doi:10.1086/316950 Am. J. Hum. Genet. 68:232–237, 2001 232 Report Gene Preference in Maple Syrup Urine Disease Mary M. Nellis1 and Dean J. Danner2 1Graduate Program in Nutrition and Health Sciences and 2Department of Genetics, Emory University School of Medicine, Atlanta Untreated maple syrup urine disease (MSUD) results in mental and physical disabilities and often leads to neonatal death. Newborn-screening programs, coupled with the use of protein-modified diets, have minimized the severity of this phenotype and allowed affected individuals to develop into productive adults. Although inheritance of MSUD adheres to rules for single-gene traits, mutations in the genes for E1a, E1b, or E2 of the mitochondrial branched-chain a-ketoacid dehydrogenase complex can cause the disease. Randomly selected cell lines from 63 individuals with clinically diagnosed MSUD were tested by retroviral complementation of branched-chain a-ketoacid dehydrogenase activity to identify the gene locus for mutant alleles. The frequencies of the mutations were 33% for the E1a gene, 38% for the E1b gene, and 19% for the E2 gene. Ten percent of the tested cell lines gave ambiguous results by showing no complementation or restoration of activity with two gene products. These results provide a means to establish a genotype/phenotype relationship in MSUD, with the ultimate goal of unraveling the complexity of this single-gene trait. This represents the largest study to date providing information on the genotype for MSUD. Maple syrup urine disease (MSUD; MSUD type Ia [MIM 248600], MSUD type II [MIM 248610], MSUD type Ib [MIM 248611]) is a rare inborn error of metabolism affecting !1 in 180,000 newborn infants in the general population (Peinemann and Danner 1994; Chuang and Shih 1995; Danner and Doering 1998). Despite its rare occurrence, MSUD is found in all racial and ethnic groups throughout the world (Danner and Elsas 1989). Neurological complications and death in the neonatal period can result from this inability to catabolize the branched-chain amino acids (BCAA) leucine, isoleucine, and valine. The pathophysiology of untreated MSUD is not well understood. Newborn-screening programs in most states and in several foreign countries detect those individuals at risk for expressing the disease. Affected individuals are protein intolerant and must be nurtured with synthetic diets that limit the intake of these essential BCAA (Elsas and Acosta 1988). These protein-modified Received October 9, 2000; accepted for publication November 6, 2000; electronically published December 7, 2000. Address for correspondence and reprints: Dr. Dean J. Danner, Department of Genetics, 1462 Clifton Road, Room 446, Emory Uni- versity School of Medicine, Atlanta, GA 30322. E-mail: ddanner @emory.edu � 2001 by The American Society of Human Genetics. All rights reserved. 0002-9297/2001/6801-0024$02.00 diets enable near-normal growth and development for affected individuals but must be tailored to the individ- ual, demonstrating one aspect of the complexity of this single-gene trait (Scriver and Waters 1999; Dipple and McCabe 2000). The inability to catabolize the BCAA results from de- fective function of the branched chain a-ketoacid de- hydrogenase (BCKD) complex. BCKD oxidatively de- carboxylates the branched-chain a-ketoacids (BCKA) formed by the transamination of the BCAA. This mi- tochondrial multienzyme complex is encoded entirely by nuclear genes. Four protein products impart catalytic function, and two other protein products are used to regulate BCKD activity-state in the different tissues (Gil- lim et al. 1983; Wagenmakers et al. 1984). This regu- lation is necessary to prevent the depletion of BCAA, since BCKD is present in all cells. The need for this wide tissue distribution of BCKD is not understood. Of the four genes that impart catalytic function, three encode proteins with specificity for the BCKA substrates. The fourth catalytic gene product also functions with other mitochondrial complexes (Yeaman 1989). Although inheritance of MSUD adheres to a simple autosomal re- cessive pattern, mutations in each of these three BCKD- specific genes have been shown to cause MSUD. Two of these genes encode subunits (E1a [Z14093] and E1b CORE Metadata, citation and similar papers at core.ac.uk Provided by Elsevier - Publisher Connector https://core.ac.uk/display/82350942?utm_source=pdf&utm_medium=banner&utm_campaign=pdf-decoration-v1 Reports 233 Table 1 BCKD Activity in Lymphoblast Lines with Defined Genotypes, from Patients with MSUD CELL LINE MUTANT GENE MUTATION MEAN � SD TOTAL BCKD ACTIVITYa (pmol 14CO2/mg protein/3 h) Not Transduced E1a E1b E2 EM408 E1a Y438Nb 0 � 0 166 � 51 0 � 0 0 � 0 EM1952c E1b N176Y/R324X 24 � 2 10 � 1 364 � 89 19 � 6 EM3069d E2 1031GrA/ (20-kb del) 71 � 3 71 � 14 82 � 21 149 � 43 DG75 None 6,158 � 626 ND ND ND a All cells were treated with 1mM a-chloroisocaproate for 10 min prior to addition of substrate to induce ability to measure full activity of the BCKD complex. ND p not done. b Known Mennonite cell line. c From McConnell et al. (1997). d From Herring et al. (1992). Figure 1 Western blot of BCKD proteins in mitochondria. Mi- tochondria were prepared from EM6229 cells without viral transduc- tion or after transduction with either the E2 or E1b vector, and 20 mg were resolved for immunoblots. Mitochondria from DG75 cells serve as a wild-type control line. [M55575]) that form a heterotetramer that uses thiamin pyrophosphate as a cofactor to decarboxylate the BCKA (Ævarsson et al. 2000). The third gene product, the acyl- transferase (E2 [X66785]) core protein, transfers the branched-chain acyl moiety to coenzyme A (Yeaman 1989; Reed and Hackert 1990). The majority of mutations described to date have been in the gene that encodes the E1a subunit (Danner and Doering 1998). It remains to be shown whether any one of these three genes has an increased tendency toward mutation. Although all the genes have been character- ized and their chromosomal location identified, no easy method is available to determine whether the E1a, E1b, or E2 gene harbors the mutant alleles. The only mutant allele found with any frequency exists in the Mennonite community and is associated with a founder effect (Dan- ner and Doering 1998). Finding a simple, reliable method to identify the spe- cific mutant gene that causes MSUD from the three pos- sible choices has presented a challenge. Western blots can be used to define individuals with MSUD who lack the E2 protein (Danner et al. 1985). However, reduction in the antigenic presence of E1a, E1b, or both can result from mutations in either gene, because the ability to form the a2b2 tetramer is affected. In some MSUD cell lines, the proteins of the complex are antigenically in- distinguishable from those seen in cells with normal ac- tivity. Thus, immunoblots have limited value in helping to identify the gene harboring the mutant alleles. Enzyme-activity measurements in cultured cells de- rived from patients are used to confirm the clinical di- agnosis of MSUD made by quantification of plasma BCAA concentration. We and others have shown that impaired BCKD activity is functionally restored to cul- tured cells by episomal addition of a wild-type cDNA for the defective gene (Litwer et al. 1989; Koyata et al. 1993; Mueller et al. 1995). Here we use this BCKD- activity-complementation approach to identify the gene harboring the mutant alleles. Retroviral plasmids hold- ing the wild-type cDNA sequence were constructed for each of the three MSUD-specific genes in order to ef- fectively and efficiently transduce both fibroblasts and lymphoblasts. Cell lines with defined mutations from individuals with MSUD were used to establish condi- tions for BCKD assay after transduction. As shown in table 1, the return of function to BCKD is clearly evident by 72 h after transduction of lymphoblasts and can spe- cifically distinguish the defective gene despite an esti- mated transduction efficiency of only 30%. Production of the subunit encoded by the viral vector was dem- onstrated by western blot of mitochondrial protein from an E2-negative cell line transduced with the E2 and E1b vectors (fig. 1). Efficiency of transduction in fibroblast cultures is 170% by 48 h; thus, assays of BCKD activity can be performed at this time (table 2). This enzyme-complementation analysis was applied to 63 cell lines from individuals with MSUD and without a known Mennonite ancestor. This represents the largest single study sample used for analysis of the MSUD ge- notype. No other qualification was placed on sample selection. The cell lines were obtained from physicians 234 Am. J. Hum. Genet. 68:232–237, 2001 Table 2 BCKD Activity in Fibroblast Lines with Unknown Genetic Mutations, from Patients with MSUD CELL LINE MEAN � SD TOTAL BCKD ACTIVITYa (pmol 14CO2/mg protein/3 h) Not Transduced E1a E1b E2 EM15496 11 � 4 2,437 � 13 8 � 0 10 � 2 EM5621 1 � 9 5 � 0 2,128 � 58 17 � 1 EM4566 22 � 3 29 � 12 26 � 4 2,112 � 453 EM1118b 3,281 � 113 ND ND ND a Conditions for the assay are described in the text and in table 2. ND p not done. b Control cell line. Figure 2 Percentage of cell lines with mutations in each of three genes that cause MSUD. A, Results of 63 cell lines using restoration of BCKD activity assay. Twenty-nine females and 34 males, excluding those with known Mennonite heritage, were tested. B, Summary of percentages based on mutations described elsewhere. throughout the world and included individuals from North, Central, and South America; Europe; and Asia. This sample includes 34 males and 29 females, in ac- cordance with the autosomal recessive nature of MSUD. As depicted in figure 2A, mutations in the gene for E1b on chromosome 6 are approximately equal to the fre- quency of E1a mutations, whereas those of E2 are some- what less common. In contrast, 52% of the previously characterized mutant alleles were the result of defects in the gene for the E1a subunit on chromosome 19, and 36% were defects in E2, as summarized in figure 2B (Chuang and Shih 1995; Danner and Doering 1998). Results from five lymphoblastoid lines and one fibro- blast line were ambiguous when this BCKD comple- mentation analysis was used. For four cell lines, no stim- ulation of residual activity was observed with any subunit in three separate transductions for each cell line. The other two cell lines exhibited !10% increases in residual activity with two different subunits. Five of the six cell lines have residual activity 12% of wild-type BCKD activity, a condition that has been used to define an “intermittent” or “intermediate” form of MSUD (Pei- nemann and Danner 1994). In contrast, the majority of the other 57 cell lines have residual activity that is !2% of wild-type BCKD activity (table 3). Western blots of mitochondrial proteins from five of the six cell lines that gave ambiguous results show the antigenic presence in all subunits. One lacked the antigenic presence of both E1a and E1b. Another cell line that was found to be ambiguous in the lymphoblast assay was resolved when fibroblasts from the patient were analyzed; this was likely a result of the higher transduction efficiency for fibroblasts. Unfortunately, fibroblasts are not available for all patients studied. Definition of the specific mu- tations in the cells with ambiguous complementation data will resolve these questions and offer a clearer un- derstanding of MSUD variability, especially regarding the “intermediate” phenotype (Peinemann and Danner 1994). The somewhat lower frequency of mutations associ- ated with the E2 gene on chromosome 1 may be real, but other explanations are possible. First, if the cell lines that give ambiguous results have mutations in the E2 gene, then an equal distribution of mutations among the three genes would result. A second explanation may be that the E2 mutations are being missed at the clinical level. As seen in table 3, individuals with E2 mutations can have higher residual activity than that found in cell lines with mutations in the other two genes. In some individuals there is no antigenic protein being produced (fig. 1) and yet some decarboxylation of the ketoacid substrate occurs (Danner et al. 1985; Ellerine et al. 1993). Mutations in E2 have also been associated with the “thiamin-responsive” form of MSUD (Danner et al. 1975; Fernhoff et al. 1985; Ellerine et al. 1993). One possible explanation for the thiamin response is that the E1 decarboxylase, which uses thiamin pyrophosphate as a cofactor in the reaction mechanism, can function in- dependently of the E2 protein. The branched-chain acyl- thiamin pyrophosphate intermediate formed in the de- carboxylation reaction is displaced by the excess thiamin pyrophosphate in the cell, thus allowing the enzyme to catalyze another round of decarboxylation (Herring et al. 1992; Danner and Doering 1998). The independent function of E1 could explain the higher residual activity observed for the E2 mutant cell lines. In some cases this activity could be high enough to prevent the full clinical phenotype, and therefore these individuals are never classified as having MSUD. The ability of the enzyme complementation analysis to identify the mutant genes was confirmed by nucleotide sequence analysis of the indicated mutant gene with the primers listed in table 4. Results for one patient from each complementing group are reported in table 5. Pa- rental cell lines for each proband were used to confirm the transmission of the mutant allele. The three examples all happened to be homozygous for the mutant allele, which is unusual for MSUD, since most probands thus far characterized are compound heterozygotes. Another Reports 235 Table 3 Distribution of MSUD Cell Lines, Based on Mutant Gene and Percent Residual Activity of BCKD Above That in the Control Lines MUTANT GENE NO. OF MSUD CELL LINES IN WHICH RESIDUAL ACTIVITY COMPARED WITH THAT IN WILD-TYPE BCKD IS !2% 2%–10% 110% E1a 19 1 1 E1b 23 0 1 E2 9 3 0 Ambiguousa 1 3 2 a BCKD activity was complemented with more than one subunit, or there was no complementation. Table 4 Primers Used to Prepare DNA for Nucleic Acid Sequence Analysis Gene and Primer Sequence (5′r3′) Position in Coding Sequence Annealing Temperature (�C) E1a: Forward CTAAACCGTGGTTTGAGCCA 32–52 54 Reverse GCGTACGCCGCCCCCACCGCC 679–658 Forward GGCGGTGGGGGCGGCGTACGC 685–679 54 Reverse GGTGGGGGTGGGCTGAGCAGG 1374–1353 E1b: Forward CGGGACTACGGGCAAACTCA 191–210 58 Reverse AGAGCCCCATGACCAACACAGCCCC 628–604 Forward TGGCTACTCAGGCTCAGG 75–92 50 Reverse TCCCCAGAGCGATAGCGATACTTG 565–542 Forward GGAATCGGAATTGCGGTCACTGGAG 444–468 58 Reverse tcgcgaattcCAGTTACGTTAATGTCAGGGG 1286–1267 E2: Forward GCTGCAGTCCGTATGCTGA 18–36 54 Reverse CTAGTAGCATAAAAGCTGGG 1456–1434 unexpected result was that the mutant allele defined for the E1a gene in proband EM5579 was the allele that accounts for 199% of MSUD in the Mennonite popu- lation. It was previously thought that this allele was not found with greater frequency than other mutations in this gene in the general population. This proband’s fam- ily has no known Mennonite heritage and the mating is not consanguineous. As more of the cell lines with mu- tations in this gene are defined, it will become possible to ascertain the true frequency of this mutant allele in the general population. Mutations described in the E1b gene are for a pro- band who is the product of a consanguineous mating, of Turkish origin. The mutant allele results in a V251A substitution in the protein. According to the crystal structure coordinates (Ævarsson et al. 2000), the valine residues are on the surface of the a2b2 tetramer, pos- tulated to be in a region of interaction with the E2 core protein. It is not yet known how this alanine substi- tution in this position disrupts the function of BCKD (Ævarsson et al. 2000). A second Turkish proband (also the product of a consanguineous mating) who complemented with the E1b subunit was homozygous for the nucleotide change 1149TrA, introducing a stop codon at Y383. These results argue against the presence of a common allele within families of Turkish decent. The homozygous mutation in the E2 gene in proband EM5591 produces an R462P amino acid substitution. At present it is not known whether this is a consan- guineous mating. The R462 residue is just C-terminal to the putative coenzyme A binding site of the protein (Danner et al. 1989). The crystal structure for the E2 protein has not been solved, so it would be highly spec- ulative to predict the way in which this amino acid sub- stitution might change the ability to form the CoA ester. The complex nature of single-gene traits has been at- tributed to variation in specific mutation sites, amino acid substitutions in the protein product, and the ge- nomic background of affected individuals (Scriver and Waters 1999; Dipple and McCabe 2000). Variation in MSUD can also result from the fact that mutations in three different genes can produce a similar phenotype. The products of these genes must interact in a multien- zyme complex to achieve their cellular function. To be- gin to unravel the complexity of MSUD we must first ask whether there is a variation in the occurrence of mutations in the three genes. Here we used this rescue technique to identify the defective subunit, and therefore the gene, with the mutant alleles. The method was ef- fective in �90% of cell lines tested. Previously, it was necessary to sequence all three genes and to exclude polymorphic changes. Now only one gene needs to be 236 Am. J. Hum. Genet. 68:232–237, 2001 Table 5 Nucleic Acid Changes in Mutant Genes Identified by BCKD Complementation Analysis CELL LINE MUTANT GENE MEAN � SD TOTAL BCKD ACTIVITYa (pmol 14CO2/mg protein/3 h) NUCLEOTIDE CHANGEb AMINO ACID CHANGEBaseline Rescued EM5579 E1a 0 � 1 136 � 6 1312TrA Y438N EM608 E1b 37 � 11 123 � 18 752TrC V251A EM5591 E2 318 � 54 569 � 29 1385GrC R462P a BCKD activity is defined in table 2. b All three probands were homozygous for the indicated change, and both parental lines in each family were heterozygous for the mutant allele. scrutinized for nucleic acid changes. It appears that mu- tations occur in each of the three genes, and analysis of additional cell lines will allow a better estimation of the exact frequency. When amino acid substitutions result from these nucleic acid changes, it may be possible, in the future, to design specific reagents that will interact with the mutant protein in a way that alters the structure into a functional form. These reagents would offer a new and improved therapy for individuals with MSUD. Acknowledgments We thank all the physicians who have sent patient material to us over the past 20� years. Without their help this study would not have been possible. Special thanks are given to Dr. Joshy Jacobs (Department of Microbiology and Immunology, Emory University School of Medicine) for his help in retrovi- ral vector construction. Lymphocyte transformations were done by the Clinical Research Facility Service Laboratory with grant support from NIH MO1-RR00039, and some fibroblast cultures were obtained through the courtesy of the Emory Ge- netics Laboratory. M.M.N. was supported in part by an NIH training grant 5T32DK07734-03. We also thank Barbara Sum- mers for her technical assistance. A preliminary report of this work was presented at the Annual Meeting of the American Society of Human Genetics, Philadelphia, October 2000. Electronic-Database Information Accession numbers and URLs for data in this article are as follows: GenBank, http://www.ncbi.nlm.nih.gov/Genbank/ (for E1a [ac- cession number Z14093], E1b [accession number M55575, and E2 [accession number X66785]) Online Mendelian Inheritance in Man (OMIM), http:// www.ncbi.nlm.nih.gov/Omim/ (for MSUD types Ia [MIM 248600], Ib [MIM 248611] and II [MIM 248610]) References Ævarsson A, Chuang JL, Wynn RM, Turley S, Chuang DT, Hol WG (2000) Crystal structure of human branched-chain a-ketoacid dehydrogenase and the molecular basis of mul- tienzyme complex deficiency in maple syrup urine disease. Structure Fold Des 8:277–291 Chuang DT, Shih VE (1995) Disorders of branched chain amino acid and keto acid metabolism. In: Scriver CR, Beau- det AL, Sly WS, Valle D (eds) The metabolic basis of in- herited disease. Vol. 1. McGraw-Hill, New York, pp 1239– 1277 Danner DJ, Armstrong N, Heffelfinger SC, Sewell ET, Priest JH, Elsas LJ (1985) Absence of branched chain acyl-trans- ferase as a cause of maple syrup urine disease. J Clin Invest 75:858–860 Danner DJ, Davidson ED, Elsas LJ (1975) Thiamine increases the specific activity of human liver branched chain a-ke- toacid dehydrogenase. Nature 254:529–530 Danner DJ, Doering CB (1998) Human mutations affecting branched chain a-ketoacid dehydrogenase. Front Biosci 3: d517–524 Danner DJ, Elsas LJ (1989) Disorders of branched chain amino acid and keto acid metabolism. In: Scriver CR, Beaudet AL, Sly WS, Valle D (eds) The metabolic basis of inherited dis- ease. McGraw-Hill, New York, pp 671–692 Danner DJ, Litwer S, Herring WJ, Pruckler J (1989) Construc- tion and nucleotide sequence of a cDNA encoding the full- length preprotein for human branched chain acyltransferase. J Biol Chem 264:7742–7746 Dipple KM, McCabe ERB (2000) Phenotypes of patients with “simple” Mendelian disorders are complex traits: Thresholds, modifiers, and systems dynamics. Am J Hum Genet 66:1729–1735 Ellerine NP, Herring WJ, Elsas LJ, McKean MC, Klein PD, Danner DJ (1993) Thiamin-responsive maple syrup urine disease in a patient antigenically missing dihydrolipoamide acyltransferase. Biochem Med Metab Biol 49:363–374 Elsas LJ, Acosta PE (1988) Nutritional management of inher- ited metabolic disorders. In: Shils ME, Young V (eds) Mod- ern nutrition in health and disease. Lea and Febiger, Phil- adelphia, pp 1337–1379 Fernhoff PM, Lubitz D, Danner DJ, Dembure PP, Schwarz HP, Hillman R, Bier DM, Elsas LJ (1985) Thiamine responsive maple syrup urine disease. Pediatr Res 19:1011–1016 Gillim SE, Paxton R, Cook GA, Harris RA (1983) Activity state of the branched chain a-ketoacid dehydrogenase com- plex in heart, liver, and kidney of normal, fasted, diabetic, and protein-starved rats. Biochem Biophys Res Commun 111:74–81 Herring WJ, McKean M, Dracopoli N, Danner DJ (1992) Reports 237 Branched chain acyltransferase absence due to an ALU- based genomic deletion allele and an exon skipping allele in a compound heterozygote proband expressing maple syrup urine disease. Biochim Biophys Acta 1138:236–242 Koyata H, Cox RP, Chuang DT (1993) Stable correction of maple syrup urine disease in cells from a Mennonite patient by retroviral-mediated gene transfer. Biochem J 295:635– 639 Litwer S, Herring WJ, Danner DJ (1989) Reversion of maple syrup urine disease phenotype of impaired branched chain a-ketoacid dehydrogenase complex activity in fibroblasts from an affected child. J Biol Chem 264:14597–14600 McConnell BB, Burkholder B, Danner DJ (1997) Two new mutations in the human E1b subunit of branched chain a- ketoacid dehydrogenase associated with maple syrup urine disease. Biochim Biophys Acta 1361:263–271 Mueller GM, McKenzie LR, Homanics GE, Watkins SC, Rob- bins PD, Paul HS (1995) Complementation of defective leu- cine decarboxylation in fibroblasts from a maple syrup urine disease patient by retrovirus-mediated gene transfer. Gene Ther 2:461–468 Peinemann F, Danner DJ (1994) Maple syrup urine disease 1954 to 1993. J Inherit Metab Dis 17:3–15 Reed LJ, Hackert ML (1990) Structure-function relation- ships in dihydrolipoamide acyltransferases. J Biol Chem 265:8971–8974 Scriver CR, Waters PJ (1999) Monogenic traits are not simple. Trends Genet 15:267–272 Wagenmakers AJM, Schepens JTG, Veldhuizen JAM, Veer- kamp JH (1984) The activity state of the branched-chain 2- oxoacid dehydrogenase complex in rat tissues. Biochem J 220:273–281 Yeaman SJ (1989) The 2-oxo acid dehydrogenase complexes: recent advances. Biochem J 257:625–632 Gene Preference in Maple Syrup Urine Disease Acknowledgments References