INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

AMP DEAMINASE (EC 3.5.4.6) catalyzes the deamination of AMP to IMP and ammonia. In skeletal muscle, this enzyme is activated during exercise when the rate of ATP utilization exceeds the potential of the cell to resynthesize ATP. Because AMP deamination displaces the equilibrium of the myokinase reaction (2 ADP  ATP + AMP) toward ATP resynthesis, a proposed role for AMP deaminase is to alleviate the exercise-induced decrease in the ATP/ADP ratio and its inhibitory effect on muscle contraction (17). Skeletal muscle AMP deaminase activities are quite variable across a wide array of neuromuscular disorders, owing to a combination of genetic and pathological factors that influence the expression of this enzyme (5, 10, 14, 22, 23, 35). Moreover, ~2% of skeletal muscle biopsies submitted for pathological evaluation have an AMP deaminase deficiency (32). Perhaps reflecting the proposed role for AMP deaminase in skeletal muscle energetics and function, one-half of these deficient individuals exhibit exercise-related muscle cramps, pain, and early fatigue but have no other neuromuscular abnormalities. Conversely, AMP deaminase deficiency in other individuals is secondary to a wide array of well-defined neuromuscular disorders. Clinical variability forms the basis of a proposal for inherited and acquired forms of skeletal muscle AMP deaminase deficiency (5).

The gene encoding the muscle-specific isoform of AMP deaminase (AMPD1) has been localized to the short arm of chromosome 1 (34). A prevalent non-sense mutation in exon 2 of the AMPD1 gene was found at a frequency of 12% in 59 Caucasians and 19% in 13 African-Americans surveyed but was not present in any of 106 Japanese individuals (22). The encoded polypeptide product of this mutant sequence would be severely truncated and catalytically inactive, thereby accounting for most, if not all, cases of inherited skeletal muscle AMP deaminase deficiency. Moreover, the high prevalence of the identified AMPD1 mutant allele also predicts a relative large group of asymptomatic, deficient individuals in the Caucasian and African-American populations (10, 22). Accordingly, AMP deaminase deficiency was observed in 2% of a healthy Swedish population sample while the remainder of these subjects exhibited a bimodal distribution of intermediate and high enzyme activities (24).

AMP deaminase expression in skeletal muscle is also dependent on fiber-type composition. Previous studies in the rat (31, 37), rabbit (28, 31), and pigeon (31) have demonstrated higher enzyme activities in muscles with a predominance of fast-twitch (white; type II) fibers compared with those with a majority of slow-twitch (red; type I) fibers. Whereas the AMPD1 isoform predominates in all mammalian skeletal muscle fibers (28-30), differential AMP deaminase gene expression may contribute to quantitative variations in total enzyme activity across muscle groups with different fiber-type compositions. AMPD1 mRNAs are relatively more abundant in fast-twitch than slow-twitch skeletal muscles of the rat (33), whereas the reverse is true for AMPD3 mRNAs (18).

Northern blot analysis of total cellular RNA isolated from mixed-fiber adult human skeletal muscle demonstrates that AMPD1 mRNAs are relatively more abundant than AMPD3 mRNAs (20). Immunohistochemical examination reveals that the more abundant isoform M (AMPD1) is most prevalent in type II fibers, whereas isoform E (AMPD3) appears localized to type I fibers (37a). Consistent with these combined data, a positive correlation has been reported between total AMP deaminase activity and the proportion of type II fibers in muscle biopsy material (6). Moreover, it has been shown that physical training lowers skeletal muscle AMP deaminase activity (11, 27) and that a high level of physical activity coincides with a higher percentage of type I fibers (13, 36). However, no difference in total AMP deaminase activities was observed when type I and type II fibers were analyzed separately (26). Furthermore, a decrease in total AMP deaminase activity was reported concurrent with an increase in the proportion of type II fibers in a sprint training study (3, 11). Therefore, AMP deaminase expression appears also to be influenced by the level of physical activity.

The goal of this study was to determine how genetic and fiber-type composition, training status, and gender influence variability of AMP deaminase activity in healthy human skeletal muscle, where pathological effects on expression are negligable. AMPD1 allele frequencies were determined in a large sample of the healthy population, and skeletal muscle biopsies were obtained from a subset of these individuals. AMP deaminase activity, fiber-type composition, and training status were assessed in both men and women according to AMPD1 genotype, thus enabling each variable to be analyzed independent of the profound genetic influence of the prevalent mutant allele.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Subjects. A healthy Swedish population sample was composed of 175 healthy, physically active Caucasian students at the College for Sports and Recreational Instructors. The main focus of the school is physical education, and the subjects were well trained but not at an elite or competitive level. Muscle biopsy samples were obtained from 39 subjects in this group and from 31 other healthy subjects taking part in various training studies in which the AMP deaminase activity has previously been determined. Subjects with high, intermediate, and low AMP deaminase activity were selected for the present study. In total, muscle biopsy samples were obtained from 70 subjects (38 men and 32 women).

Each of these 70 individuals completed a questionnaire regarding their health and physical activity habits. All considered themselves healthy, and none expressed complaints of exercise-induced muscle cramps or myalgias. The assessed physical activity level was summarized in a training index 1-4, for which number of hours per week of engagement in physical activity and the intensity of training were considered. Training index 1 = sedentary subjects, training <1 h/wk; 2 = moderately trained subjects, training 1-3 h/wk; 3 = well-trained subjects, training 4-10 h/wk; and 4 = subjects at national elite level, training at very high intensity of >10 h/wk.

The age, height, and weight (mean and range) of the men were, respectively, 25 (20-35) yr, 180 (167-194) cm, and 76 (57-98) kg and of the women were, respectively, 24 (20-31) yr, 168 (158-182) cm, and 65 (51-84) kg. The subjects were fully informed of procedures of the experiment and potential risks before giving their consent to participate. The study was approved by the Ethics Committee of the Karolinska Institutet.

Muscle and blood sampling, treatment, and analysis. Percutaneous muscle samples were taken from the muscle quadriceps femoris vastus lateralis at rest by using a needle-biopsy technique described by Bergstrom (2). One muscle biopsy sample from each individual was frozen in isopentane precooled with liquid nitrogen and subsequently analyzed histochemically for fiber types by using a myofibrillar ATPase stain (4), which was used to distinguish type I and type II fibers. The second muscle biopsy sample was immediately frozen in liquid nitrogen and subsequently freeze-dried. A small portion of the freeze-dried sample was dissected free from visible connective tissue and blood under a dissection microscope, weighed, homogenized in 0.1 M phosphate buffer, pH 7.7, and used for the analysis of AMP deaminase activity by HPLC, as previously described (25). Blood samples were collected from all individuals for extraction of genomic DNA for genotype analysis.

Genotype analysis. Genomic DNA was isolated from fresh whole blood according to a previously described method (12). The region immediately surrounding exon 2 of the AMPD1 gene was amplified by using a modification of a previously described method (9). A sense strand amplimer (5'-CTTCATACAGCTGAAGAGACA-3') was designed to create an NspI restriction endonuclease site when the mutant allele (characterized by the CT transition at nucleotide +34 located at the exon 2/intron 2 boundary of the AMPD1 gene) is used as the template, i.e., ACATgt. The anti-sense strand amplimer (5'-GAATCCAGAAAAGCCATGAGC-3') begins at nucleotide 192 of intron 2. A polymerase chain reaction (PCR) was performed in a DNA thermal cycler (Ericomp) for 50 cycles at 94°C denaturing temperature for 1 min, 50°C annealing temperature for 1 min, and 72°C extension temperature for 2 min. Each 100-µl PCR reaction contained 20 mM Tris · HCl, pH 8.4, 1.5 mM MgCl₂, 200 µM of each deoxyribonucleoside triphosphate, 200 ng of each primer, 2.5 U of Taq polymerase (GIBCO-BRL), and 10 µl of genomic DNA. After PCR, each sample was phenol extracted, ethanol precipitated, and resuspended in 20 µl of DNase-free water. Four microliters of each resuspended PCR product were run on a 1% agarose gel to visualize relative yields of the expected 214-bp product. Typically, one-tenth of the resuspended PCR product was then digested with 1.5 U of NspI (Amersham Life Sciences) overnight at 37°C in buffer provided by the supplier. Approximately one-half of each digest was loaded onto an 8% acrylamide gel, fractionated for 3 h at 150 V, and visualized after ethidium bromide staining. Product amplified from the mutant allele is cut by NspI into two fragments of 191 and 23 bp. The 23-bp fragment is barely detectable, so genotype is determined as follows: homozygote for the normal allele exhibits a single 214-bp band; heterozygotes display 214- and 191-bp bands; homozygote for the mutant allele is identified by a single 191-bp band (see Fig. 1 for representative genotyping results).

View larger version (31K): [in this window] [in a new window]   Fig. 1.   Polymerase chain reaction (PCR)-based genotyping of AMPD1 mutant allele. A 214-bp region surrounding exon 2 of AMPD1 gene was amplified by PCR. Reaction products were digested with Nsp1 and fractionated on an 8% acrylamide gel. The 214- and 191-bp bands represent the normal (+) and mutant () alleles, respectively. Lane 1, cloned mutant homozygote control; lane 2, mutant homozygote; lane 3, heterozygote; lane 4, size standard; lanes 5 and 6, normal homozygotes; lane 7, cloned normal homozygote control.

Statistical analysis. ANOVA or an unpaired Student's t-test was applied to determine differences between groups (genotype and gender). Correlation between variables was tested by linear-regression analysis (single or multiple). In the multiple-regression analysis, AMP deaminase activity was chosen as a dependent variable (y), and the independent variables were x1 = genotype, x2 = gender, x3 = relative proportion of type I fibers, and x4 = training index.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Genotype analysis of 175 healthy Swedish individuals selected at random was based on the previously identified mutant allele involving a C34T transition in the AMPD1 open-reading frame (22). This point mutation converts a glutamine codon (CAA) into a premature stop codon (T[U]AA) and is located at the last nucleotide base pair in exon 2. The normal and mutant alleles can be distinguished by restriction endonuclease analysis of PCR-amplified genomic DNA (9, 22). Table 1 shows that 130 subjects were homozygotes for the normal allele, 42 were heterozygotes, and 3 were homozygotes for the mutant allele. These data can be used to derive a mutant allele frequency of 13.7% (48 mutant alleles/350 total alleles) and an incidence of 1.7% for the inherited deficiency in this healthy population sample. Using this mutant allele frequency, the Hardy-Weinberg principle predicts (rounded off to the nearest whole number) 130 homozygotes for the normal allele, 41 heterozygotes, and 3 homozygotes for the mutant allele in a population sample of 175 individuals.

AMP deaminase activity was quantitated in crude homogenates prepared from muscle biopsies obtained from 4 homozygotes for the mutant allele and 66 additional healthy subjects (47 homozygotes for the normal allele and 19 heterozygotes). A frequency-distribution diagram of enzyme activity according to AMPD1 genotype is presented in Fig. 2. Homozygotes for the normal allele uniformly have high enzyme activities [mean 1,560 mmol · min¹ · kg dry mass¹ (range 1,010-2,200 mmol · min¹ · kg dry mass¹)]. With one exception, heterozygotes display intermediate activities [mean 559 mmol · min¹ · kg dry mass¹ (range 337-947 mmol · min¹ · kg dry mass¹)]. Furthermore, the four homozygotes for the mutant allele have very low enzyme activities [mean 10 mmol · min¹ · kg dry mass¹ (range 6-14 mmol · min¹ · kg dry mass¹)]. Only one individual in this subset of the healthy population sample, a heterozygote with very low enzyme activity (4 mmol · min¹ · kg dry mass¹), varies from this relationship between AMPD1 genotype and relative AMP deaminase activity. Consequently, this individual was not used in any statistical analysis where AMPD1 genotype was considered as a variable.

View larger version (43K): [in this window] [in a new window]   Fig. 2.   Skeletal muscle AMP deaminase activity in healthy subjects with different AMPD1 genotypes. Histogram illustrates AMP deaminase activity (mmol · min1 · kg dry mass1) and AMPD1 genotype in 70 healthy subjects. Mutant homozygotes and atypical heterozygote all have <15 mmol · min1 · kg dry mass1 of AMP deaminase activity.

Fiber typing was also performed by using remaining biopsy material obtained from the 70 subjects. There was no statistical difference in the relative proportion of type I fibers between the three different genotypes as determined by factorial ANOVA analysis. The mean (range) percentage of type I fibers is 56.8% (27.1-85.9%) in homozygotes for the normal allele, 50.5% (28.7-69.0%) in heterozygotes, and 61.0% (55.8-69.0%) in homozygotes for the mutant allele. The relative proportion of type I fibers is lower in men than in women, as determined by an unpaired two-tailed t-test (P < 0.05). The mean (range) percentage of type I fibers is 52.9% (27.1-85.9%) in men and 59.0% (30.6-74.9%) in women. Although their relative small number precludes a demonstration of statistical significance, it is perhaps relevant to note that all five subjects with very low AMP deaminase activities had a higher than mean percentage of type I fibers for their gender; the two male subjects have 55.8 and 56.2% of type I fibers, respectively, and the three women have 62.0, 62.8, and 69% of type I fibers, respectively.

AMP deaminase activity vs. percentage of type I fibers is graphed according to AMPD1 genotype in Fig. 3. There is a negative correlation between enzyme activity and the relative proportion of type I fibers in homozygotes for the normal allele (r = 0.34, P < 0.05; Fig. 3A) and a negative tendency in heterozygotes (r = 0.24, P = 0.35; Fig. 3B). Conversely, there is a positive correlation between residual AMP deaminase activity and relative type I fiber composition in homozygotes for the mutant allele (r = +0.99, P < 0.05; Fig. 3C), despite the small number of individuals with this genotype. Reciprocal correlations for each AMPD1 genotype are also evident for AMP deaminase activity and the relative proportion of type II fibers [r = +0.34, P < 0.05 (normal homozygotes); r = +0.24, P < 0.35 (heterozygotes); r = 0.99, P < 0.05 (mutant homozygotes)].

View larger version (10K): [in this window] [in a new window]   Fig. 3.   Correlative analysis of skeletal muscle AMP deaminase activity vs. fiber-type composition according to AMPD1 genotype. Enzyme activities (mmol · min1 · kg dry mass1) are plotted vs. %type I fibers for normal homozygotes (A; r = 0.34, P < 0.05, n = 46), heterozygotes (B; r = 0.24, P = 0.35, n = 17), and mutant homozygotes (C; r = +0.99, P < 0.05, n = 4). Best fit line parameters: A, y = 2,021.2  8.2x; B, y = 705.4  2.9x; C, y = 26.3 + 0.6x. Note: due to the unknown effect of a suspected rare mutant allele in atypical heterozygote with low levels of skeletal muscle AMP deaminase activity, this individual was excluded from any statistical analysis involving AMPD1 genotype.

When analyzed by AMPD1 genotype, skeletal muscle AMP deaminase activities in women are on average only 85-88% of those in men. The mean (±SD) AMP deaminase activity in homozygotes for the normal allele is 1,673 ± 255 mmol · min¹ · kg dry mass¹ for men and 1,420 ± 318 mmol · min¹ · kg dry mass¹ for women and in heterozygotes 592 ± 188 mmol · min¹ · kg dry mass¹ for men and 519 ± 83 mmol · min¹ · kg dry mass¹ for women.

Multiple-regression analysis was applied to investigate the correlation between AMP deaminase activity on the one hand and genotype, gender, training index, and fiber-type composition (type I) on the other. Genotype (P < 0.0001), gender (P < 0.01), and training index (P < 0.01), but not fiber-type composition (P > 0.05), make a significant contribution to total AMP deaminase activity. The lack of statistical significance for fiber-type composition in the multiple-regression analysis is most likely explained by a stronger single correlation between AMP deaminase activity and training index than between AMP deaminase activity and relative type I fiber proportion. Furthermore, there is a simple correlation between training index and the relative type I fiber proportion (P = 0.01) in this healthy population sample.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

An inherited deficiency of skeletal muscle AMP deaminase activity is found in 1-2% of all individuals referred for neurological evaluation and is associated with a generalized exercise-induced myalgia (32). A mutant allele that introduces a premature stop codon into the AMPD1 open-reading frame has been identified in all reported myopathic patients who have been genotyped (10, 22). However, this non-sense mutation is equally prevalent in the general population (10, 22). This study has confirmed this observation and related it to the observed variation in skeletal muscle AMP deaminase activity in a healthy population sample (24).

Skeletal muscle AMP deaminase activity is variable across the healthy population, with nearly 2% of individuals exhibiting <2% of the mean value determined for the rest of the group (24). Data presented in this study demonstrate that this variability is predominantly due to the previously identified non-sense mutation at nucleotide +34 in the AMPD1 open-reading frame. With only one exception, 70 individuals in a healthy population sample exhibit high, intermediate, or low skeletal muscle AMP deaminase activities that can be attributed to their AMPD1 genotype. All 47 subjects in the high-activity group are homozygotes for the normal allele, all 18 individuals in the intermediate-activity group are heterozygotes, and 4 of 5 subjects in the low-activity group are homozygotes for the mutant allele. The sole exception, a heterozygote with very low enzyme activity, suggests the presence of a rare second AMPD1 mutation in this individual. Furthermore, Western blot analysis indicates that this putative rare mutant allele encodes an extremely truncated or unstable polypeptide, as judged from the lack of any detectable immunoreactive band in a skeletal muscle extract prepared from this individual (data not shown).

AMPD1 genotype analysis of 175 healthy individuals selected at random reveals a mutant allele frequency of 13.7% and an incidence of 1.7% for the homozygous mutant genotype. These data are consistent with previous studies that have reported AMPD1 mutant allele frequencies of 11.9-12.3% in different, randomly selected Caucasian population samples (10, 22) and skeletal muscle AMP deaminase deficiency in 2% of a healthy population sample (24). The present data link these two observations in the same population sample and reinforce the contention that there is a relatively large group of asymptomatic, myoadenylate deaminase-deficient individuals that harbor the same AMPD1 mutation found in patients with an associated exercise-induced myopathy (10, 22).

When different AMPD1 genotypes are separately analyzed, relative fiber-type composition is also seen to influence skeletal muscle AMP deaminase activity. A negative correlation between the percentage of type I fibers and AMP deaminase activity is observed in homozygotes for the AMPD1 normal allele. Conversely, there is a positive correlation between these two parameters in homozygotes for the AMPD1 mutant allele. These seemingly divergent observations may be resolved by using available information regarding human AMP deaminase gene and isoform expression in skeletal muscle. AMP deaminase is manifest as a multigene family in humans that produces mRNAs encoding the following parental isoforms: AMPD1, isoform M (myoadenylate deaminase); AMPD2, isoform L; and AMPD3, isoform E. Whereas the AMPD1 gene is expressed predominantly, if not exclusively, in skeletal muscle (1), AMPD2 (1) and AMPD3 (19) are widely expressed, although AMPD3 mRNAs are most abundant in skeletal muscle. A comparative Northern blot demonstrates that AMPD1 mRNAs are more abundant relative to AMPD3 mRNAs in total cellular RNA isolated from adult skeletal muscle (20). Finally, immunocytochemical analysis of adult human skeletal muscle sections has shown greater anti-M reactivity in type II fibers, whereas anti-E reactivity appears confined to type I fibers (37a).

Based on this combined information, the majority of AMP deaminase activity in skeletal muscle of homozygotes and heterozygotes for the AMPD1 normal allele can be attributed to isoform M. Accordingly, >90% of AMP deaminase activity in extracts prepared from normal skeletal muscle is immunoprecipitated by anti-M serum (7). Therefore, the observed negative correlation between skeletal muscle AMP deaminase activity and relative proportion of type I fibers in homozygotes for the AMPD1 normal allele and in heterozygotes is consistent with the relative lower abundance of isoform M in these fibers.

Conversely, homozygotes for the AMPD1 mutant allele exhibit a positive correlation between residual AMP deaminase activity and the proportion of type I fibers. Owing to an alternative splicing event that removes sequence containing the non-sense mutation from 0.6-2% of primary transcripts (21), homozygotes for the AMPD1 mutant allele are expected to produce detectable levels of functional isoform M. However, the observed positive correlation is most likely due to isoform E expression in type I fibers (37a). Isoform E reportedly contributes 18-76% of residual AMP deaminase activity in symptomatic myoadenylate deaminase deficiency (7). Consequently, a positive correlation of residual AMP deaminase activity and the proportion of type I fibers might be anticipated in healthy homozygotes for the AMPD1 mutant allele.

Multiple-regression analysis performed on data derived from the studied individuals shows that the training index also contributes to total AMP deaminase activity. This is in agreement with previous studies, which have demonstrated that both sprint (11) and endurance (27) training can lead to decreased levels of AMP deaminase activity in human skeletal muscle and that a high level of physical activity coincides with a high percentage of type I fibers (13, 36). Therefore, the observed negative correlation between total AMP deaminase activity and the percentage of type I fibers in homozygotes for the normal AMPD1 allele and heterozygotes may also be related to the training status of the studied individuals (well-trained subjects having lower AMP deaminase activity and higher percentage of type I fibers).

Gender is another factor affecting skeletal muscle AMP deaminase activity. On average, 14-18% higher AMP deaminase activities are found in skeletal muscle of men compared with women. This is similar to what has been observed for other enzymes, such as lactate dehydrogenase and phosphofructokinase (8, 15, 36) and suggests a coupling between AMP deaminase activity and glycolytic capacity of human skeletal muscle.

In conclusion, data presented in this study document that variation in skeletal muscle AMP deaminase activity across a healthy population sample is primarily due to a prevalent AMPD1 mutant allele. Additional contributors to variation are fiber-type composition, training status, and gender. These results may impact a body of literature that has focused on IMP accumulation and the accompanying production of ammonia as indicators of altered skeletal muscle energy balance in a variety of physiological settings. Based on several estimations of AMPD1 allele frequencies, including this study, one of four healthy individuals is either a carrier of, or homozygote for, a prevalent mutant allele and thus may have reduced capacities for adenylate catabolism, independent of the training status, gender, or age. It is difficult to predict what effect AMPD1 genotype may have had on previously published results and interpretations. However, the results of this study reveal that current protocols using IMP accumulation and ammonia production as indicators of skeletal muscle energy metabolism would do well to evaluate AMPD1 genotype as a prerequisite to defining their healthy volunteer groups.