There are several isoenzymes of AMP deaminase in mammalian tissues. Different tissues contain varying proportions of these isoenzymes, and the kinetic, physical, and immunologic properties of these isoenzymes are distinct. In humans, four isoenzymes of AMP deaminase are known.72 Isoenzyme M, or myoadenylate deaminase, is found only in skeletal muscle; isoenzyme L is the predominant form in liver and brain; isoenzymes E1 and E2 are found in erythrocytes. Tissues such as heart, kidney, and spleen contain isoenzymes L, E1, and E2. Consequently, skeletal muscle is the only affected tissue in myoadenylate deaminase-deficient patients.9,13,73 At least three isoenzymes of AMP deaminase have been identified in the rat74,75 and, as in humans, one is found exclusively in skeletal muscle. Molecular studies have demonstrated that all the mammalian myoadenylate deaminase isoenzymes are encoded by transcripts produced from different genes (see “AMP Deaminase Genes and Transcripts,” below).
During the course of development, the total amount of AMP deaminase activity in a given tissue may vary, as may the relative distribution of isoenzyme types.76– 83 Human skeletal muscle exhibits a greater than fivefold increase in total AMP deaminase activity from the 11-week-old fetus to the adult.81 In rat skeletal muscle, the observed increase from birth is greater than eightfold.82 In both cases, there is a switch in AMP deaminase isoforms produced in skeletal muscle at different stages of development. Moreover, adult skeletal muscle fiber types exhibit differential AMP deaminase activities and patterns of isoenzyme expression. For example, rat skeletal muscles composed predominantly of fast-twitch white (type IIb) fibers contain three- to tenfold more AMP deaminase activity84 than do those with a majority of slow-twitch red (type I) fibers. Adult human skeletal muscle exhibits predominant anti-isoform M and anti-isoform E sera reactivity in type II and type I fibers, respectively.85 In exercising human skeletal muscle, calculated AMP deaminase activity86 and NH3 production87 are directly related to the percentage of fast-twitch fibers in postexercise biopsy. A shift in isoenzyme type has also been observed in human,88 rat,83 and chick77 skeletal myocytes grown in culture, which appears to explain the findings of AMP deaminase activity in primary myoblasts grown in culture from muscle biopsy specimens of patients with well-documented myoadenylate deaminase deficiency.13,88
Myoadenylate deaminase (isoform M) has been purified to homogeneity from many mammalian sources, including humans.72,89– 91 The native molecular mass of the enzyme from human skeletal muscle is approximately 300 kDa,72,91 and following treatment with denaturing agents, such as guanidine hydrochloride or sodium dodecylsulfate, the enzyme dissociates into a single polypeptide with a reported molecular mass of 71 to 72 kDa.72,91 However, these and many other purifications of mammalian skeletal muscle AMP deaminase were performed in the absence of protease inhibitors that can yield proteolyzed polypeptides lacking up to nearly 100 N-terminal amino acid residues.92 Purification of rat skeletal muscle AMP deaminase performed in the presence of protease inhibitors yielded a subunit molecular mass of 80 kDa.93 This latter estimation is closer to the calculated molecular mass of 86 to 87 kDa as predicted from available rat94 and human45 myoadenylate deaminase cDNA sequences. Nevertheless, all data are consistent with the conclusion that AMP deaminase from skeletal muscle is composed of four polypeptide chains, and tryptic maps suggest that these subunits are identical. Protein kinase C-mediated phosphorylation of rat skeletal muscle myoadenylate deaminase has been reported.95 Other post-translational modifications, such as glycosylation or fatty acylation of skeletal muscle myoadenylate deaminase, have not been investigated.
Myoadenylate deaminase is closely associated with the contractile apparatus in skeletal myocytes. Histochemical studies have demonstrated that myoadenylate deaminase is bound to the myofibril in the region of the A band 96 and that 2 mol of native enzyme bind to 1 mol of myosin.97 Myoadenylate deaminase binds to a specific region of rabbit myosin heavy chain, that is, heavy meromyosin or subfragment 2, and an association between myoadenylate deaminase and myosin appears to have a functional significance in controlling this enzymatic activity (see below). In vitro, the rat 80-kDa muscle myoadenylate deaminase peptide binds much more tightly to myosin than does the 66-kDa proteolyzed fragment.93
In vitro and in vivo studies suggest that myoadenylate deaminase exhibits a striking increase in activity in skeletal muscle during exercise (see “Purine Nucleotide Cycle,” above). Reconstitution experiments that attempt to mimic conditions found in resting muscle in vivo suggest that myoadenylate deaminase is inhibited by as much as 80 to 90 percent under these conditions, whereas conditions that mimic contracting muscle are associated with marked increases in activity and flux through the purine nucleotide cycle.60 Many factors that affect myoadenylate deaminase activity in vitro have been identified: adenylate energy charge; ratio of purine nucleoside triphosphates to diphosphates to monophosphates; concentration of K+, H+, Pi, and creatine phosphate; and binding to myosin.60,97– 100 Changes in one or more of these variables are thought to lead to the release of inhibition of myoadenylate deaminase and the increase in activity of this enzyme during exercise.
It is not established how each of the above factors influences myoadenylate deaminase activity, but the data obtained by Ashby and Frieden99 from kinetic and binding studies led them to propose a model that explains many of the regulatory properties of this enzyme. They suggest that the enzyme has three distinct types of purine nucleotide binding sites: a catalytic site that binds AMP, an inhibitory site that binds purine nucleoside triphosphates, and a stimulatory site that binds all types of purine nucleotides, but diphosphates in preference to monophosphates, and monophosphates in preference to triphosphates. K+ ions affect the activity of the enzyme through cooperative effects on the catalytic site.
Nucleoside triphosphates (ATP and GTP) bind avidly to the inhibitory site and produce effects on cooperativity at the catalytic site. Nucleoside diphosphates and monophosphates (ADP, GDP, AMP, and GMP) bind avidly to the stimulatory site and indirectly decrease the affinity of the inhibitory site for nucleoside triphosphates. IMP, a product of the reaction, also binds to the stimulatory site and may, under some conditions, lead to enzyme activation. In resting muscle, purine nucleoside triphosphates are present in considerable excess relative to diphosphates and monophosphates. Nucleoside triphosphate content decreases following vigorous work, producing a lower concentration of ligands for binding at the inhibitory site. Nucleoside diphosphate and monophosphate content increases, leading to higher ligand concentrations for binding at the stimulatory site and secondarily decreasing the affinity of the inhibitory site for nucleoside triphosphates. AMP content increases, providing more substrate, and probably enhanced enzyme activity, by binding of this monophosphate at the stimulatory site.
This model accommodates a close correlation between the increase in myoadenylate deaminase activity and the drop in adenylate energy charge that occurs with exercise.61,101 The drop in myocyte pH that follows exercise also contributes to the increase in myoadenylate deaminase activity, especially in fast-twitch glycolytic fibers.102 The pH optimum of myoadenylate deaminase (i.e., pH 6.5) reflects the role of metabolic acidosis in AMP deaminase activation.103 Changes in creatine phosphate concentration probably do not play a role in controlling the activity of this enzyme.104 The association of myoadenylate deaminase with myosin, on the other hand, represents a physical interaction with potential physiological significance. In vitro binding studies indicate an association of the myosin components with the nucleoside triphosphate, or inhibitory site, of AMP deaminase.97
Studies performed with rat skeletal muscle have provided in vivo evidence for the significance of the myofibrillar binding of myoadenylate deaminase. In resting muscle or during contractions, when energy balance is well maintained, nearly all AMP deaminase is found in the cytosol. During intense contraction, however, conditions favor binding, which increases in a first-order manner and precedes initiation of IMP formation.105 Furthermore, myosin-bound AMP deaminase exhibits bimodal kinetics, including a component that allows for a higher rate of AMP deamination at the low [AMP] found physiologically.106 Together, these data imply that the myofibrillar association of myoadenylate deaminase is critical to defining its role in skeletal muscle function. Studies performed with human skeletal muscle, however, have observed higher percentages of bound AMP deaminase in resting skeletal muscle that do not increase during high-intensity exercise.107,108
Isoenzyme L of AMP deaminase has been purified to apparent homogeneity from platelets109 and autopsy liver.72 The native molecular mass, determined on the latter, is also approximately 300 kDa. Under denaturing conditions, however, the preparation from platelets exhibited a subunit molecular mass of 83 to 85 kDa, whereas that from liver was 68 kDa. Isoenzyme L cDNA sequence110 would predict that the former estimation is the more accurate and that the latter may reflect proteolysis during preparation. Among all the human AMP deaminase variants, isoenzyme L exhibits the unique property of allosteric activation by ATP in the presence of 150 mM KCl.72
Isoenzymes E1 and E2 display similar kinetic and immunologic characteristics. Deficiency of AMP deaminase activity restricted to erythrocytes has been described in several asymptomatic families in Japan111,112 and is characterized by a lack of both E1 and E2.111 The combined information about the E isoenzymes has been used to hypothesize that both are the product of the same gene and that the relatively less abundant E2 is merely a proteolytic, or otherwise modified, derivative of E1.111 Accordingly, an AMPD3 mutant allele has now been identified in Japanese individuals who have a deficiency of erythrocyte AMP deaminase.113 Anti-isoenzyme E1 serum reactivity in adult human skeletal muscle sections appears confined primarily to type I fibers,85 and a similarly reactive component is the major residual activity in extracts prepared from skeletal muscle of patients with inherited myoadenylate deaminase deficiency.114 Isoenzyme E1, purified to apparent homogeneity from human erythrocytes, also exhibits a native molecular mass of approximately 300 kDa, which is converted to 80 kDa under denaturing conditions.72,115 Available cDNA sequence predicts subunit molecular masses ranging from 88 to 90 kDa for alternative forms of isoenzyme E1.116
Availability of cDNA sequences for three of the human AMP deaminase isoenzymes45,110,116 allows for alignments of their predicted primary amino acid sequences.110,116 The C-terminal domains, encompassing approximately two-thirds of the polypeptide chains, exhibit 62 to 73 percent amino acid identity. Included in this region is the AMP deaminase signature sequence found in all AMP deaminases in organisms ranging from yeast to humans, SLSTDDP.45,94,110,116– 120 This motif also is highly conserved (i.e., S[I/L]NTDDP) in adenosine deaminases in organisms ranging from bacteria to humans.121 Significantly, x-ray crystallographic analysis of murine adenosine deaminase has revealed that the consecutive aspartate residues in this motif are involved in the catalytic site of the enzyme.122 By analogy, a catalytic role has been proposed for the conserved aspartate residues in AMP deaminase,121 and site-directed mutagenesis that replaces an aspartate residue leads to complete loss of enzymatic activity.123 Unlike C-terminal domains, the N-terminal third of different human isoenzymes of AMP deaminase exhibits divergent sequences. Several gaps must be inserted into these alignments to generate any similarities.110,116 However, alignments between rat and human sequences reveal high identities along the entire AMPD1121 and AMPD3121 predicted polypeptides, respectively. These data suggest a functional significance for isoenzyme-specific N-terminal sequences of AMP deaminase. Identified roles for divergent N-terminal domains include contributions to isoform-specific kinetic properties,124 protein-protein interaction,92 and post-translational modification (unpublished data).