Molecular Basis of AMP Deaminase Deficiency in Skeletal Muscle

Home , AMP deaminase

Proc. Natl. Acad. Sci. USA Vol. 89, pp. 6457-6461, July 1992 Medical Sciences Molecular basis of AMP deaminase deficiency in skeletal muscle TAKAYUKI MORISAKI*t, MANFRED GROSS*tt, HIROKO MORISAKI*t, DIETER PONGRATZ§, NEPOMUK ZOLLNERt, AND EDWARD W. HOLMES*t¶ *Departments of Medicine and Biochemistry, Duke University Medical Center, Durham, NC 27706; tDepartments of Medicine and Human Genetics, Seymour Gray Molecular Medicine Laboratory, University of Pennsylvania, Philadelphia, PA 19104-4283; tMedizinische Poliklinik der Universitat Munchen, Munich, Federal Republic of Germany; and §Friedrich-Baur-Institut bei der Medizinischen Klinik der Universitat Munchen, Munich, Federal Republic of Germany Communicated by James B. Wyngaarden, April 9, 1992

ABSTRACT AMP deaminase (AMPD; EC 3.5.4.6) is en- inherited defect in the AMPDJ gene since the enzyme defi- coded by a multigene family in mammals. The AMPDI gene is ciency has been reported in several members of the same expressed at high levels in skeletal muscle, where this enzyme family (13, 20). However, acquired deficiency of AMPD has is thought to play an important role in energy metabolism. also been described (14), raising the possibility that the Deficiency of AMPD activity in skeletal muscle is associated absence of AMPD activity may be secondary to other ab- with symptoms of a metabolic myopathy. Eleven unrelated normalities. To determine the molecular basis for this poten- individuals with AMPD deficiency were studied, and each was tially common abnormality we have studied 11 unrelated shown to be homozygous for a mutant allele characterized by individuals with AMPD deficiency. All of these individuals a C -. T transition at nucleotide 34 (codon 12 in exon 2) and are homozygous for the same mutant allele. In randomly at nucleotide 143 (codon 48 in exon 3). The C -* T transition selected Caucasians and African-Americans we have found at codon 12 results in a nonsense mutation predicting a severely this mutant allele in 13% of 144 alleles tested, but we have not truncated AMPD peptide. Consistent with this prediction, no found this mutant allele in 106 DNA samples from Japanese immunoreactive AMPD1 peptide is detectable in skeletal mus- subjects. cle of these patients. This mutant allele is found in 12% of Caucasians and 19% of African-Americans, whereas none of the 106 Japanese subjects surveyed has this mutant allele. We METHODS AND MATERIALS conclude from these studies that this mutant allele is present at Patients. The index case for these studies is an 18-year-old a sufficiently high frequency to account for the 2% reported German female, who first noted calf pain at 4 years of age, incidence of AMPD deficiency in muscle biopsies. The re- usually related to exercise. Persistence of these symptoms stricted distribution and high frequency of this doubly mutated along with weakness of the upper arms eventually led to a allele suggest it arose in a remote ancestor of individuals of muscle biopsy, which exhibited absence of AMPD activity Western European descent. with normal phosphorylase and phosphofructokinase activ- ities. This patient's muscle biopsy, as well as DNA samples AMP deaminase (AMPD; EC 3.5.4.6), an enzyme that cat- from other family members, was studied in detail. Muscle alyzes deamination of AMP to IMP, and the purine nucleo- biopsies or DNA samples from 10 other unrelated individuals tide cycle, of which AMPD is one component, play a central with AMPD deficiency and variable symptoms (Table 1) were role in purine nucleotide interconversion in eukaryotic cells. also analyzed. All patients were identified and referred to us As a consequence, AMPD activity can be a determinant of because muscle biopsies performed for the indicated symp- adenylate energy charge and energy metabolism in the cell (1, toms (Table 1) were found to have reduced levels of AMPD 2). In mammals, AMPD is encoded by a multigene family (3), activity. DNA samples from 59 Caucasians, 13 African- which accounts in part for the tissue-specific and stage- Americans, and 106 Japanese were obtained from normal specific isoforms of AMPD that have been identified (4, 5). volunteers or from investigators who had no knowledge of The activity ofAMPD in skeletal muscle is -100 times higher the donors' medical history. than that of other organs, a consequence of the high level of Protein Analyses. AMPD activity was quantified either by expression of the AMPDJ gene in this tissue (1, 4, 5). a radiochemical assay or by a spectrophotometric assay (5), Since Fishbein et al. (6) first reported 5 patients with and the method employed is noted in the text or figure legend. AMPD deficiency, >100 patients with this enzyme defect Immunoblots were performed with antiserum raised to have been described (7-21). Several centers have reported AMPD purified from rat skeletal muscle (5) using an en- AMPD deficiency in up to 2% of randomly selected muscle hanced chemiluminescence detection system (Amersham). biopsies (6, 7, 9, 12). A review of reported cases of AMPD Nucleic Acid Analyses. Northern hybridization of RNA deficiency noted that 88% of these individuals with AMPD extracted from skeletal muscle was performed as described deficiency describe exercise-related symptoms, including from this laboratory (3). First-strand cDNA (22) was synthe- muscle aches, cramps, and early fatigue (1). Symptoms are sized from patient-derived RNA samples using an oligonu- variable, however, with some reports of asymptomatic indi- cleotide complementary to bases 2239-2258 (5'-TTGGTT- viduals and descriptions ofother patients who exhibit a range TACTTTTTTTTATTC-3') in this 2.3-kilobase (kb) mRNA of neuromuscular disorders. In the few patients studied in (the sequence of human AMPD is reported in ref. 23). detail (6, 17, 18), the deficiency of AMPD activity has been Single-strand cDNA was amplified by the polymerase chain restricted to skeletal muscle, consistent with high-level ex- reaction (PCR) (24, 25) in two separate reactions; oligonu- pression of the AMPD1 gene being restricted to skeletal cleotides corresponding to bases -20 to -1 (5'-AATCAAG- muscle (3, 4). GATCCCAGCAACA-3') and 881-900 (5'-CACCTTCCTG- The molecular basis for AMPD deficiency is not known, CAGTTATAAA-3') were used to synthesize the 5' region; but in some individuals it is presumed to be the result of an Abbreviation: AMPD, AMP deaminase. The publication costs of this article were defrayed in part by page charge ITo whom reprint requests should be addressed at: Department of payment. This article must therefore be hereby marked "advertisement" Medicine, University of Pennsylvania, 100 Centrex, 3400 Spruce in accordance with 18 U.S.C. §1734 solely to indicate this fact. Street, Philadelphia, PA 19104-4283.

6457 Downloaded by guest on September 30, 2021 6458 Medical Sciences: Morisaki et al. Proc. Natl. Acad Sci. USA 89 (1992) Table 1. AMPD-deficient patients AMPD activity,* Age at onset of Patient Sex Age, years units/g Clinical presentation symptoms, years lt 9 18 5.4 Calf pain and muscle weakness 4 2 9 16 12.5 Progressive weakness and cramps in legs 11 3 25 18.8 Rhabdomyolysis after viral infection 25 4 d 31 13.5 Pain in both legs after exercise 30 5 d 32 8.8 Easy fatigue and pain in both legs 29 6 45 6.2 Pain in both arms after exercise 41 7 9 45 4.4 Pain in arms and shoulders after exercise 33 8 d 51 3.4 Right shoulder pain after exercise 50 9 d 53 6.6 Weakness and pain in legs during exercise 50 10 d 62 8.3 Muscle pain after exercise 55 11 9 68 0 Pain in arms and legs aggravated by exercise 55 *AMPD activity normal range is 60-300 units/g of noncollagen protein. tIndex case. oligonucleotides corresponding to bases 881-900 (5'- To evaluate the significance of the missense mutation at TTATAACTGCAGGAAGGTG-3') and bases 2239-2258 (5'- position 143, wild-type and mutant cDNAs were ligated into TTGGTTTACTTTTTTTTATTC-3') were used to synthe- a prokaryotic expression vector (pKK233-2) and E. coli size the 3' region. This approach was selected to reduce the lysates were assayed for AMPD activity. The 5' terminal 133 probability of errors induced by PCR oflong DNA sequences nucleotides of the mutant cDNA were replaced with the (26) and to facilitate functional analysis of individual muta- wild-type sequence to remove the nonsense mutation at tions that might be found in the 5' and 3' regions of this position 34, resulting in a cDNA that had the single C -+ T transcript. PCR products were subcloned into pBS (Strata- mutation at position 143. Prokaryotes do not exhibit detect- gene) for sequencing by the dideoxynucleotide chain- able AMPD activity and any activity in the lysate from E. coli termination method (27). Thirty independently isolated pBS transformed with the expression vector is presumably de- subclones were pooled for initial sequencing to obviate rived from the plasmid introduced into E. coli. This was PCR-induced errors. Mutations identified by this screening confirmed by immunoprecipitation of the AMPD activity procedure were subsequently confirmed by sequencing this using an antiserum specific for the AMPDJ gene product (5). region in individual clones. The PCR products were also E. coli transformed with the wild-type cDNA exhibit AMPD subcloned into a prokaryotic vector, pKK233-2 (Pharmacia), activity in the range of 2-8 munits/mg of protein, and the for expression in Escherichia coli. This approach permits a AMPD activity in lysates from E. coli transformed with the rapid functional analysis of mutations since prokaryotes do cDNA mutated only at position 143 is not detectably differ- not exhibit AMPD activity. Genomic DNA, isolated from ent. patients and volunteers by standard methods (22), was am- Although the mutation at position 143 appears to have little plified by PCR with oligonucleotides corresponding to intron effect on catalytic activity of AMPD, the nonsense mutation 1 and intron 2 (5'-GCAATCTACATGTGTCTACC-3') and at position 34 in this patient would be expected to give rise to 5'-ATAGCCATGTTTCTGAATTA-3') for evaluation of a severely truncated peptide, only 11 amino acid residues exon 2; oligonucleotides corresponding to intron 2 and intron compared to 747 residues in the control. Muscle extract from 3 (5'-AGCAGAAACTCTCAGCTGAC-3' and 5'-CTCCT- this patient has no detectable immunoreactive AMPD using TAGGTGCCAATATAC-3') were used for evaluation of polyclonal antiserum, whereas an easily detectable 86-kDa exon 3. These PCR products were analyzed by direct se- band is visualized in muscle extract from the normal control quencing (28), hybridization with allele-specific oligonucle- using this antiserum (Fig. 3). Deletion and point mutation otides (29), and/or restriction analysis as described in the analyses of human AMPD1 have shown that the active site text. for catalysis in this enzyme is located 3' to nucleotide 531, codon 177 (unpublished observation). Thus, the nonsense mutation at position 34 would be expected to result in RESULTS complete loss of AMPD activity, and the absence of immu- Evaluation of Index Family. AMPD1 transcript size is not noreactive protein can also be explained by this mutation. detectably altered in skeletal muscle from the index patient, Genomic DNA from the index patient's mother, father, and and transcript abundance is not reduced (Fig. 1). To the brother, all of whom are asymptomatic, was also sequenced contrary, the AMPD1 transcript, when normalized to creatine kinase M abundance, is approximately three times Ct Pt greater in this patient than a normal control. Studies of a heterozygote for AMPD deficiency also demonstrate that the --m mutant AMPD1 transcript is present in greater abundance AMPDDI than the wild-type transcript (unpublished observations). Comparison ofthe cDNA sequence for the index patient to for a normal control reveals only two nucleotide differ- that - ences (Fig. 2). Numbering from the first in-frame AUG, the CK-M presumptive translation start site, nucleotide 34 is a T in the patient and it is a C in the control; nucleotide 143 is a T in the FIG. 1. AMPD1 mRNA abundance in muscle. Four micrograms patient and it is a C in the control. The deviation in sequence of total RNA from control muscle (Ct) and muscle from the index of the patient cDNA at position 34 (exon 2) changes this patient (Pt) was resolved on a 1% agarose gel containing formalde- codon from CAA to TAA-i.e., a stop codon instead of hyde, transferred to Nytran paper (Schleicher & Schuell), and glutamine. The sequence difference at position 143 (exon 3) hybridized with a human AMPD1 cDNA probe (AMPD1) (23). After in the patient predicts an amino acid substitution of a leucine washing the filter, the same filter was used for hybridization with a for proline. human creatine kinase M (CK-M) cDNA probe (30). Downloaded by guest on September 30, 2021 Medical Sciences: Morisaki et al. Proc. Natl. Acad. Sci. USA 89 (1992) 6459

A B Exon 2 Exon 2 Exon 3 CONTROL PATIENT 3, G GATC GATC T A G T A "/ G T T A -A- A A Stop A I A 34 Cl A A

A A IG T Exon 3 \G CONTROL PATIE N T A GATC G A T C

5, .t ws'

FIG. 2. Nucleotide sequence of cDNA and gene for AMPD1. (A) Sequence of AMPD1 cDNA from a control and the index patient. The left

panel shows the mutation in exon 2 (C-. T at nucleotide 34) that results in a nonsense mutation at codon 12 (Gln -. Stop). The right panel shows

the mutation in exon 3 (C -* T at nucleotide 143) that results in a missense mutation at codon 48 (Pro -* Leu). (B) AMPD1 sequence of PCR-amplified genomic DNA. The upper panel shows the base substitution (C -* T at nucleotide 34) in exon 2. The lower panel shows the base

substitution (C -. T at nucleotide 143) in exon

in the relevant regions of exon 2 and exon 3 (Fig. 4). The that the C -* T transitions at positions 34 and 143 are present AMPDJ genes for all three of these individuals contain both on the same allele. cytosine and thymidine nucleotides at positions 34 and 143, Population Studies. Genomic DNA obtained from 59 Cau- indicating they are heterozygotes. casians, 13 African-Americans, and 106 Japanese was ana- Studies of Other Patients with AMPD Deficiency. Muscle lyzed by two techniques for the nonsense mutation at posi- from 1 of the 10 other unrelated individuals with AMPD tion 34 (Fig. 5). PCR-amplified DNA from all subjects was deficiency was used to prepare cDNA as in the index case and immobilized on Nytran filters and hybridized to a wild-type this cDNA was sequenced in its entirety. Genomic DNA or a mutant oligonucleotide for distinguishing the two types from all 11 AMPD-deficient subjects was sequenced in the of alleles (29). PCR-amplified DNA was also restricted with regions of exon 2 and 3. The only deviations from the control Mae II in 31 subjects. This restriction enzyme recognizes the

sequence are C -. T transitions at positions 34 and 143 (Table sequence ACGT, found only once in this region ofthe normal 2). The AMPD activity levels in skeletal muscle and the AMPD1 gene. The mutation at position 34 alters the sequence symptoms in each of these individuals are listed in Table 1. of this restriction site, and the PCR product from the mutant An immunoblot of muscle lysate from one of these individ- allele (AIGT) is not a substrate for this restriction endonu- uals confirmed the absence of immunoreactive AMPD pep- clease. Both tests give identical results in all cases. Seventeen tide in this patient. percent of Caucasians and 23% of African-Americans are Muscle was also available from one heterozygous individ- heterozygous for the nonsense mutation in exon 2, whereas ual and cDNA was prepared by reverse transcription PCR as none of the Japanese examined have this mutant allele. In described above. Sequencing of individual subclones of the addition, two Caucasians and one African-American were PCR products prepared from this individual demonstrated found to be homozygous for this mutation. Each of these individuals with the exon 2 nonsense mutation also has a C Ct Pt -- T transition at position 143 of exon 3. Not one of the 34 chromosomes analyzed with a C at position 34 has a T at position 143. Thus, in the 74 chromosomes examined, there is no evidence for disconcordance between a C at positions 34 and 143 or a T at both of these positions. AMPD -- DISCUSSION Eleven unrelated individuals with AMPD deficiency were evaluated in this and both alleles of the AMPDJ gene _ study, have the same two mutations in all 11 individuals-i.e., a C -+ T transition at positions 34 and 143. The C -- T transition FIG. 3. Immunoreactive AMPD peptide in muscle lysate. Thirty at position 34, in exon 2, results in a nonsense mutation that micrograms of protein from muscle lysate of a control (Ct) or the a truncated peptide as the product of this index patient (Pt) was resolved on an 8% SDS polyacrylamide gel, predicts severely transferred to nitrocellulose paper, and incubated with polyclonal mutant transcript. This truncated peptide terminates prior to antiserum raised to rat AMPD1 (5). This antiserum does not react the synthesis of the catalytic domain of AMPD, providing an with other AMPD isoforms (5). The 86-kDa AMPD1 peptide in explanation for the significant decrease of this enzyme ac- control lysate is indicated. tivity in these individuals. Either this truncated peptide is not Downloaded by guest on September 30, 2021 6460 Medical Sciences: Morisaki et al. Proc. Natl. Acad. Sci. USA 89 (1992)

lF\(oll 2 34I)II1 i t11I I I *I 4 *1' 1'34 ( T3'e3 ( ('34A I11. 1 RU() IId1.i)}.IL ()

i It (. ( A T C ()1t1g I

Ofigo * * %,~~~~~~~~~~~~~~~~~~I*RId awo'X'\i- |~~~~~i.t i omi oimtroHX iom I~~~~~~~~~~tl B T1.34 (:1-34 (1:434 F _ \~~~~~~~~~~~~~.k

.. A - 198 hp

\'l~~~~~~~~~~~~~~~~~~~~~~~~~~ -- 1I 1)1) FAMILY G - 87 hip

FIG. 5. Screening method for detecting the mutation at nucleotide 34. (A) Allele-specific oligonucleotides. PCR-amplified genomic DNA from the region ofexon 2 ofAMPD1 was fixed to Nytran paper and hybridized to a radiolabeled wild-type oligonucleotide (Oligo 1; 5'-ATACTCAC-jTTTCTCTTCAG-3') or a radiolabeled mutant oligonucleotide (Oligo 2; 5'-ATACTCACATTTCTCTTCAG-3'). DNA from an individual homozygous for a C at nucleotide 34 (Homo C/C) hybridizes only to oligonucleotide 1, DNA from an individual homozygous for a T at nucleotide 34 (Homo T/T) hybridizes only to oligonucleotide 2, and DNA from a heterozygote (Hetero C/T) FIG. 4. AMPD1 sequence in family members of index case. hybridizes to both oligonucleotides. (B) Restriction endonuclease (Upper) The sequence of PCR-amplified genomic DNA in the region mapping. PCR-amplified genomic DNA from the region of exon 2 of exon 2 (nucleotide 34) is illustrated on the left and the sequence was digested with Mae II. This enzyme recognizes the sequence in the region of exon 3 (nucleotide 143) is illustrated on the right. ACGT that occurs once in the wild-type PCR fragment, but it is Positions 34 and 143 both exhibit the presence of a C and T at these absent from the mutant PCR fragment as consequence of the C -. T sites, confirming this individual (the index patient's mother) is a transition in exon 2. The undigested PCR fragment is 198 nucleotides heterozygote. (Lower) The genotype of each family member is in length; the digested PCR fragments are 111 and 87 nucleotides in depicted in this pedigree. The index patient is shown by the alrow. length. Abbreviations for homozygote and heterozygote are the same as in A. bp, Base pairs. recognized by the available antiserum or it is labile, explain- ing the absence of a discernible AMPD signal in immunoblots tors (14, 31). We assume as they do that most of this residual performed with muscle extracts from these patients. Our activity reflects AMPD produced in nonmyocytes in the immunological observations in AMPD-deficient patients are muscle tissue by one of the other AMPD genes since this similar to those reported in prior studies by Fishbein (14) and activity is not reactive with antisera specific for the AMPD1 Sabina et al. (31). The variable residual AMPD activity found gene product (14, 31). However, we cannot exclude the in our patients is similar to that reported by other investiga- possibility that a small fraction of the residual activity is produced in myocytes from the mutant AMPDJ gene as a Table 2. Population study consequence of alternative splicing (see below). Nucleotide 34 Nucleotide 143 The C -b T transition at nucleotide 143 in the AMPD1 Group transcript of these patients is apparently a silent mutation Control C C based on studies performed with recombinant peptides pro- AMPD deficient T (11) T (11) duced in a prokaryotic expression system. Although detailed Caucasian C (47) C (14) kinetic studies have not been carried out with the wild-type C/T (10) C/T (4) and mutant peptides, which differ by only a proline or a T (2) T (1) leucine at codon 48, these two enzymes have comparable African-American C (9) C (1) activity under the assay conditions employed. The apparent C/T (3) C/T (3) normal catalytic activity exhibited by the AMPD peptide T (1) T (1) harboring a mutation in codon 48 could assume clinical Japanese C (106) C (2) importance, iffuture studies demonstrate the human AMPD1 The AMPD-deficient patients are described in Table 1. The DNA transcript is subject to alternative splicing, which deletes samples from 59 Caucasians, 13 African-Americans, and 106 Japa- exon 2. In rat, exon 2 is deleted from the majority of nese were obtained from normal volunteers or from investigators transcripts produced from the AMPDJ gene in embryonic who had no knowledge of the donors' medical history. The number muscle (32) and in response to changes in neural innervation in parentheses is the number of individuals for whom the DNA of skeletal muscle (unpublished observations). Since the sequence was determined at the indicated nucleotide on PCR sam- of the AMPDJ has been ples prepared from genomic DNA as described in the legend of Fig. primary sequence gene highly 5. Every DNA sample was screened for a C or T at nucleotide 34 conserved in man and rat (23), a similar pattern ofalternative through allele-specific oligonucleotide hybridization and confirmed splicing may occur in embryonic human muscle or in re- by restriction digestion and/or direct sequencing. A limited number sponse to external signals. Alternative splicing of exon 2 in of DNA samples were screened for a C or T at nucleotide 143 since the human AMPD1 transcript would delete the nonsense this required direct sequencing of the PCR product. mutation specified by the C -* T transition at nucleotide 34. Downloaded by guest on September 30, 2021 Medical Sciences: Morisaki et al. Proc. Natl. Acad. Sci. USA 89 (1992) 6461

Although we have not detected alternative splicing in adult Beaudet, A. L., Sly, W. S. & Valle, D. (McGraw-Hill, New human skeletal muscle (33), we cannot exclude at this time York), Vol. 2, pp. 1077-1084. the possibility that it may occur at an early stage of skeletal 2. Sabina, R. L., Swain, J. L., Olanow, C. W., Bladley, W. G., muscle development, thereby ameliorating the consequences Fishbein, W. N., DiMauro, S. & Holmes, E. W. (1984) J. Clin. Invest. 73, 720-730. of the nonsense mutation in exon 2. 3. Morisaki, T., Sabina, R. L. & Holmes, E. W. (1990) J. Biol. The absence of a C -- T transition at nucleotide 143 in the Chem. 265, 11482-11486. 34 normal chromosomes examined makes it unlikely that this 4. Ogasawara, N., Goto, H., Yamada, Y. & Watanabe, T. (1978) mutation is a common polymorphism in these populations. Eur. J. Biochem. 87, 297-304. This fact, coupled with the 100%6 concordance of C -- T 5. Marquetant, R., Desai, N. M., Sabina, R. L. & Holmes, E. W. transitions at nucleotides 34 and 143 in 40 chromosomes (1987) Proc. Nati. Acad. Sci. USA 84, 2345-2349. analyzed from AMPD-deficient subjects and heterozygotes, 6. Fishbein, W. N., Armbrustmacher, V. W. & Griffin, J. L. suggests one of two explanations. Either there is a significant (1978) Science 200, 545-548. 7. Shumate, J. B., Katnik, R., Ruiz, M., Kaiser, K., Frieden, C., rate of spontaneous and coordinated mutation at both of Brooke, M. H. & Carroll, J. E. (1979) Muscle Nerve 2, 213- these positions or the doubly mutated allele is present at a 216. relatively high frequency in some populations. The latter 8. Scholte, H. R., Busch, H. F. N. & Luyt-Houwen, E. M. explanation seems more plausible since it is difficult to (1981) J. Inher. Metab. Dis. 4, 169-170. envision a mechanism that results in spontaneous mutations 9. Kar, N. C. & Pearson, C. M. (1981)Arch. Neurol. 38, 279-281. affecting the same two nucleotides in all of these alleles. 10. Mercelis, R., Martin, J. J., Dehaene, I., de Barsy, Th. & Van Furthermore, this mechanism would have to be restricted to den Berghe, G. (1981) J. Neurol. 225, 157-166. 11. Hayes, D. J., Summers, B. A. & Morgan-Hughes, J. A. (1982) certain populations since the doubly mutated allele has not J. Neurol. Sci. 53, 125-136. been observed in the Japanese population. A more plausible 12. Kelemen, J., Rice, D. R., Bradley, W. G., Munsat, T. L., explanation in our opinion is that the doubly mutated allele DiMauro, S. & Hogan, E. L. (1982) Neurology 32, 857-863. arose at some time in the remote past, and it has become 13. Gertler, P. A. & Jacobs, R. P. (1984) Arthritis Rheum. 27, widely disseminated in individuals of Western European 586-590. descent. Studies are necessary to compare the frequency of 14. Fishbein, W. N. (1985) Biochem. Med. 33, 158-169. this mutant allele in different ethnic groups in Western 15. Lally, E. V., Frieden, J. H. & Kaplan, S. R. (1985) Arthritis to Rheum. 28, 1298-1302. Europe, as well as African-Americans and native Africans, 16. Goebel, H. H., Bardosi, A., Conrad, B., Kuhlendahl, H. D., gain additional insight into the origin of this allele. The high DiMauro, S. & Rumpf, K. W. (1986) Klin. Wochenschr. 64, frequency of this allele in some populations may prove useful 342-347. in studying the relationship between different ethnic groups. 17. DiMauro, S., Miranda, A. F., Hays, A. P., Franck, W. A., Recognizing that the frequency ofAMPD heterozygosity is Hoffman, G. S., Schoenfeldt, R. S. & Singh, N. (1980) J. -20% in Caucasians and African-Americans, it is not sur- Neurol. Sci. 47, 191-202. prising that several centers have reported 1-3% of randomly 18. Fishbein, W. N., David, J. I., Nagarajan, K., Winkert, J. W. & enzyme activity Foellmer, J. W. (1980) Arch. Biochem. Biophys. 205, 360-364. sampled muscle biopsies are deficient in this 19. Heller, S. L., Kaiser, K. K., Planer, G. J., Hagberg, J. M. & (6, 7, 9, 12). Clearly 1-3% of individuals in these populations Brooke, M. H. (1987) Neurology 37, 1039-1042. do not have symptoms that are clinically severe enough to be 20. Sinkeller, S. P. T., Joosten, E. M. G., Wevers, R. A., Oei, classified as a metabolic myopathy. Thus, the frequency of T. L., Jacobs, A. E. M., Veerkamp, J. H. & Hamel, B. C. J. this mutant allele in these populations raises a number of (1988) Muscle Nerve 11, 312-317. questions about the clinical implications of this form of 21. Kaletha, K. & Nowak, G. (1990) Clin. Chim. Acta 190,147-156. AMPD deficiency. One could conclude this mutant allele is 22. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular a harmless polymorphism. On the other hand, this mutation Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold could be for in some tissues or at some stages Spring Harbor, NY), 2nd Ed. compensated 23. Sabina, R. L., Morisaki, T., Clarke, P., Eddy, R., Shows, of development by removal of exon 2 through alternative T. B., Morton, C. C. & Holmes, E. W. (1990) J. Biol. Chem. splicing. The residual AMPD activity observed in muscle 265, 9423-9433. tissue of these patients, presumably a product of another 24. Saiki, R. K., Scharf, S., Faloona, F., Mullis, K. B., Horn, member of this multigene family (3, 14, 31), might also G. T., Erlich, H. A. & Arnheim, N. (1985) Science 230, 487- compensate for the mutation in the AMPDJ gene, especially 491. if this activity were present in myocytes. Other interpreta- 25. Veres, G., Gibbs, R. A., Schere, S. E. & Caskey, C. T. (1987) tions include the possibility that clinically significant myo- Science 237, 415-417. pathic symptoms develop only in individuals who have 26. Saiki, R. K., Gelfand, D. H., Stoffel, S. J., Higuchi, R., Horn, or abnormality in energy metab- G. T., Mullis, K. B. & Erlich, H. A. (1988) Science 239, another inherited acquired 487-491. olism. Answers to these questions will require additional 27. Sanger, F., Coulson, A. R., Garrell, B. G., Smith, A. J. H. & physiological and molecular studies in controls and individ- Roe, B. A. (1980) J. Mol. Biol. 143, 161-178. uals with this mutant allele. 28. Wong, C., Dowling, C. E., Saiki, R. K., Higuchi, R. G., Er- lich, H. A. & Kazazian, H. H. Jr. (1987) Nature (London) 330, T.M., M.G., and H.M. contributed equally to this report. We 384-386. thank Dr. Ingrid Paetzke (Klinische Chemie Stddt Krankenhaus 29. Saiki, R. K., Bugaman, T. L., Horn, G. T., Mullis, K. B. & Munchen-Schwabing), who kindly provided the results on AMPD Erlich, H. A. (1986) Nature (London) 324, 163-166. activity of the muscle biopsies of the 11 patients included in this 30. Perryman, M. B., Kerner, S. A., Bohlmeyer, T. J. & Roberts, study, and Dr. Hisaichi Fujii (Tokyo Women's Medical College), R. (1986) Biochem. Biophys. Res. Commun. 140, 981-989. who generously supplied the DNA samples for the Japanese subjects. 31. Sabina, R. L., Fishbein, W. N., Pezeshkpour, G., Clarke, This work was supported by grants to E.W.H. (DK-12413, National P. R. H. & Holmes, E. W. (1992) Neurology 42, 170-179. Institutes of Health) and M.G. (Deutsche Forschungsgemeinschaft). 32. Sabina, R. L., Ogasawara, N. & Holmes, E. W. (1989) Mol. Cell. Biol. 9, 2244-2246. 1. Sabina, R. L., Swain, J. L. & Holmes, E. W. (1989) in The 33. Mineo, I., Clarke, P., Sabina, R. L. & Holmes, E. W. (1990) Metabolic Basis of Inherited Disease, eds. Scriver, C. R., Mol. Cell. Biol. 10, 5271-5278. Downloaded by guest on September 30, 2021