Genetic basis of galactosemia

HUMAN MUTATION 1:190-196 (1992)

MUTATION UPDATE

Genetic Basis of Galactosemia Juergen K.V. Reichardt Howard Hughes Medical Institute and Department of Cell Biology, Baylor College of Medicine, Texas Medical Center, Houston, Texas 77030-3498

Communicated by Haig H. Kazazian, Jr.

Classic galactosemia is an inborn error of galactose metabolism and results from deficiency of the ubiquitously expressed enzyme galactose- 1 -phosphate uridyltransferase (GALT). Nine missense mu. tations, three splicing mutations, three GALT protein polymorphisms, and one silent nucleotide substitution have been identified to date. Most of the disease-causing mutations are rare among patients. The most common mutation, Q188R, has a frequency of only one-fourth in the patient population examined. Three classes of disease-causing mutations have been reported: CRM + missense mutations (the most common class), CRM- missense mutations, and splicing mutations. Thus, galactosemia is heterogeneous at the molecular level, which is noteworthy in light of the well-documented clinical variability observed in this disorder. It has also been shown that eight of nine galactosemia missense mutations occur in evolutionarily well-conserved domains, suggesting that they affect functionally and/or structurally important residues. In contrast, all protein polymorphisms alter variable amino acids which presumably are not important for the enzyme’s function. o 1992 WiIey-Liss, Inc.

KEY WORDS: Inborn error of metabolism, Mutational analysis, Structure-function analysis, PCR

INTRODUCTION

The galactosemias are a group of three inborn errors of metabolism resulting in the inability to metabolize galactose (Segal, 1989). Deficiency of each one of the three galactose-converting enzymes, galactokinase (GALK) , galactose- 1 -phosphate uridyltransferase (GALT), and UDP-Gal4’- epimerase (GALE), can result in elevated levels of galactose and its metabolites, hence the term “galactosemia. ” However, the most common and most severe form of the disease results from deficiency of galactose- 1 -phosphate uridyltransferase (GALT; EC 2.7.7.12; Segal, 1989; McKusick 230400). Early symptoms include vomiting, diarrhea, failure to thrive, and hepatomegaly. Cataracts and Esche- richia coli sepsis are also commonly observed (Segal, 1989). These symptoms disappear upon the insti- tution of a galactose-restricted diet. Therefore, most U.S. states and many Western countries have instituted newborn screening programs for galactosemia. However, long-term complications such as neurologic abnormalities and ovarian failure persist in many well-managed patients (Waggoner et al., 1990). Genetically, galactosemia is inherited as an autosomal recessive disease with a frequency of about 1:60,000 (Levy and Hammersen, 1978). The GALT gene has been mapped to 9p13-21 (Segal, 1989; Reichardt, 1989) while the other two galac-

tose-converting enzymes, GALK and GALE, are located on different chromosomes.

The molecular analysis of galactosemia began with the cloning (Reichardt and Berg, 1988a) and sequence characterization (Flach et al., 1990a) of a full-length, expressible GALT cDNA. This cDNA was used to probe Southern and Northern blots of galactosemic patients and no abnormalities could be detected (Reichardt, 1989, 1991). Finally, the cDNA was overexpressed as a trpE fusion protein in E. coli and used as an antigen to raise a poly- clonal antiserum (Reichardt, 1989). This serum detected full-length GALT CRM (cross-reacting material) on Western blots of 13 patients (Rei- chardt, 1989, 1991). This result led to the suggestion that galactosemia is caused predominantly by missense mutations since each patient examined by Western blotting had to bear at least one allele

Received May 2, 1992; accepted June 1, 1992

Present address of J.K.V. Reichardt: Institute for Genetic Medi- cine and Department of Biochemistry, Univ. of So. California School of Medicine, Los Angeles, CA 90033.

Abbreviations: ASO, allele-specific oligonucleotide; CRM, cross- reacting material; D, Duarte allele; ES, embryonal stem; GALE, UDP-Gal 4‘-epimerase; GALK, galactokinase; GALT, galactose-l- phosphate uridyltransferase; GGTB, P-1,4-glycoprotein galactosyl- transferase; RFLP, restriction fragment length polymorphism; PCR, polymerase chain reaction; PKU, phenylketonuria; STR, short tandem repeat; UDP-Gal, uridine diphosphogalactose.

0 1992 WILEY-LISS, INC.

GENETIC BASIS OF GALACTOSEMIA 191

TABLE 1. Galactosemia Mutations and GALT Polymorohisms"

Residual Change CPG activity

Galactosemia missense mutations CRM' V44M No <1 L74P No <1 M142K No 4 F171S No 7 Q188R No 10 L195P No <1 R231H Yes n/a R333W Yes <1 CRM- R148W Yes <1

i1087 Yes n/a two n/a n/a

L62M No 108 S135L Yes 93 N314D No 129

Galactosemia splicing mutations

GALT protein polymorphisms

Frequency

Normal Galactosemia Reference

018 0114 018 0114 0/10 0114 nla 0110

0/14

0120 n/a

018 0114 118

1/16 1/26 2/16 2/26'' 5/20 2/24 112 1/20

1/24

1/20 n/a

1/16 2/26b 1/16

Reichardt and Woo (1991) Reichardt et al. (1992b) Reichardt and Woo (1991) Reichardt et al. (1992b) Reichardt et al. (1991) Reichardt et al. (1992a) Okano (personal communication) Reichardt et al. (1991)

Reichardt et al. (1992a)

Wadelius et al. (1992) Flach et al. (1990b)

Reichardt and Woo (1991) Reichardt et al. (1992b) Reichardt and Woo (1991)

"CpG denotes mutations altering a CpG dinucleotide by C to T transitions. Residual activity is expressed as percent of normal in cos cells overexpressing the normal or mutant polypeptides. Frequencies are shown as positive alleles identified per total number of alleles examined. Missense mutations are identified by the single letter amino acid code for the changed residue and i stands for insertion at the particular nucleotide number given. bLinked alleles found in one Caucasian and one African-American patient; n/a, not available.

with a missense mutation (Reichardt, 1989, 1991). This conclusion is supported by earlier data (Banroques et al., 1983) and has since been con- firmed by the mutational analysis of galactosemia patients. In fact, nine galactosemia missense mutations have been reported as well as three splicing mutations, three GALT protein polymorphisms and one silent nucleotide substitution (Flach et al., 1990b; Okano, personal communication; Rei- chardt and Woo, 1991; Reichardt et al., 1991, 1992a,b; Wadelius et al., 1992).

This review focuses on recent developments in the molecular analysis of galactosemia. Molecular genetic studies of galactosemia patients have opened new avenues of research in the field and have led to interesting findings. Galactosemia mutations and GALT polymorphisms will be dis- cussed along with their implications for past and future galactosemia research.

MUTATIONS AND POLYMORPHISMS IN HUMAN GALT

The cloning and sequence characterization of a human GALT cDNA (Reichardt and Berg, 1988a; Flach et al., 1990a) has allowed investigators to study mutations and polymorphisms in GALT. Ex- amination of patients by Southern, Northern, and Western blotting failed to detect any gross abnormalities in GALT (Reichardt, 1989, 1991). The

presence of full-length protein in all patients examined (Banroques et al., 1983; Reichardt, 1989, 1991) led to the suggestion that most galactosemia mutations would be of the missense type (Rei- chardt, 1989, 1991).

The following strategy has been used success- fully in the past to identify galactosemia mutations and GALT polymorphisms: mRNA is prepared from cells of interest and then reverse-transcribed into total cDNA. The entire GALT coding region is then amplified from this uncloned library by the polymerase chain reaction (PCR; RT-PCR) and directly sequenced to eliminate PCR artifacts (Rei- chardt and Woo, 1991). This procedure leads to the characterization of nucleotide substitutions which, however, may be either disease-causing mutations or benign polymorphisms. Therefore, it is imperative that nucleotide substitutions be next evaluated genetically by screening normal and galactosemic individuals for the presence or absence of each novel nucleotide substitution. Screening is done either by allele-specific oligonucleotide (ASO) hybridization or by scoring fortuitously altered restriction sites. Finally, each nucleotide substitution is analyzed biochemically by recreating it by site-directed mutagenesis, overexpressing the encoded polypeptide in cos cells, and assaying its biochemical properties. These two criteria, genetic and biochemical, allow investigators to categorize nucleotide substitutions as disease-causing muta-

192 REICHARDT

tions or protein polymorphisms (Reichardt and Woo, 1991).

To date, a total of 12 galactosemia mutations, three polymorphisms, and one silent nucleotide substitution have been reported (Table 1). Three of the galactosemia mutations are splicing aberrations (Flach et al., 1990b; Wadelius et al., 1992) and nine disease-causing mutations, V44M, L74P, M142K, R148W, F171S, Q188R, L195P, R231H, and R333W, are of the missense type (Table 1; Fig. 1A; Okano, personal communication; Rei- chardt and Woo, 1991; Reichardt e t al., 1991, 1992a,b).

The three splicing mutations can either lead to deletions (Flach et al., 1990b) or insertions (Wadelius et al., 1992). I t is noteworthy that the latter splicing mutation is predicted to result in an insertion of 18 amino acids in a conserved area of the enzyme near the carboxy terminus by activat- ing a cryptic splice donor site.

The homologous enzymes galT from E. coli, GAL7 from bakers’ yeast, and GALT from humans share about 35% sequence identity (Flach et al., 1990a). Mutations can be either in completely conserved domains, e.g., R333W (Fig. lA) , or in totally nonconserved domains, e.g., R148W (Fig. 1A). However, the most common case is substantial sequence conservation as for the F171S mutation (Fig. 1A) or partial conservation as seen with the L195P mutation (Fig. 1A). In fact, eight galactosemia missense mutations are found in conserved areas. Three mutations, F171S, Q188R, L195P, cluster around the putative active site nucleophile, histidine 186, highlighting this critical domain of the enzyme (Fig. 1A; Reichardt et al., 1992a,b). The six other galactosemia missense mutations are dispersed throughout the protein (Fig. 1A; Okano, personal communication; Reichardt and Woo, 1991; Reichardt et al., 1991, 1992a,b). Thus, amino acid residues critical for the enzyme’s structure and/or function are conserved by natural selection throughout evolution and their mutation in humans results in disease-causing mutations. Finally, one galactosemia missense mutation, R148W, encodes an unstable polypeptide resulting in a CRM- phenotype and it is found in a nonconserved area of the enzyme (Reichardt et al., 1992a).

Only one galactosemia mutation, Q188R, has been reported to account for a significant number of mutant alleles, about 26% in the patient population examined (Table 1; Reichardt et al., 1991). All other disease-causing mutations appear to be rare among patients. Thus, galactosemia seems to

be heterogeneous at the molecular level. Three different molecular mechanisms have been found to cause the disease to date: CRM+ missense mutations, CRM- missense mutations, and splicing aberrations. The latter two categories are null mutations while the former is sometimes associated with residual activity in cos cells. Furthermore, galactosemia can be caused by a myriad of different mutations. This molecular heterogeneity is noteworthy in light of the well-documented clinical variability found in this disorder (Waggoner et al., 1990).

The CpG dinucleotide has been shown to be a “hotspot” for mutations in humans since it can undergo oxidative deamination of 5-methyl- cytosine (Cooper and Youssoufian, 1988). Five CpG mutations have been reported in human GALT, the i1087 splicing aberration, the R148W, R23 lH, and R333W galactosemia missense mutations, and the S135L GALT polymorphism (Table 1). All four are C to T transitions as predicted for deaminations of CpG dinucleotides. Thus, 5 of 16 nucleotide substitutions in human GALT are CpG mutations, which is in the range of the average mutation frequency compiled by Cooper and Yous- soufian (1988).

All galactosemia mutations and GALT polymorphisms reported to date are point mutations (Table 1) and no gene rearrangements have been detected by Southern blotting (Reichardt, 1989, 1991). This may be due to the small size of the GALT gene, reportedly only about 4 kb (Leslie et al., 1991). The high frequency of point mutations in galactosemia probably explains the measurable reversion frequency of the galactosemic phenotype in tissue culture (Benn et al., 1981; Kelley et al., 1983). Furthermore, at least three mutations cre- ate new CpG dinucleotides (L74P, Ql88R, L195P) which could undergo oxidative deaminations in culture yielding revertant clones. Finally, the large number of missense mutations might explain the phenomenon of interallelic complemen- tation observed in some galactosemic cell fusions (Nadler e t al., 1970).

Three galactosemia missense mutations encode substantial residual activities when overexpressed in cos cells: M142K, F171S, and Ql88R (Table 1; Reichardt and Woo, 1991; Reichardt et al., 1991, 1992b). In fact, each fully genotyped patient is predicted to bear at least one allele with residual activity in cos cells but frequently galactosemia patients are compound heterozygotes with a second inactive mutation. Furthermore, low levels of residual activity have been measured in galactosemic


A) GALACTOSEMI A MUTATIONS

pmwsed adive site msidaoes

cys phe his pro his \ Y

NH2 r; V44M L74P M142K

m P k H . s a r , i n e s vLVSa npLCP PlMSV S.ce revi s i ae ILVSP CyLCP PqMkq E. coli ILVSP CfLCa PelSV

F171 S S

QIFEN Q I FEN Q-FEN

/ r L195P

P SFLPd esiPs SFLPn

R148W Q188R R231Q W r q

eirav HcQvW KeRlV dLvhI HGQi W K SRvV aLteI HGQaW gSRtV

R333W W

TVRKL TVRKL TVRKL

GALT POLYMORPHISMS

FIGURE 1: Mutations and polymorphisms in human GALT. (A) Alignment of nine galactosemia missense mutations. The homologous proteins galT from E. coli, GAL7 from yeast, and GALT from humans were aligned in the area surrounding each mutation. The overall amino acid sequence identity for these three species is about 35% (Flach et a]., 1990a). Res- idues conserved in two species are in capital letters while residues identical in all three species are in bold. A dash indicates noncontinuity for homology maximization. Putative

active site residues are shown (Reichardt and Berg, 1988b) and the proposed active site nucleophile, histidine 186, is italicized (Kim et al., 1990; Reichardt et al., 1992a). Note that eight of the nine missense mutations occur in conserved domains and that three mutations cluster around the putative active site nucleophile. (B) Alignment of three GALT polymorphisms. The GALT protein polymorphisms occur in nonconserved areas of the enzyme and are dispersed throughout.

194 REICHARDT

cells and tissues (Russell and DeMars, 1967; Rog- ers et al., 1970; Segal et al., 1971; Kelley et al., 1989; Reichardt, 1991). Future therapies aimed at ameliorating the long-term complications of the disease may focus on enhancement of this residual activity.

Well-managed galactosemic patients have a deficit of UDP-Gal (uridine diphosphogalactose; Ng et al., 1990), one of the products of GALT. While this deficiency has been questioned recently (Kirk- man, 1992), it is reflected in macromolecules, such as glycoproteins and glycolipids, that depend on UDP-Gal for their biosynthesis (Holton et al., 1990; Petry et al., 1991). The UDP-Gal deficiency and the galactosylation defect may be related to the long-term complications of galactosemia (Kaufman et al., 1988). Finally, from a genetic perspective it is noteworthy that GALT, which synthesizes UDP-Gal, is no more than 200 kb from the GGTB (p- 1,4-glycoprotein galactosyltrans- ferase) locus, which utilizes UDP-Gal as the donor of galactosyl moieties in the galactosylation reactions it catalyzes (Reichardt, 1989). This synteny may be of functional significance.

Three GALT protein polymorphisms, L62M, S135L, and N314D (Table 1; Fig. IB), have been described thus far (Reichardt and Woo, 1991; Reichardt et al., 1992b). They occur scattered throughout the protein and alter nonconserved residues. Thus, residues which were found noncrit- ical in evolution when mutated in people lead to benign polymorphisms.

GALACTOSEMIA VARIANTS AND RACIAL DIVERSITY

Many variants have been described in human GALT based on either aberrant electrophoretic mobility and/or altered enzymatic properties (Segal, 1989). In fact, one GALT protein polymorphism, N314D (Reichardt and Woo, 1991), has been described which is predicted to result in an overall charge change perhaps producing altered electrophoretic properties. Molecular characterization may help understand these variants bet- ter and may be clinically useful. The Duarte variant (D) of galactosemia is a common cause of a positive newborn screening test but generally does not require treatment. Therefore, genetic delinea- tion of this variant may help in diagnosing people with the D allele.

The “Black” variant of galactosemia is found commonly in African-Americans and it is charac- terized by a substantial residual ability to metabolize galactose (Segal, 1989). One galactosemia mu-

tation and one GALT polymorphism have been described in an African-American patient (FI 71s and S135L; Table I ; Reichardt et al., 199213). Both substitutions were also found in a Caucasian patient suggesting linkage and racial admixture in this particular situation. Racial admixture ac- counts for about one-third of all alleles in African- Americans (Reed, 1969). However, more patients from diverse backgrounds will have to be examined in order to allow definitive statements on ethnic and racial diversity in galactosemia.

PRENATAL DIAGNOSIS

Genetic analysis of galactosemia families have obvious implications for family planning decisions. Molecular data can and have been used for prenatal diagnosis (Reichardt and Packman, unpublished). The ongoing characterization of more mutations will further help in this endeavor.

Unfortunately no restriction fragment length polymorphisms (RFLPs) have been found in the human GALT gene (Reichardt, 1989) and no short tandem repeat (STRs) have been identified to date (Reichardt and Woo, unpublished). There- fore, prenatal diagnosis of galactosemia relies at present on testing for the common Q188R mutation (Table 1) and detection of novel mutations.

Characterization of the GALT gene, which has been reported in abstract form (Leslie et al., 1991), will be useful by enabling researchers to PCR amplify from genomic DNA and may uncover DNA polymorphisms which thus far have been elusive.

GENOTYPE-PHENOTYPE CORRELATIONS

The variable clinical outcome of galactosemia makes a prospective genotype-phenotype correlation particularly desirable. Such a study proved very rewarding in the case of phenylketonuria (PKU; Okano et al., 1991).

So far no clear correlation between biochemical parameters and clinical outcome has been found (Waggoner et al., 1990). However, the aforemen- tioned molecular heterogeneity may explain why genotyping of patients to categorize them might become necessary. Therefore, it is anticipated that the characterization of as many mutations as pos- sible will be necessary before a cooperative genotype-phenotype study can be initiated.

IMPLICATIONS FOR STRUCTURE-FUNCTION ANALYSIS OF GALT

Galactosemia missense mutations seem to highlight functionally or structurally important amino


acids in human GALT. In contrast, polymorphisms mark less critical residues (Fig. 1B). Thus, a detailed mutational analysis of human GALT in galactosemic patients may provide valuable in- sights into this interesting enzyme. Earlier it had been predicted that conserved residues would be useful in structure-function studies because they might pinpoint structurally or functionally important residues (Reichardt and Berg, 1988b). This prediction has been born out by the galactosemia experience and may provide a useful paradigm for the study of other housekeeping enzymes.

Work by Frey and co-workers has biochemically dissected the E. coli galT enzyme which is about 45% homologous to human GALT (Flach et al., 1990a). In particular the E. coli enzyme has a ping- pong reaction mechanism with a covalent uridy- late intermediate attached to histine 166 (Frey et al., 1982; Field et al. , 1988; Kim et al., 1990). A detailed biochemical analysis of human GALT will presumably follow the overall E. coli galT example and use a microorganismal overexpression and selection system (e.g., gal7 yeast).

CONCLUDING REMARKS

The molecular analysis of galactosemia has un- covered a multiplicity of mutations and polymorphisms (Table 1). The latter could be expected given the large number of normal and abnormal GALT variants described in the literature (Segal, 1989). The mutational diversity is noteworthy in light of the clinical heterogeneity. It also is anticipated that a detailed mutational analysis will be helpful in diagnosis and in a genotype-phenotype correlation. The residual activity associated with some galactosemia mutations can perhaps be en- hanced in patients to alleviate their long-term symptoms.

Mutations and protein polymorphisms highlight different domains in human GALT. CRM+ missense mutations, the most common form of galactosemia mutation, tend to occur in evolutionarily conserved areas which are presumably functionally and/or structurally important. Polymorphisms, in contrast, delineate less critical residues. Thus, a mutational analysis of galactosemia may pave the way to a detailed biochemical analysis of this interesting enzyme.

The high preponderance of missense mutations and the residual activity associated with some of them when overexpressed in cos cells are noteworthy. This may be due to the small gene size but may also lead to an interesting speculative question: is human GALT an essential gene? Clearly its ho-

mologues, galT in E. coli and GAL7 in yeast, are not essential genes since they can be deleted with- out adverse effects on the organism except in the presence of galactose as a nutrient (Segal, 1989). However, multicellular organisms such as humans are different from unicellular microorganisms since minimal GALT activity may be required to provide a threshold-type concentration of UDP-Gal, the substrate for galactosylation reactions. In fact, the low frequency of null mutations, splicing, and CRM- mutations supports this proposition since they can be balanced by either a normal allele in a carrier or by a mutant allele with residual activity in a galactosemia patient. The hypothesis of the essentiality of GALT will be addressed in suitable animal model systems such as Galt-deficient mice created by homologous recombination in embryonal stem (ES) cells. Such animals would also provide an invaluable tool for studying the pathophys- iology of the disease and the testing of novel therapies aimed at ameliorating the long-term complications of galactosemia.

ACKNOWLEDGMENTS

I am grateful to Savio Woo (Baylor) for his sup- port and encouragement over the years. I thank Yoshiyuki Okano (Osaka) and Claes Wadelius (Uppsala) for sharing unpublished information and Randy Eisensmith (Baylor) for critical reading of this manuscript. I was supported as an Associate of the Howard Hughes Medical Institute.

REFERENCES Banroques J , Schapira F, Gregori C, Dreyfus JC (1983) Molecular

studies on galactose- 1 -phosphate uridyltransferase from normal and mutant subjects. Ann Hum Genet 47:151-155.

Benn PA, Kelley RI, Mellman WJ, Amer AL, b c h e s FS, Markus HB, Nichols W, Hoffman B (1981) Reversion from deficiency of galactose-1-phosphate uridyltransferase (GALT) in an SV40-transformed human fibroblast line. Som Cell Genet 7: 667-682.

Cooper DN, Youssoufian H (1988) The CpG dinucleotide and human genetic disease. Hum Genet 78:151-155.

Dobbie JA, Holton JB, Clamp JR (1990) Defective galactosylation of proteins in cultured fibroblasts from galactosaemic patients. Ann Clin Biochem 27:274-275.

Field TS, Reznikoff WS, Frey PA (1989) Galactose-1-phosphate uridylyltransferase: Identification of histidine 164 and histidine 166 as critical residues by site-directed mutagenesis. Biochem- istry 28:2094-2099.

Flach JE, Reichardt JKV, Elsas LJ (1990a) Sequence of a cDNA encoding human galactose-1-phosphate uridyl transferase. Mol Biol Med 7:365-369.

Flach J , Dembure P, Elsas L (1990b) Transferase deficiency galactosemia: A molecular analysis. Am J Hum Genet 47:A607.

Kaufman FR, Xu YK, Ng WG, Donnell GN (1988) Correlation of ovarian function with galactose-1 -phosphate uridyl transferase activity in galactosemia. J Pediat 112:754-756.

196 REICHARDT

Kelley RI, Harris H, Mellman WJ (1983) Characterization of normal and abnormal variants of galactose-1-phosphate uridylyltransferase (EC 2.7.7.12) by isoelectric focusing. Hum Genet 63:274-279.

Kelley RI, Feinberg DM, Segal S (1989) Galactose-1-phosphate uridyl transferase in density-fractionated erythrocytes. Hum Genet 82:99-103.

Kim J , Ruzicka F, Frey PA (1990) Remodeling hexose-1-phosphate urdilylyltransferase: Mechanism-inspired mutation into a new enzyme, UDP-hexose-synthase. Biochemistry 29:10590- 10593.

Kirkman HN ( 1992) Erythrocytic uridine diphosphate galactose in galactosaemia. J lnher Metabol Dis 15:4-16.

Levy HL, Hammersen G (1978) Newborn screening for galactosemia and other galactose metabolic defects. J Pediatr 92: 871-877.

Leslie ND, Flach J , Florez M, Elsas L (1991) The cDNA and gene encoding human galactose-1-phosphate uridyl transferase (GALT). FASEB J 5:A2238.

Nadler HL, Chacko CM, Rachmeler M (1970) Interallelic com- plementation in hybrid cells derived from human diploid strains deficient in galactose- 1-phosphate uridyl transferase activity. Proc Natl Acad Sci USA 67:976-982.

Ng WG, Xu YK, Kaufman FR, Donnell GR (1989) Deficit of uridine diphosphate galactose in galactosaemia. J Inher Me- tabol Dis 12:257-266.

Okano Y, Eisensmith RC, Guettler FR, Lichter-Konecki U, Ko- necki D, Trefz FK, Dasovich M, Wang T, Henriksen K, Lou H, Woo SLC (1991) Molecular basis of phenotypic heterogeneity in phenylketonuria. N Engl J Med 324:1232-1238.

Petry K, Greinix HT, Nudelman E, Hakomori S, Levy HL, Rei- chardt JKV (1990) Characterization of a novel biochemical abnormality in galactosemia: Deficiency of glycolipids contain- ing galactose or N-acetyl galactosamine and accumulation of precursors in brain and lymphocytes. Biochemistry Med Me- tabol Biol 46:93-104.

Reichardt JKV (1989) Molecular biology of galactosemia. PhD Thesis, Stanford University, Stanford, CA.

Reichardt JKV (1991) Molecular analysis of 11 galactosemia patients. N ~ i d Acids Res 19:7049-7052.

Reichardt JKV, Berg P (1988a) Cloning and characterization of acDNA encoding human galactose- 1 -phosphate uridyl transferase. Mol Biol Med 5:107-122.

Reichardt JKV, Berg P (199813) Conservation of short patches of amino acid sequence between proteins with a common function but evolutionarily distinct origins: Implications for cloning genes and structure-function analysis. Nucl Acids Res 16: 9017-9026.

Reichardt JKV, Woo SLC (1991) Molecular basis of galactosemia: Mutations and polymorphisms in the gene encoding human galactose-1-phosphate uridylyltransferase. Proc Natl Acad Sci USA 88: 2633 -263 7.

Reichardt JKV, Packman S, Woo SLC (1991) Molecular characterization of two galactosemia mutations: Correlation of mutations with highly conserved domains in galactose-1-phosphate uridyl transferase. Am J Hum Genet 49:860-867.

Reichardt JKV, Belniont JW, Levy HL, Woo SLC (1992a) Char- acterization of two missense mutations in human galactose- 1 -phosphate uridyl transferase: Different molecular mechanisms in galactosemia. Genomics 12:596-600.

Reichardt JKV, Levy HL, Woo SLC (199210) Molecular characterization of two galactosemia mutations and one polymorphism: Implications for structure-function analysis of human galactose-1-phosphate uridyl transferase. Biochemistry 3 1: 5430-5433.

Rogers S, Holtzapple PG, Mellman WS, Segal S (1970) Charac- terization of galactose- 1-phosphate uridyl transferase in intes- tinal mucosa of normal and galactosemic individuals. Metab- olism 19:701-708.

Russell JD, DeMars R (1967) UDP-g1ucose:a-D-galactose-l-phos- phatetransferase activity in cultured human fibroblasts. Bio- chem Genet 1:ll-24.

Segal S (1989) Disorders of galactose metabolism. In Scriver CR, Beaudet AL, Sly WS, Valle D (eds): The Metabolic Basis of Inherited Disease. New York: McGraw-Hill, 453-480.

Wadelius C , Lagerkvist A, Larsson A, von Doebeln U, Petterson U (1992) Galactosemia caused by point mutation that acti- vates cryptic donor splice site in the galactose-I -phosphate uridyl transferase gene. Submitted.

Waggoner DD, Buist NRM, Donnell GV (1990) Long-term prog- nosis in galactosaemia: Results of a survey of 350 cases. J lnher Metabol Dis 13:802-818.

Documents

Genetic basis of galactosemia