infant pig: an investigation of alternative tracer approaches. Am. J. Physiol. 1994; 267:R221-R227[Abstract/Free Full Text]
• Konishi Y., Ichihara A. Transfer of purines from liver to erythrocytes. J. Biochem. 1979; 85:295-301[Abstract/Free Full Text]
• Lardeux B. R., Kleydrick S. J., Mortimore G. E. RNA degradation in perfused rat liver as determined from the release of [¹⁴C]cytidine. J. Biol. Chem. 1987; 262:14507-14513[Abstract/Free Full Text]
• Leleiko N. S., Bronstein A. D., Baliga B. S., Munro H. N. De novo purine nucleotide synthesis in the rat small and large intestine: effect of dietary protein and purines. J. Pediatr. Gastroenterol. Nutr. 1983; 2:313-319[Medline]
• Leleiko, N. S., Bronstein, A. & Munro, H. N. (1979) RNA synthesis in the small intestine: utilization of luminal purines. Gastroenterology 76: 1183 (abs.).
• Mackinnon A. M., Deller D. J. Purine nucleotide biosynthesis in gastrointestinal mucosa. Biochim. Biophys. Acta 1973; 319:1-4[Medline]
• Morley S. J. Signal transduction mechanisms in the regulation of protein synthesis. Mol. Biol. Rep. 1994; 19:221-231[Medline]
• Pain V. M., Clemens M. J., Garlick P. J. The effect of dietary protein deficiency on albumin synthesis and on the concentration of active albumin messenger ribonucleic aid in rat liver. Biochem. J. 1978; 172:129-135[Medline]
• Plagemann P.G.W. Nucleotide pools of Novikoff hepatoma cells growing in suspension culture. II Independent nucleotide pools for nucleic acid synthesis. J. Cell. Physiol. 1971; 77:241-258[Medline]
• Plagemann P.G.W. Nucleotide pools of Novikoff hepatoma cells growing in suspension culture. III. Effects of nucleosides in medium on levels of nucleotides in separate pools for nuclear and cytoplasmic RNA synthesis. J. Cell Biol. 1972; 52:131-146[Abstract/Free Full Text]
• Reeds P. J., Burrin D. G., Davis T. A., Fiorotto M. L. Postnatal growth of gut and muscle competitors or collaborators. Proc. Nutr. Soc. 1993; 52:57-67[Medline]
• Robins S. P., Reeds P. J. Measurement of a-ketoacids in plasma using an amino acid analyzer. Anal. Biochem. 1984; 142:24-27[Medline]
• Sander G., Topp H., Wieland J., Heller-Schoch G., Schoch G. Possible use of urinary modified RNA metabolites in the measurement of RNA turnover in the human body. Hum. Nutr. Clin. Nutr. 1986; 40:103-118[Medline]
• Savaiano D. A., Ho C. Y., Chu V., Clifford A. J. Metabolism of orally and intravenously administered purines in rats. J. Nutr. 1980; 110:1793-1804
• Schneeman, B. O. (1982) Pancreatic and digestive function. In: Dietary Fiber in Health and Disease (Vahouny, G. V. & Kritchevsky, D., eds.), pp. 73-83. Plenum Press, New York, NY.
• Schroder E. W., Rapaport E. Retinoic acid alters subcellular compartmentalization of ATP pools in 3T3 cells but not in HeLa cells. J. Cell. Physiol. 1984; 120:204-210[Medline]
• Shikama N., Brack C. Changes in the expression of genes involved in protein synthesis during Drosophila aging. Gerontology 1996; 42:123-136[Medline]
• Simmons H. A., Rising T. J., Cadenhead A., Hatfield P. J., Jones A. S., Cameron J. S. Radioisotopic studies of purine metabolism during administration of guanine and allopurinol in the pig. Biochem. Pharmacol. 1973; 22:2553-2563
• Sonoda T., Tatibana M. Metabolic fate of pyrimidines and purines in dietary nucleic acids ingested by mice. Biochim. Biophys. Acta 1979; 521:55-66
• Vahouny, G. V. & Cassidy, M. M. (1984) Dietary fiber and intestinal adaptation. In: Dietary Fiber in Health and Disease (Vahouny, G. V. & Kritchevsky, D., eds.), pp. 181-209. Plenum Press, New York, NY.
• Young E. G., Conway C. F. On the estimation of allantoin by the Rimini-Schryver reaction. J. Biol. Chem. 1942; 142:839-853[Free Full Text]

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences