Main

A noninvasive method for determining the whole-body degradation rates of tRNA, rRNA, and mRNA in human beings and other mammalian species was developed by us (1–5). This method is based on measuring (with HPLC) the renal excretion rates of special RNA catabolites. These are the modified ribonucleosides t6A, m22G, hU, and ψ, as well as the nucleobases m7Gua and oh8m7Gua. Modified ribonucleosides are formed posttranscriptionally by enzymatic modification of the primarily unmodified RNA precursor molecules during the maturation to functional t-, r-, and mRNA. Within mammalian species, the kinds and positions of modifications in RNA are highly conservative, and the average frequencies of occurrence of modified ribonucleosides within the different RNA classes have been calculated by us from published sequence data of RNA (2, 3). Generally, as a consequence of RNA turnover, the modified ribonucleosides are liberated and not reutilized for de novo RNA synthesis. Furthermore, the above mentioned modified RNA catabolites are virtually quantitatively excreted in urine (3, 4, 6). However, the nucleobase m7Gua is partially converted to oh8m7Gua before excretion in some species, including man (7, 8). On the basis of these prerequisites, we have developed calculation methods for determining the degradation rates of the different RNA classes by measuring the renal excretion rates of the above mentioned RNA catabolites. The details of the calculations are given in the Methods.

First, we determined the whole-body degradation rates of t-, r-, and mRNA in seven different mammalian species of various BW, including human adults and preterm infants. In total, the BW ranged from 30 g (mouse) to 125 kg (pig). We found that the degradation rates of the different RNA classes are highly correlated with the RMR (2, 4–6). The respective RMR of the different species were estimated on the basis of their BW using an empirical formula taken from literature (9). From these findings, we have hypothesized that at metabolic equilibrium there is a causal relationship between the degradation rates of RNA and the RMR. Consequently, noninvasive determinations of the whole-body degradation rates of RNA at metabolic equilibrium could turn out to be useful for assessing the RMR. Furthermore, in healthy subjects, modified RNA catabolites as indicators of the metabolic rate could turn out to be useful references to standardize other metabolic markers in (spontaneous) urine samples.

In the present pilot study, we have determined in 3-, 6-, 10-, 14-, and 18-y-old male and female children and adolescents the whole-body degradation rates of t-, r-, and mRNA (in micromoles per day per kilogram BW) under normal living conditions as well as the correlations between the degradation rates of the different RNA classes and the RMR (in kilojoules per day per kilogram BW). The RMR of the subjects were calculated on the basis of their BW using published empirical formulas (10).

METHODS

Subjects and urine collection.

Fourteen female and 14 male healthy subjects for each age group (3, 6, 10, 14, and 18 y) were enrolled in the study. All subjects were participants of the Dortmund Nutritional and Anthropometric Longitudinally Designed Study (DONALD), in which the normal development of children and adolescents is investigated. The survey has been approved by the international Scientific Commission of the Forschungsinstitut für Kinderernährung. Food intake was ad libitum. The average BW and values of BMI (kilogram BW per square meter height) of the subjects in the different age groups are given in Table 1. By taking published values of BMI (11) as reference values, recommended by the European Childhood Obesity Group, the BMI of seven female and five male subjects in our study was >97th percentile. These subjects can be classified as adipose. One pooled urine per subject per day was collected quantitatively with controlled collection times (mean ± SD, 1373 ± 145 min). The urine samples were frozen at −20°C until analysis. For the calculation of the degradation rates of the different RNA classes, the determined amounts of the modified RNA catabolites (t6A, m22G, hU, ψ, m7Gua, and oh8m7Gua) were extrapolated to 24-h values.

Table 1 Average BW, BMI, and calculated RMR in 3-, 6-, 10-, 14-, and 18-y-old female and male subjects Values are mean ± SD of n = 14 individuals per age group and sex. * Formulas for calculating RMR are given in Methods.
Full size table

Determination of the modified ribonucleosides and nucleobases inurine.

The analytical procedures for determining the amounts of the RNA catabolites in urine used for determining the RNA degradation rates and the evaluation of the methods have been described elsewhere (3, 4, 12). Summing up, standards of the ribonucleosides hU and ψ were purchased from Sigma Chemical Co. (Deisenhofen, Germany), and m22G, from Pharmacia P-L Biochemicals (Freiburg, Germany); t6A was a kind gift from Professor Schlimme (Federal Dairy Research Centre, Kiel, Germany). In the first step, the ribonucleosides were enriched from urine by affinity chromatography using a boronate gel (Affi-Gel 601, Bio-Rad, München, Germany). Then, HPLC analysis of the ribonucleoside fraction was performed using a reversed-phase gel (Nucleosil 120–5 C18, Macherey-Nagel, Düren, Germany) and a ternary gradient composed of ammonium dihydrogen phosphate, methanol, and acetonitrile. ψ, m22G, and t6A were quantified by their absorption at 254 and 280 nm, and hU, at 220 and 230 nm (4). The nucleobase standard m7Gua was purchased from Sigma Chemical Co. The nucleobase standard oh8m7Gua was gained from the ribonucleoside oh8m7G (8-hydroxy-7-methylguanosine) by hydrolysis with perchloric acid; oh8m7G was obtained from Ortho Pharmaceutical Corporation (Raritan, NJ). m7Gua was enriched from urine on a cation-exchange column with aromatic sulfonic acid as active groups (SPE system, Baker, Griesheim, Germany). Finally, an aliquot of the m7Gua-containing fraction was analyzed by cation-exchange HPLC (DC-6A, Benson, Reno, NV) using ammonium dihydrogen phosphate with methanol as the eluent. m7Gua was quantified by its absorption at 254 and 280 nm (3). The method for determining oh8m7Gua has not been described before and is therefore given in more detail here. The first step using the cation-exchange column with aromatic sulfonic acid as active groups (SPE system, Baker) was the same as has been previously described by us for m7Gua (3). However, whereas m7Gua remained bound to the gel during the first elution with 6 mL of 100 mM formic acid, oh8m7Gua was eluted quantitatively under these conditions. This fraction was collected and evaporated to dryness at 95°C, then redissolved in 500 μL of H2O. Then oh8m7Gua was determined by HPLC under the following conditions: injected sample volume, 5 μL; cation-exchange column, DC-6A (Benson), 250 × 4.6 mm; elution, 0.02 M ammonium dihydrogen phosphate in 5% acetonitrile, pH 6; flow rate, 0.2 mL/min; column temperature, 25°C; detection at 254 and 280 nm; and retention time, 36 min.

Determination of the whole-body degradation rates oftRNA, rRNA, and mRNA.

The details of our calculation methods for determining the degradation rates of the different RNA classes have been described previously (2–4). Therefore, only the essentials are outlined here and the final formulas for determining the degradation rates of the different RNA classes are given. There are three different markers that are specific for determining the degradation of tRNA (2, 4). These are t6A, m22G, and hU. Their average frequencies of occurrence per tRNA molecule are 0.22 residues t6A, 0.5 residues m22G, and 2.6 residues hU. Therefore, the whole-body degradation rate of tRNA can be calculated in three ways:MATH 1 MATH 2 MATH 3

The whole-body degradation rate of rRNA is calculated on the basis of excreted ψ (2, 3). There are on average 95 residues of ψ per rRNA molecule (28S + 18S + 5.8S RNA; 5S RNA is unmodified). However, ψ occurs also in tRNA, and the fraction of ψ in urine stemming from degraded tRNA must be first subtracted from the total excreted ψ. Thereafter, the degradation of rRNA can be determined by dividing the residual ψ by 95. The fraction of ψ in urine from degraded tRNA can be determined on the basis of excreted t6A, m22G, or hU because the intramolecular ratios of residues ψ:t6 (3.02:0.22 = 13.7), ψ:m22G (3.02:0.5 = 6), or ψ:hU (3.02:2.6 = 1.16) per average tRNA molecule are known (2, 3). Therefore, ψ from degraded tRNA can be obtained by multiplying urinary t6A with 13.7, m22G with 6, or hU with 1.16. Consequently, the whole-body degradation rate of rRNA can also be calculated in three ways:MATH 4 MATH 5 MATH 6

The whole-body degradation rate of mRNA is calculated on the basis of excreted m7Gua, which in the 3- to 18-y-old subjects is partly (30–40%) oxidized to oh8m7Gua before excretion (8). Therefore, to calculate the degradation rate of mRNA, the sum of m7Gua + oh8m7Gua in urine must be formed. There is one residue of m7Gua per mRNA molecule. However, m7Gua also occurs in tRNA and rRNA, and the fractions in urine of m7Gua (and oh8m7Gua) from degraded tRNA and rRNA must be subtracted from totally excreted m7Gua (and oh8m7Gua) before determining the degradation rate of mRNA by the residual m7Gua (and oh8m7Gua). The fractions of m7Gua (and oh8m7Gua) in urine from degraded tRNA can be calculated because again the ratios of m7Gua:t6A (0.44:0.22 = 2), m7Gua:m22G (0.44:0.5 = 0.88), or m7Gua:hU (0.44:2.6 = 0.17) per average tRNA molecule are known (2, 3). Furthermore, an rRNA molecule (28S + 18S + 5.8S RNA) contains one residue of m7Gua and, as already mentioned, 95 residues of ψ (2, 3). Therefore, dividing the amount of ψ in urine from degraded rRNA by 95 yields m7Gua (and oh8m7Gua) in urine from degraded rRNA. Consequently, the whole-body degradation rate of mRNA can be calculated again in three ways:MATH 7 MATH 8 MATH 9

Calculation of the resting metabolic rates.

The RMR were calculated on the basis of the BW using the following formulas after Schofield (10):MATH 10 MATH 11 MATH 12 MATH 13

Statistical analysis.

The effects of age (3, 6, 10, 14, 18 y) and sex on the RMR and the degradation rates of the different RNA classes were investigated using ANOVA (SAS 6.12, SAS Institute Inc., Cary, NC). Linear regression analyses were performed using SAS 6.12.

RESULTS

The calculated RMR per kilogram BW are on average 2.2 and 2 times higher, respectively in the 3-y-old female and male subjects compared with the same-sex 18-y-old subjects (Table 1). The average degradation rates of tRNA, rRNA, and mRNA in the different age groups are given in Table 2. By using the three different degradation markers for tRNA, namely m22G, t6A, and hU, we obtained very similar degradation rates of tRNA. The same is true for the degradation rates of rRNA determined by ψ (or mRNA determined on the basis of m7Gua + oh8m7Gua), and corrected for by the different degradation markers of tRNA (see Methods). The maximum difference by using the different RNA degradation markers exists in the case of the degradation rates of tRNA, in which the values determined by hU were on average 16% higher in comparison to the respective values determined by m22G. The average degradation rates of the different RNA classes per kilogram BW decrease continuously from the 3-y-old to the 18-y-old subjects. They are on average 1.8 times higher in the youngest compared with the oldest group. For all investigated parameters (RMR as well as degradation rates of t-, r-, and mRNA), the decrease with age is highly significant (p < 0.0001). Furthermore, the RMR as well as the degradation rates of tRNA and rRNA are significantly higher (p < 0.0001) in the boys than in the girls. In contrast, the degradation rates of mRNA tend to be higher in the girls with the exception of the 18-y-old group. The effect of sex on the degradation rate of mRNA is only significant (p < 0.05) for the values given in the last two columns of Table 2, for which the determinations of mRNA degradation were performed by urinary m7Gua + oh8m7Gua, ψ, and either t6A or hU. For all measurements, there was no significant age or sex interaction.

Table 2 Average whole-body degradation rates of tRNA, rRNA, and mRNA in 3-, 6-, 10-, 14-, and 18-y-old female and male subjects Values are mean ± SD of n = 14 individuals per age group and sex. Formulas for calculating the degradation rates of the different RNA classes by the renal excretion rates of the RNA catabolites m22G, t6A, hU, ψ, m7Gua, and oh8m7Gua are given in Methods.
Full size table

The degradation rates of the different RNA classes and the RMR per day per kilogram BW in all female or male subjects are highly correlated (Table 3). The correlation coefficients range from 0.68 to 0.82 (p < 0.0001). In Figure 1, the regression lines between the RMR and degradation rates of tRNA, rRNA, and mRNA in the female subjects are presented. By using the different RNA degradation markers, the resulting regression lines of a given RNA class run largely parallel.

Table 3 Correlations between the RMR (kJ/d/kg BW; x values) and the whole-body degradation rates of tRNA, rRNA, and mRNA (μmol/d/kg BW; y values) in 3-, 6-, 10-, 14-, and 18-y-old female and male subjects as well as regression formulas
Full size table
Figure 1
figure 1

Regression lines between the calculated RMR and the whole-body degradation rates of tRNA (A), rRNA (B), and mRNA (C) per day per kilogram BW in 3-, 6-, 10-, 14-, and 18-y-old female subjects (n = 14 per age group). The degradation rates of tRNA (A) were calculated by the renally excreted m22G (♦, line 3), t6A (Δ, line 2), and hU (○, line 1); those of rRNA (B) by ψ and m22G (♦, line 1), ψ and t6A (Δ, line 2), and ψ and hU (○, line 3); and those of mRNA (C) by m7Gua + oh8m7Gua, ψ, and m22G (♦, line 1), m7Gua + oh8m7Gua, ψ, and t6A (Δ, line 2), and m7Gua + oh8m7Gua, ψ, and hU (○, line 3). Details are given in Methods.

Full size image

DISCUSSION

In the current pilot cross-sectional study, we present basic whole-body degradation rates of t-, r-, and mRNA per day per kilogram BW in healthy 3- to 18-y-old subjects. Our present results demonstrate for the first time that, in female and male children and adolescents, the degradation rates of the different RNA classes are highly correlated with the RMR estimated by BW (10). These findings are in accordance with our previous corresponding findings in various mammalian species, which differed in their average BW by four orders of magnitude (2, 4–6). The modified RNA catabolites in urine used as markers for the whole-body degradation rates of t-, r-, and mRNA originate from the whole-body lean mass. The degradation rates of a given RNA class obtained on the basis of the different calculation procedures (see Methods) are remarkably similar. The most pronounced deviations (on average 16%) are observed in the case of tRNA when the degradation rates are determined either by hU or m22G (Table 2). This may be because the abundance of hU and m22G in the idealized average tRNA molecule that we modeled (see Methods) is slightly different from the abundance of hU and m22G in an average tRNA molecule in vivo. In practice, one can calculate the average of the differing degradation rates of tRNA determined by the different markers. In the case of rRNA and mRNA, the differences between the mean degradation rates determined by the different degradation markers are even smaller (Table 2). Also, differences in the regression lines calculated by the various RNA degradation markers (Fig. 1, Table 3) are most pronounced in the case of tRNA, which also may partly be explained by the reason given above.

Interestingly, the average degradation rates of tRNA and rRNA per kilogram BW are significantly higher (on average approximately 13 to 24%) in male subjects when compared with female subjects in the various age groups (Table 2). Also, the calculated average RMR per unit BW in the different age groups (Table 1) as well as data taken from literature for RMR (13–15), total energy expenditure (14–17), and lean tissue mass (16–18) indicate that these variables per unit BW tend to be slightly higher in male than in female children. Even though an exact comparison of the different variables is impossible, we conclude that the gender-specific differences in the metabolic rate per kilogram BW of the different age groups is reflected by the degradation rates of tRNA and rRNA. In contrast, the whole-body degradation rates of mRNA are slightly higher in the female than in the male subjects, except in the 18-y-old group (Table 2). This indicates that the mRNA turnover in girls per unit metabolized energy might be comparatively higher than in boys. However, for definite conclusions, exact determinations of the RMR and total energy expenditure are necessary. Generally, this pilot study leaves questions unanswered, e.g. what are the correlations between the degradation rates of the different RNA classes and the RMR measured by indirect calorimetry, what are the correlations between the RNA degradation rates and the total energy expenditure measured by the double-labeled water method, and what is the relationship of these measurements to the lean body mass measured by dual-energy x-ray absorptiometry. A joint project to answer these questions is under way.

The facts that the degraded RNA stem from the metabolically active body mass and are highly correlated with the RMR let one conclude that in healthy subjects, RNA degradation rates could be useful to assess the metabolic rate. Furthermore, the above mentioned properties make the RNA degradation rates or the excreted modified RNA catabolites candidates for normalizing in healthy subjects other metabolites of interest in urine to these simultaneous indicators of the metabolically active body mass and its activity. This could particularly be of interest in studies in which only spontaneous urine sample collection is possible. Generally, the concentration of creatinine in spontaneously voided and often variously concentrated urine samples is used as a reference for the relation of other metabolites of interest. By doing this, the concentration of a metabolite of interest can be compared between various spontaneous urine samples. Urinary creatinine stems mainly from the energy metabolism of the skeletal muscle and has been proposed as a marker for the estimation of the skeletal muscle mass (19, 20) as well as the lean body mass (21, 22) in healthy subjects. However, skeletal muscle can constitute an individually different, age- and sex-dependent part of the total whole-body lean mass, and urinary creatinine is partly dependent on the supply of creatine and creatinine from the diet (19, 22). For example, vegetarians excrete about 30% less creatinine in their urine in comparison with a nonvegetarian reference population (23).

The usefulness of modified RNA catabolites as a reference depends on showing that the amounts excreted in urine are derived only from tissue metabolism and are not influenced by intake from the food. It has been demonstrated that, in contrast to creatine and creatinine, the tRNA-specific degradation marker t6A, when given orally, is virtually not excreted in urine (24), which could be because t6A is not absorbed or because it is metabolized by the microorganisms of the gut. Furthermore, the average molar quotients of the tRNA-specific catabolites hU/t6A (12.5) and m22G/t6A (2.1) in the analyzed urine samples (not shown) are very similar to the corresponding ratios in average mammalian cytoplasmic tRNA of 11.8 and 2.2, respectively, calculated by us on the basis of published sequence data of cytoplasmic tRNA [see Methods] (2, 4). This finding makes it very probable that hU and m22G are also not absorbed from the food or the microflora of the gut because the above given molar quotients of the tRNA-specific modified nucleosides are highly conserved in mammalian tRNA but are different in tRNA of plants or bacteria. In the case of ψ, it was shown by Schöch et al. (25, 26) that in two children, an isocalorically virtually nucleic acid–free diet for 5 d had no significant effect on the excretion of ψ in comparison to ad libitum eating. Only a diet extremely rich in nucleic acids led to an increase in the excretion of ψ of 8 and 10%, whereas the corresponding increase in creatinine was 18 and 35% (25, 26). Hence, the contribution of ψ in the diet to the excretion of ψ is negligible under normal nutritional conditions. This is also supported by the fact that the molar quotient of ψ/t6A of 43.4 ± 2.4 is very constant in the 140 ad libitum eating subjects investigated in the current study (not shown) and very similar to the quotients of ψ/t6A determined by us in other mammalian species (with different nutritional behavior) of 46 (gorilla), 39 (orangutan), 48 (gibbon), 48 (baboon), 46 (chimpanzee), 42 (mangabey), 40 (langur), 46 (marmoset), 36 (goat), 48 (sheep), and 50 (camel; Topp H, Schöch G, unpublished results). In the above mentioned study by Schöch et al. (26), neither giving a virtually nucleic acid–free nor a nucleic acid–extremely rich diet led to significant alterations in the renal excretion of m7Gua in comparison to ad libitum eating. Also, in adults, no significant effect on m7Gua excretion was found with a purine-poor, meat-free diet or a diet consisting only of glucose (27). Investigating the concentrations of ψ, m22G, and t6A in breast milk of mothers with preterm infants led to the conclusion that the supplied amounts of the modified nucleosides virtually do not contribute to the excreted amounts in urine (28). On the other hand, model calculations indicate that creatinine and creatine supplied with breast milk or infant formulas can be considerable (29). We have further proved the endogenous origin of the modified RNA catabolites ψ, t6A, and m7Gua by comparing the excretion rates between conventional and germ-free Wistar rats (30). These findings establish the validity of using modified RNA catabolites as a reference for other metabolites in urine samples.