Molecular basis of AMP deaminase deficiency in skeletal muscle - NCBI

6 downloads 0 Views 1MB Size Report
Brooke, M. H. & Carroll, J. E. (1979) Muscle Nerve 2, 213-. 216. 8. Scholte, H. R. ... DiMauro, S. & Hogan, E. L. (1982) Neurology 32, 857-863. 13. Gertler, P. A. ...
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

inherited defect in the AMPDJ gene since the enzyme deficiency has been reported in several members of the same family (13, 20). However, acquired deficiency of AMPD has also been described (14), raising the possibility that the absence of AMPD activity may be secondary to other abnormalities. To determine the molecular basis for this potentially common abnormality we have studied 11 unrelated individuals with AMPD deficiency. All of these individuals are homozygous for the same mutant allele. In randomly selected Caucasians and African-Americans we have found this mutant allele in 13% of 144 alleles tested, but we have not found this mutant allele in 106 DNA samples from Japanese subjects.

AMP deaminase (AMPD; EC 3.5.4.6) is enABSTRACT coded by a multigene family in mammals. The AMPDI gene is expressed at high levels in skeletal muscle, where this enzyme is thought to play an important role in energy metabolism. Deficiency of AMPD activity in skeletal muscle is associated with symptoms of a metabolic myopathy. Eleven unrelated individuals with AMPD deficiency were studied, and each was shown to be homozygous for a mutant allele characterized by a C -. T transition at nucleotide 34 (codon 12 in exon 2) and at nucleotide 143 (codon 48 in exon 3). The C -* T transition at codon 12 results in a nonsense mutation predicting a severely truncated AMPD peptide. Consistent with this prediction, no immunoreactive AMPD1 peptide is detectable in skeletal muscle 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 conclude from these studies that this mutant allele is present at a sufficiently high frequency to account for the 2% reported incidence of AMPD deficiency in muscle biopsies. The restricted distribution and high frequency of this doubly mutated allele suggest it arose in a remote ancestor of individuals of Western European descent.

METHODS AND MATERIALS Patients. The index case for these studies is an 18-year-old German female, who first noted calf pain at 4 years of age, usually related to exercise. Persistence of these symptoms along with weakness of the upper arms eventually led to a muscle biopsy, which exhibited absence of AMPD activity with normal phosphorylase and phosphofructokinase activities. This patient's muscle biopsy, as well as DNA samples from other family members, was studied in detail. Muscle biopsies or DNA samples from 10 other unrelated individuals with AMPD deficiency and variable symptoms (Table 1) were also analyzed. All patients were identified and referred to us because muscle biopsies performed for the indicated symptoms (Table 1) were found to have reduced levels of AMPD activity. DNA samples from 59 Caucasians, 13 AfricanAmericans, and 106 Japanese were obtained from normal volunteers or from investigators who had no knowledge of the donors' medical history. Protein Analyses. AMPD activity was quantified either by a radiochemical assay or by a spectrophotometric assay (5), and the method employed is noted in the text or figure legend. Immunoblots were performed with antiserum raised to AMPD purified from rat skeletal muscle (5) using an enhanced chemiluminescence detection system (Amersham). Nucleic Acid Analyses. Northern hybridization of RNA extracted from skeletal muscle was performed as described from this laboratory (3). First-strand cDNA (22) was synthesized from patient-derived RNA samples using an oligonucleotide complementary to bases 2239-2258 (5'-TTGGTTTACTTTTTTTTATTC-3') in this 2.3-kilobase (kb) mRNA (the sequence of human AMPD is reported in ref. 23). Single-strand cDNA was amplified by the polymerase chain reaction (PCR) (24, 25) in two separate reactions; oligonucleotides corresponding to bases -20 to -1 (5'-AATCAAGGATCCCAGCAACA-3') and 881-900 (5'-CACCTTCCTGCAGTTATAAA-3') were used to synthesize the 5' region;

AMP deaminase (AMPD; EC 3.5.4.6), an enzyme that catalyzes deamination of AMP to IMP, and the purine nucleotide cycle, of which AMPD is one component, play a central role in purine nucleotide interconversion in eukaryotic cells. As a consequence, AMPD activity can be a determinant of adenylate energy charge and energy metabolism in the cell (1, 2). In mammals, AMPD is encoded by a multigene family (3), which accounts in part for the tissue-specific and stagespecific isoforms of AMPD that have been identified (4, 5). The activity of AMPD in skeletal muscle is -100 times higher than that of other organs, a consequence of the high level of expression of the AMPDJ gene in this tissue (1, 4, 5). Since Fishbein et al. (6) first reported 5 patients with AMPD deficiency, >100 patients with this enzyme defect have been described (7-21). Several centers have reported AMPD deficiency in up to 2% of randomly selected muscle biopsies (6, 7, 9, 12). A review of reported cases of AMPD deficiency noted that 88% of these individuals with AMPD deficiency describe exercise-related symptoms, including muscle aches, cramps, and early fatigue (1). Symptoms are variable, however, with some reports of asymptomatic individuals and descriptions of other patients who exhibit a range of neuromuscular disorders. In the few patients studied in detail (6, 17, 18), the deficiency of AMPD activity has been restricted to skeletal muscle, consistent with high-level expression of the AMPD1 gene being restricted to skeletal muscle (3, 4). The molecular basis for AMPD deficiency is not known, but in some individuals it is presumed to be the result of an

Abbreviation: AMPD, AMP deaminase. ITo whom reprint requests should be addressed at: Department of Medicine, University of Pennsylvania, 100 Centrex, 3400 Spruce Street, Philadelphia, PA 19104-4283.

The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 6457

6458

Proc. Natl. Acad Sci. USA 89 (1992)

Medical Sciences: Morisaki et al.

Table 1. AMPD-deficient patients AMPD activity,* Clinical presentation units/g Patient Sex Age, years 9 5.4 Calf pain and muscle weakness 18 lt Progressive weakness and cramps in legs 9 12.5 2 16 25 18.8 3 Rhabdomyolysis after viral infection 4 Pain in both legs after exercise 31 13.5 d 32 8.8 5 Easy fatigue and pain in both legs d 45 6.2 Pain in both arms after exercise 6 9 45 4.4 Pain in arms and shoulders after exercise 7 Right shoulder pain after exercise 51 3.4 8 d 6.6 9 53 Weakness and pain in legs during exercise d Muscle pain after exercise 8.3 62 d 10 9 68 0 Pain in arms and legs aggravated by exercise 11 *AMPD activity normal range is 60-300 units/g of noncollagen protein.

Age at onset of symptoms, years 4 11 25 30 29 41 33 50 50 55 55

tIndex case.

oligonucleotides corresponding to bases 881-900 (5'TTATAACTGCAGGAAGGTG-3') and bases 2239-2258 (5'TTGGTTTACTTTTTTTTATTC-3') were used to synthesize the 3' region. This approach was selected to reduce the probability of errors induced by PCR of long DNA sequences (26) and to facilitate functional analysis of individual mutations that might be found in the 5' and 3' regions of this transcript. PCR products were subcloned into pBS (Stratagene) for sequencing by the dideoxynucleotide chaintermination method (27). Thirty independently isolated pBS subclones were pooled for initial sequencing to obviate PCR-induced errors. Mutations identified by this screening procedure were subsequently confirmed by sequencing this region in individual clones. The PCR products were also subcloned into a prokaryotic vector, pKK233-2 (Pharmacia), for expression in Escherichia coli. This approach permits a rapid functional analysis of mutations since prokaryotes do not exhibit AMPD activity. Genomic DNA, isolated from patients and volunteers by standard methods (22), was amplified by PCR with oligonucleotides corresponding to intron 1 and intron 2 (5'-GCAATCTACATGTGTCTACC-3') and 5'-ATAGCCATGTTTCTGAATTA-3') for evaluation of exon 2; oligonucleotides corresponding to intron 2 and intron 3 (5'-AGCAGAAACTCTCAGCTGAC-3' and 5'-CTCCTTAGGTGCCAATATAC-3') were used for evaluation of exon 3. These PCR products were analyzed by direct sequencing (28), hybridization with allele-specific oligonucleotides (29), and/or restriction analysis as described in the text.

RESULTS Evaluation of Index Family. AMPD1 transcript size is not detectably altered in skeletal muscle from the index patient, and transcript abundance is not reduced (Fig. 1). To the contrary, the AMPD1 transcript, when normalized to creatine kinase M abundance, is approximately three times greater in this patient than a normal control. Studies of a heterozygote for AMPD deficiency also demonstrate that the mutant AMPD1 transcript is present in greater abundance than the wild-type transcript (unpublished observations). Comparison ofthe cDNA sequence for the index patient to that for a normal control reveals only two nucleotide differences (Fig. 2). Numbering from the first in-frame AUG, the 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 patient and it is a C in the control. The deviation in sequence of the patient cDNA at position 34 (exon 2) changes this codon from CAA to TAA-i.e., a stop codon instead of glutamine. The sequence difference at position 143 (exon 3) in the patient predicts an amino acid substitution of a leucine for proline.

To evaluate the significance of the missense mutation at position 143, wild-type and mutant cDNAs were ligated into a prokaryotic expression vector (pKK233-2) and E. coli lysates were assayed for AMPD activity. The 5' terminal 133 nucleotides of the mutant cDNA were replaced with the wild-type sequence to remove the nonsense mutation at position 34, resulting in a cDNA that had the single C -+ T mutation at position 143. Prokaryotes do not exhibit detectable AMPD activity and any activity in the lysate from E. coli transformed with the expression vector is presumably derived from the plasmid introduced into E. coli. This was confirmed by immunoprecipitation of the AMPD activity using an antiserum specific for the AMPDJ gene product (5). E. coli transformed with the wild-type cDNA exhibit AMPD activity in the range of 2-8 munits/mg of protein, and the AMPD activity in lysates from E. coli transformed with the cDNA mutated only at position 143 is not detectably different.

Although the mutation at position 143 appears to have little effect on catalytic activity of AMPD, the nonsense mutation at position 34 in this patient would be expected to give rise to a severely truncated peptide, only 11 amino acid residues compared to 747 residues in the control. Muscle extract from this patient has no detectable immunoreactive AMPD using polyclonal antiserum, whereas an easily detectable 86-kDa band is visualized in muscle extract from the normal control using this antiserum (Fig. 3). Deletion and point mutation analyses of human AMPD1 have shown that the active site 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 complete loss of AMPD activity, and the absence of immunoreactive protein can also be explained by this mutation. Genomic DNA from the index patient's mother, father, and brother, all of whom are asymptomatic, was also sequenced Ct Pt

AMPD DI --m

CK-M FIG. 1. AMPD1 mRNA abundance in muscle. Four micrograms of total RNA from control muscle (Ct) and muscle from the index patient (Pt) was resolved on a 1% agarose gel containing formaldehyde, transferred to Nytran paper (Schleicher & Schuell), and hybridized with a human AMPD1 cDNA probe (AMPD1) (23). After washing the filter, the same filter was used for hybridization with a human creatine kinase M (CK-M) cDNA probe (30).

Medical Sciences: Morisaki et al.

Proc. Natl. Acad. Sci. USA 89 (1992)

6459

B

A

Exon 2

Exon 3

Exon 2

PATIENT

CONTROL GATC

3, G

GATC

T

A G

T

"/

A

G T T A A Stop

-A-

A A

I

A

34

Cl

A

A

A A A

A

A

A

IG T

Exon 3 PATIE N T

CONTROL GATC

\G A

GATC

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

-*

-*

-.

in the relevant regions of exon 2 and exon 3 (Fig. 4). The AMPDJ genes for all three of these individuals contain both cytosine and thymidine nucleotides at positions 34 and 143, indicating they are heterozygotes. Studies of Other Patients with AMPD Deficiency. Muscle from 1 of the 10 other unrelated individuals with AMPD deficiency was used to prepare cDNA as in the index case and this cDNA was sequenced in its entirety. Genomic DNA from all 11 AMPD-deficient subjects was sequenced in the regions of exon 2 and 3. The only deviations from the control sequence

are

C

-.

T transitions at

positions

34 and 143

(Table

2). The AMPD activity levels in skeletal muscle and the symptoms in each of these individuals are listed in Table 1. An immunoblot of muscle lysate from one of these individuals confirmed the absence of immunoreactive AMPD peptide in this patient. Muscle was also available from one heterozygous individual and cDNA was prepared by reverse transcription PCR as described above. Sequencing of individual subclones of the PCR products prepared from this individual demonstrated Ct Pt

that the C

-*

T transitions at

are

present

DISCUSSION Eleven unrelated individuals with AMPD deficiency were evaluated in this study, and both alleles of the AMPDJ gene have the same two mutations in all 11 individuals-i.e., a C -+

FIG. 3. Immunoreactive AMPD peptide in muscle lysate. Thirty micrograms of protein from muscle lysate of a control (Ct) or the index patient (Pt) was resolved on an 8% SDS polyacrylamide gel, transferred to nitrocellulose paper, and incubated with polyclonal antiserum raised to rat AMPD1 (5). This antiserum does not react with other AMPD isoforms (5). The 86-kDa AMPD1 peptide in control lysate is indicated.

34 and 143

--

AMPD --

_

positions

the same allele. Population Studies. Genomic DNA obtained from 59 Caucasians, 13 African-Americans, and 106 Japanese was analyzed by two techniques for the nonsense mutation at position 34 (Fig. 5). PCR-amplified DNA from all subjects was immobilized on Nytran filters and hybridized to a wild-type or a mutant oligonucleotide for distinguishing the two types of alleles (29). PCR-amplified DNA was also restricted with Mae II in 31 subjects. This restriction enzyme recognizes the sequence ACGT, found only once in this region of the normal AMPD1 gene. The mutation at position 34 alters the sequence of this restriction site, and the PCR product from the mutant allele (AIGT) is not a substrate for this restriction endonuclease. Both tests give identical results in all cases. Seventeen percent of Caucasians and 23% of African-Americans are heterozygous for the nonsense mutation in exon 2, whereas none of the Japanese examined have this mutant allele. In addition, two Caucasians and one African-American were found to be homozygous for this mutation. Each of these individuals with the exon 2 nonsense mutation also has a C 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. on

T transition at

positions

34 and 143. The C

--

T transition

at position 34, in exon 2, results in a nonsense mutation that

predicts a severely truncated peptide as the product of this mutant transcript. This truncated peptide terminates prior to the synthesis of the catalytic domain of AMPD, providing an explanation for the significant decrease of this enzyme activity in these individuals. Either this truncated peptide is not

6460

~\'l~ ~ ~I

Medical Sciences: Morisaki et al. lF\(oll 2 I11. 1 RU() i

It (.

Proc. Natl. Acad. Sci. USA 89 (1992) 1 34I )II

i

t11I

I I *I 4

*1' 1'34 ( T3'e3 ( ('34A

d1.i)}.

I I IL () ( ATC

()1t1g I

II

%,

*

*

*RId |t~~~~~i.

aw io'X'\

I~ ~ ~ ~ ~ t l

F

Ofigo

_

\~ ~ ~ ~ ~ ~ ~.

B

i om i oimtro H X iom T1.34 (:1-34 (1:434

k

. .

FAMILY G

A

FIG. 4. AMPD1 sequence in family members of index case. (Upper) The sequence of PCR-amplified genomic DNA in the region of exon 2 (nucleotide 34) is illustrated on the left and the sequence in the region of exon 3 (nucleotide 143) is illustrated on the right. Positions 34 and 143 both exhibit the presence of a C and T at these sites, confirming this individual (the index patient's mother) is a heterozygote. (Lower) The genotype of each family member is depicted in this pedigree. The index patient is shown by the alrow.

recognized by the available antiserum or it is labile, explaining the absence of a discernible AMPD signal in immunoblots performed with muscle extracts from these patients. Our immunological observations in AMPD-deficient patients are similar to those reported in prior studies by Fishbein (14) and Sabina et al. (31). The variable residual AMPD activity found in our patients is similar to that reported by other investigaTable 2. Population study

Nucleotide 143 Nucleotide 34 C C T (11) T (11) C (14) C (47) C/T (4) C/T (10) T (2) T (1) C (1) African-American C (9) C/T (3) C/T (3) T (1) T (1) C (2) C (106) Japanese The AMPD-deficient patients are described in Table 1. The DNA samples from 59 Caucasians, 13 African-Americans, and 106 Japanese were obtained from normal volunteers or from investigators who had no knowledge of the donors' medical history. The number in parentheses is the number of individuals for whom the DNA sequence was determined at the indicated nucleotide on PCR samples prepared from genomic DNA as described in the legend of Fig. 5. Every DNA sample was screened for a C or T at nucleotide 34 through allele-specific oligonucleotide hybridization and confirmed by restriction digestion and/or direct sequencing. A limited number of DNA samples were screened for a C or T at nucleotide 143 since this required direct sequencing of the PCR product.

Group

Control AMPD deficient Caucasian

-

--

198 hp

1I 1)1) hip

- 87

FIG. 5. Screening method for detecting the mutation at nucleotide 34. (A) Allele-specific oligonucleotides. PCR-amplified genomic DNA from the region of exon 2 of AMPD1 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) hybridizes to both oligonucleotides. (B) Restriction endonuclease mapping. PCR-amplified genomic DNA from the region of exon 2 was digested with Mae II. This enzyme recognizes the sequence ACGT that occurs once in the wild-type PCR fragment, but it is absent from the mutant PCR fragment as consequence of the C -. T transition in exon 2. The undigested PCR fragment is 198 nucleotides in length; the digested PCR fragments are 111 and 87 nucleotides in length. Abbreviations for homozygote and heterozygote are the same as in A. bp, Base pairs.

tors (14, 31). We assume as they do that most of this residual activity reflects AMPD produced in nonmyocytes in the muscle tissue by one of the other AMPD genes since this activity is not reactive with antisera specific for the AMPD1 gene product (14, 31). However, we cannot exclude the possibility that a small fraction of the residual activity is produced in myocytes from the mutant AMPDJ gene as a consequence of alternative splicing (see below). The C -b T transition at nucleotide 143 in the AMPD1 transcript of these patients is apparently a silent mutation based on studies performed with recombinant peptides produced in a prokaryotic expression system. Although detailed kinetic studies have not been carried out with the wild-type and mutant peptides, which differ by only a proline or a leucine at codon 48, these two enzymes have comparable activity under the assay conditions employed. The apparent normal catalytic activity exhibited by the AMPD peptide harboring a mutation in codon 48 could assume clinical importance, if future studies demonstrate the human AMPD1 transcript is subject to alternative splicing, which deletes exon 2. In rat, exon 2 is deleted from the majority of transcripts produced from the AMPDJ gene in embryonic muscle (32) and in response to changes in neural innervation of skeletal muscle (unpublished observations). Since the primary sequence of the AMPDJ gene has been highly conserved in man and rat (23), a similar pattern of alternative splicing may occur in embryonic human muscle or in response to external signals. Alternative splicing of exon 2 in the human AMPD1 transcript would delete the nonsense mutation specified by the C -* T transition at nucleotide 34.

Medical Sciences: Morisaki

et

Proc. Natl. Acad. Sci. USA 89 (1992)

al.

Although we have not detected alternative splicing in adult human skeletal muscle (33), we cannot exclude at this time the possibility that it may occur at an early stage of skeletal muscle development, thereby ameliorating the consequences of the nonsense mutation in exon 2. The absence of a C -- T transition at nucleotide 143 in the 34 normal chromosomes examined makes it unlikely that this mutation is a common polymorphism in these populations. This fact, coupled with the 100%6 concordance of C -- T transitions at nucleotides 34 and 143 in 40 chromosomes analyzed from AMPD-deficient subjects and heterozygotes, suggests one of two explanations. Either there is a significant rate of spontaneous and coordinated mutation at both of these positions or the doubly mutated allele is present at a relatively high frequency in some populations. The latter explanation seems more plausible since it is difficult to envision a mechanism that results in spontaneous mutations affecting the same two nucleotides in all of these alleles. Furthermore, this mechanism would have to be restricted to certain populations since the doubly mutated allele has not been observed in the Japanese population. A more plausible explanation in our opinion is that the doubly mutated allele arose at some time in the remote past, and it has become widely disseminated in individuals of Western European descent. Studies are necessary to compare the frequency of this mutant allele in different ethnic groups in Western Europe, as well as African-Americans and native Africans, to gain additional insight into the origin of this allele. The high frequency of this allele in some populations may prove useful in studying the relationship between different ethnic groups. Recognizing that the frequency of AMPD heterozygosity is -20% in Caucasians and African-Americans, it is not surprising that several centers have reported 1-3% of randomly sampled muscle biopsies are deficient in this enzyme activity (6, 7, 9, 12). Clearly 1-3% of individuals in these populations do not have symptoms that are clinically severe enough to be classified as a metabolic myopathy. Thus, the frequency of this mutant allele in these populations raises a number of questions about the clinical implications of this form of AMPD deficiency. One could conclude this mutant allele is a harmless polymorphism. On the other hand, this mutation could be compensated for in some tissues or at some stages of development by removal of exon 2 through alternative splicing. The residual AMPD activity observed in muscle tissue of these patients, presumably a product of another member of this multigene family (3, 14, 31), might also compensate for the mutation in the AMPDJ gene, especially if this activity were present in myocytes. Other interpretations include the possibility that clinically significant myopathic symptoms develop only in individuals who have another inherited or acquired abnormality in energy metabolism. Answers to these questions will require additional physiological and molecular studies in controls and individuals with this mutant allele. T.M., M.G., and H.M. contributed equally to this report. We thank Dr. Ingrid Paetzke (Klinische Chemie Stddt Krankenhaus Munchen-Schwabing), who kindly provided the results on AMPD activity of the muscle biopsies of the 11 patients included in this study, and Dr. Hisaichi Fujii (Tokyo Women's Medical College), who generously supplied the DNA samples for the Japanese subjects. This work was supported by grants to E.W.H. (DK-12413, National Institutes of Health) and M.G. (Deutsche Forschungsgemeinschaft). 1. Sabina, R. L., Swain, J. L. & Holmes, E. W. (1989) in The Metabolic Basis of Inherited Disease, eds. Scriver, C. R.,

2. 3.

4. 5. 6. 7.

8. 9. 10. 11.

12. 13. 14. 15. 16.

17. 18. 19.

20. 21. 22. 23. 24. 25. 26. 27.

28. 29. 30. 31.

32. 33.

6461

Beaudet, A. L., Sly, W. S. & Valle, D. (McGraw-Hill, New York), Vol. 2, pp. 1077-1084. Sabina, R. L., Swain, J. L., Olanow, C. W., Bladley, W. G., Fishbein, W. N., DiMauro, S. & Holmes, E. W. (1984) J. Clin. Invest. 73, 720-730. Morisaki, T., Sabina, R. L. & Holmes, E. W. (1990) J. Biol. Chem. 265, 11482-11486. Ogasawara, N., Goto, H., Yamada, Y. & Watanabe, T. (1978) Eur. J. Biochem. 87, 297-304. Marquetant, R., Desai, N. M., Sabina, R. L. & Holmes, E. W. (1987) Proc. Nati. Acad. Sci. USA 84, 2345-2349. Fishbein, W. N., Armbrustmacher, V. W. & Griffin, J. L. (1978) Science 200, 545-548. Shumate, J. B., Katnik, R., Ruiz, M., Kaiser, K., Frieden, C., Brooke, M. H. & Carroll, J. E. (1979) Muscle Nerve 2, 213216. Scholte, H. R., Busch, H. F. N. & Luyt-Houwen, E. M. (1981) J. Inher. Metab. Dis. 4, 169-170. Kar, N. C. & Pearson, C. M. (1981) Arch. Neurol. 38, 279-281. Mercelis, R., Martin, J. J., Dehaene, I., de Barsy, Th. & Van den Berghe, G. (1981) J. Neurol. 225, 157-166. Hayes, D. J., Summers, B. A. & Morgan-Hughes, J. A. (1982) J. Neurol. Sci. 53, 125-136. Kelemen, J., Rice, D. R., Bradley, W. G., Munsat, T. L., DiMauro, S. & Hogan, E. L. (1982) Neurology 32, 857-863. Gertler, P. A. & Jacobs, R. P. (1984) Arthritis Rheum. 27, 586-590. Fishbein, W. N. (1985) Biochem. Med. 33, 158-169. Lally, E. V., Frieden, J. H. & Kaplan, S. R. (1985) Arthritis Rheum. 28, 1298-1302. Goebel, H. H., Bardosi, A., Conrad, B., Kuhlendahl, H. D., DiMauro, S. & Rumpf, K. W. (1986) Klin. Wochenschr. 64, 342-347. DiMauro, S., Miranda, A. F., Hays, A. P., Franck, W. A., Hoffman, G. S., Schoenfeldt, R. S. & Singh, N. (1980) J. Neurol. Sci. 47, 191-202. Fishbein, W. N., David, J. I., Nagarajan, K., Winkert, J. W. & Foellmer, J. W. (1980) Arch. Biochem. Biophys. 205, 360-364. Heller, S. L., Kaiser, K. K., Planer, G. J., Hagberg, J. M. & Brooke, M. H. (1987) Neurology 37, 1039-1042. Sinkeller, S. P. T., Joosten, E. M. G., Wevers, R. A., Oei, T. L., Jacobs, A. E. M., Veerkamp, J. H. & Hamel, B. C. J. (1988) Muscle Nerve 11, 312-317. Kaletha, K. & Nowak, G. (1990) Clin. Chim. Acta 190,147-156. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, NY), 2nd Ed. Sabina, R. L., Morisaki, T., Clarke, P., Eddy, R., Shows, T. B., Morton, C. C. & Holmes, E. W. (1990) J. Biol. Chem. 265, 9423-9433. Saiki, R. K., Scharf, S., Faloona, F., Mullis, K. B., Horn, G. T., Erlich, H. A. & Arnheim, N. (1985) Science 230, 487491. Veres, G., Gibbs, R. A., Schere, S. E. & Caskey, C. T. (1987) Science 237, 415-417. Saiki, R. K., Gelfand, D. H., Stoffel, S. J., Higuchi, R., Horn, G. T., Mullis, K. B. & Erlich, H. A. (1988) Science 239, 487-491. Sanger, F., Coulson, A. R., Garrell, B. G., Smith, A. J. H. & Roe, B. A. (1980) J. Mol. Biol. 143, 161-178. Wong, C., Dowling, C. E., Saiki, R. K., Higuchi, R. G., Erlich, H. A. & Kazazian, H. H. Jr. (1987) Nature (London) 330, 384-386. Saiki, R. K., Bugaman, T. L., Horn, G. T., Mullis, K. B. & Erlich, H. A. (1986) Nature (London) 324, 163-166. Perryman, M. B., Kerner, S. A., Bohlmeyer, T. J. & Roberts, R. (1986) Biochem. Biophys. Res. Commun. 140, 981-989. Sabina, R. L., Fishbein, W. N., Pezeshkpour, G., Clarke, P. R. H. & Holmes, E. W. (1992) Neurology 42, 170-179. Sabina, R. L., Ogasawara, N. & Holmes, E. W. (1989) Mol. Cell. Biol. 9, 2244-2246. Mineo, I., Clarke, P., Sabina, R. L. & Holmes, E. W. (1990) Mol. Cell. Biol. 10, 5271-5278.