Galactosemia - Scriver. MMBID 2006

Chapter 72: Galactosemia
PART 7: CARBOHYDRATES

Judith L. Fridovich-Keil, John H. Walter
Abstract
1. Galactosemia results from an impaired ability to metabolize galactose.

2. Humans encounter the monosaccharide galactose from both dietary and endogenous sources. The
predominant source of dietary galactose, especially for infants, is the disaccharide lactose, which is
abundant in milk and milk products. Lactose is hydrolyzed to its constituent monosaccharides, glucose
and galactose, prior to absorption from the intestine. Galactose is also found at significant levels in some
nondairy foods.
3. "Free" dietary galactose may exist in either the alpha or beta conformation. "Bound" dietary galactose is
released in the beta conformation from the breakdown of lactose, complex sugars, or glycoconjugates.
4. Once absorbed, dietary galactose may be converted into glucose-1-phosphate (Glc-1-P), as described
below, and metabolized to release energy. Alternatively, galactose may be converted into UDP-galactose
(UDP-Gal) and its derivatives, which serve as key substrate donors for the biosynthesis of glycoproteins
and glycolipids.
5. The conversion of β-D-galactose to Glc-1-P is a four-step process that has been conserved from bacteria
to mammals. In the first step, galactose mutarotase (EC 5.1.3.3) catalyzes the conversion of β-D-galactose
to α-D-galactose. In the second step, galactokinase (GALK, EC 2.7.1.6) phosphorylates α-D-galactose to
produce galactose-1-phosphate (Gal-1-P). In the third step, galactose-1-phosphate uridylyltransferase
(GALT, EC 2.7.7.12) transfers a UMP group from UDP-glucose (UDP-Glc) to Gal-1-P, releasing Glc-1-P
and producing UDP-Gal. In the final step, uridine diphosphate galactose 4′-epimerase (GALE, EC 5.1.3.2)
epimerizes the 4′ carbon of UDP-Gal to form UDP-Glc. There are autosomal recessive genetic diseases
resulting from impairment of each of the GALK, GALT, and GALE enzymes.
6. The biochemical consequences of impaired galactose metabolism are abnormally high concentrations of
galactose and its derivative metabolites in body tissues and fluids. Abnormal glycosylation of
glycoproteins and/or glycolipids also may occur. The severity of clinical features spans a broad range, with
outcome a function of which enzyme is impaired, the degree of functional impairment, and environmental
or other factors, many of which remain poorly understood.
7. The gene encoding human galactokinase (Reference Assembly NC_000017.9) maps to chromosome
17p24; it consists of 8 exons spanning about 7.3 kb of genomic DNA and shows many of the features of a
housekeeping gene. At least 25 different mutations have now been identified in the GALK loci of
galactokinase-deficient patients. The human enzyme includes 392 amino acids with a subunit molecular
mass of 42 kDa; structural coordinates as determined by x-ray crystallography are publicly available
(1WUU; Protein Data Bank).
8. Profound galactokinase deficiency occurs with a frequency of fewer than 1 in 100,000 live births. A
hallmark of galactokinase deficiency is cataracts that are usually bilateral and detectable in the early
weeks of life. Pseudotumor cerebri, or elevated intracranial pressure, also has been described in several
cases of galactokinase deficiency. The clinical symptoms of galactokinase deficiency generally will
self-resolve on dietary restriction of galactose.
9. The gene encoding human galactose-1-phosphate uridylyltransferase (Reference Assembly
NC_000009.10) maps to chromosome 9p13 and includes 11 exons that span about 4 kb of genomic DNA.
Copyright © The McGraw-Hill Companies, Inc. All rights reserved. -1- Any use is subject to the Terms of Use on www.ommbid.com.
The cDNA is approximately 1.3 kb in length and encodes a polypeptide of 379 amino acids with an
estimated subunit molecular mass of 44 kDa. The active enzyme is a homodimer of molecular mass 88
kDa. Structures have been solved for the Escherichia coli enzyme by high-resolution x-ray crystallography
and are publicly available (1GUP and 1GUQ, Protein Data Bank). More than 200 different mutations have
now been identified in the GALT loci of patients, although the functional significance of most mutations
remains unconfirmed. A few of these mutations are common, although most are rare. Of those mutations
that have been characterized in vitro or through whole-body galactose oxidation studies of homozygotes,
it is clear that some are true nulls, whereas others are hypomorphs.
10. Profound GALT deficiency, termed classic galactosemia, occurs with a frequency of approximately 1 in
30,000 to 1 in 60,000 live births, although this frequency varies over a wide range between among
geographic populations. Acute symptoms of classic galactosemia generally appear in the first weeks of life
and include poor feeding and weight loss, vomiting, diarrhea, lethargy, and hypotonia; liver dysfunction,
bleeding tendencies, cataracts, and septicemia also may occur. These symptoms generally will
self-resolve on stringent dietary restriction of galactose and may be prevented altogether by
presymptomatic diagnosis via newborn screening. Unfortunately, many affected individuals go on to suffer
serious long-term cognitive, female reproductive, and/or neurologic complications despite early detection
and careful intervention. Whether the long-term complications observed result from subtle developmental
abnormalities initiated in utero, chronic exposure to endogenously produced galactose, or other factors
remains unknown.
11. The gene encoding human uridine diphosphate galactose 4′-epimerase (GALE, Reference Assembly
NG_007068.1) maps to chromosome 1p36. The gene includes 12 exons that span about 4 kb of genomic
DNA. The cDNA is approximately 1.5 kb in length and encodes a polypeptide of 348 amino acids; the
predicted subunit molecular mass is 38 kDa. The active enzyme is a homodimer. At least 20 distinct
mutations have now been reported in the GALE loci of affected individuals. Structures of the human
epimerase enzyme have been solved by high-resolution x-ray crystallography and are available in the
Protein Data Bank (1ek5, 1ek6).
12. The population frequency of GALE deficiency remains unknown, largely owing to variable expressivity
and inconsistent newborn screening protocols, many of which are inadequate for identifying patients with
GALE deficiency. The majority of patients described with GALE deficiency appear to have an enzyme
defect restricted to their erythrocytes and circulating white blood cells; these patients are said to have
peripheral epimerase deficiency and are believed to be asymptomatic. A generalized, severe form of
GALE deficiency with a clinical presentation similar to classic galactosemia also has been described; this
condition is extremely rare. Finally, recent data demonstrate that even clinically well infants diagnosed
with hemolysate GALE deficiency may not be entirely peripheral in their enzyme impairment; these
individuals are said to have an intermediate form of GALE deficiency. The clinical significance of
intermediate GALE deficiency remains unclear.
BACKGROUND
Galactose-intolerant individuals were described in the medical literature as early as 1908. Since that time,
numerous publications have elaborated on the symptoms and signs of the disorder commonly referred to
as classic galactosemia, 54, 261, 393, 467 which results from profound impairment of the enzyme
galactose-1-phosphate uridylyltransferase (GALT). Other forms of galactosemia also have been
described; these can by caused by partial impairment of GALT or by impairment of one of the other
enzymes in the Leloir pathway of galactose metabolism, namely, galactokinase (GALK) or UDP-galactose
4′-epimerase (GALE).
Although all forms of galactosemia are the result of impaired galactose metabolism, clinical presentation
and severity vary widely among these patients. Furthermore, although neonatal diagnosis and lifelong
dietary restriction of galactose now can resolve or prevent the potentially lethal symptoms for patients
born in countries that offer the appropriate neonatal screening, the majority of classic galactosemia
patients in all countries nonetheless go on to experience serious long-term cognitive, neurologic, and/or
female reproductive complications. Despite decades of research, the underlying bases of pathophysiology
in galactosemia remain unknown, hampering the development of better intervention strategies.
SOURCES OF GALACTOSE
Dietary Sources of Galactose

The principal sources of galactose in most normal diets derive from the disaccharide lactose found in
mammalian milk and dairy products. When ingested, lactose is hydrolyzed rapidly to its constituent
monosaccharides, galactose and glucose, by a lactase enzyme contained within the brush border of the
small intestine. 161 Although galactose has been found free in many fruits and vegetables and in glycosidic
linkage in other plant and animal products (e.g., see refs. 2 and 166), these sources (or potential sources)
of galactose generally are considered insignificant when compared with the galactose load derived from
dairy products. 32
Galactose and glucose are absorbed from the intestinal lumen by a sodium-coupled, energy-dependent
system. 160 In normal individuals, galactose loads are absorbed and cleared from the blood very
rapidly. 428 Studies using 14 C- or 13 C-labeled galactose have shown that the label soon appears in the
glucose pool and subsequently in the CO 2 of expired air. 34– 36, 38, 39, 396 Indeed, the close correlation
between the rate of galactose clearance and hepatic blood flow, 457 coupled with the successful early use
of galactose tolerance as a liver function test, 428 indicated that the liver is a major organ responsible for
galactose metabolism in humans. This conclusion is especially important given that most clinical
diagnostic tests are performed on erythrocytes, lymphoblasts, fibroblasts, or amniotic fluid cells, which
may not be representative of the liver. This conclusion also may explain the observation that infants with
patent ductus venosus (PDV) account for a significant proportion of those identified by newborn screening
to have neonatal mild hypergalactosemia. 323
Endogenous Production of Galactose

Almost 40 years ago, Gitzelmann proposed the existence of a pathway for the endogenous synthesis of
galactose and galactose derivatives. 145 According to this model, UDP-glucose (UDP-Glc) is epimerized
by GALE to produce UDP-galactose (UDP-Gal), which is then cleaved by UTP-glucose/galactose
pyrophosphorylase (UGP; EC 2.7.7.10) to produce UTP and galactose-1-phosphate (Gal-1-P). Finally,
Gal-1-P is cleaved by inositol monophosphatase (IMPase) to yield free galactose. This pathway is
independent of the GALT enzyme and enables the production of galactose and essential
galactose-derivative molecules, such as UDP-Gal, under conditions or at developmental times when
environmental sources of galactose are insufficient. Indeed, over the last 15 years, the existence of an
endogenous pathway for the synthesis of galactose or galactose derivatives has been established beyond
any doubt. 31, 37, 319, 383 The magnitude of this endogenous galactose synthesis has been measured in
both normal individuals and GALT-deficient adults using stable isotope infusions. 31, 37, 319 In
GALT-deficient patients, the calculated production rate of galactose was between 0.49 and 1.09 mg/kg
per hour. 319 More recent studies of endogenous galactose production have corroborated these values but
have gone further to demonstrate that the rate of synthesis is age-dependent such that children
synthesize significantly greater amounts of endogenous galactose than do their adult counterparts. 31, 383
PATHWAYS OF GALACTOSE METABOLISM

Glucose is the primary metabolic fuel for humans and most other species. By comparison, galactose (Fig.
72-1) is less stable and is more susceptible to the formation of nonspecific glycoconjugates, perhaps
explaining the evolution and strong conservation of a pathway for the rapid conversion of galactose to
glucose. 62 In addition to the main Leloir pathway, described below, humans and many other species also
have a number of alternate pathways for the catabolism of galactose. These alternate pathways are
described at the end of this section.
The structures of galactose and glucose.
The Leloir Pathway
Conversion of β-D-galactose to α-D-galactose

Prior to further metabolism, the β-D-galactose derived from lactose or other "bound" sources first must be
converted to α-D-galactose. This reaction is catalyzed by the bidirectional enzyme mutarotase (EC
5.1.3.3), which was first isolated from Escherichia coli and characterized more than 40 years ago. 466
Mutarotase now has been observed in organisms ranging from bacteria and fungi to plants and
mammals. 438 Whereas most galactose mutarotase enzymes exist as distinct proteins, the mutarotase of
Saccharomyces cerevisiae is expressed as part of a fusion protein with UDP-galactose 4′-epimerase,
encoded by the gene GAL10. 436 Crystal structures for the human 438 and yeast 436 mutarotase enzymes
have been solved by x-ray crystallography and are available in the Protein Data Bank (1SNZ and 1SO0
and 1Z45).
Conversion of α-D-galactose to glucose-1-phosphate

Once in the alpha conformation, D-galactose is converted to glucose-1-phosphate via the sequential
reactions of a three-step pathway elucidated initially in yeasts and bacteria 65, 257 and named after Leloir,
who made a major contribution to this work (Fig. 72-2). The Leloir pathway consists of three enzymes:
galactokinase (GALK, EC 2.7.1.6), which phosphorylates α-D-galactose to produce Gal-1-P;
galactose-1-phosphate uridylyltransferase (GALT, EC 2.7.7.12), which transfers a UMP group from
UDP-glucose to Gal-1-P, thereby releasing Glc-1-P and forming UDP-Gal; and uridine diphosphate
galactose 4′-epimerase (GALE, EC 5.1.3.2), which interconverts UDP-Gal and UDP-Glc. Human GALE
also interconverts UDP-N-acetylgalactosamine (UDP-GalNAc) and UDP-N-acetylglucosamine
(UDP-GlcNAc). Combined, the Leloir enzymes convert 1 mole of α-D-galactose + 1 mole of ATP into 1
mole of Glc-1-P + 1 mole of ADP. 189 Along the way, the enzymes of this pathway also establish and
maintain intracellular levels and ratios of UDP-Glc/UDP-Gal and UDP-GalNAc/UDP-GlcNAc, all of which
are essential substrates for the biosynthesis of glycoproteins and glycolipids in humans. 11, 370
The metabolic pathways of galactose. The main pathway of galactose metabolism, from α-D-galactose to
glucose-1-phosphate, involves three enzymes (blue), galactokinase (GALK), galactose-1-phosphate
uridylyltransferase (GALT), and uridine diphosphate galactose-4′-epimerase (GALE). The
pyrophosphorylase (UGP) pathway is believed to operate both as a GALT-independent bypass pathway
for the conversion of Gal-1-P to UDP-Gal and also in the reverse direction (involving UGP, GALE, UGP
again, and IMPASE) to enable the endogenous synthesis of galactose from Glc-1-P and UTP. Human
GALE also interconverts UDP-GalNAc and UDP-GlcNAc. UDP-Gal, UDP-Glc, UDP-GalNAc, and
UDP-GlcNAc all serve as key sugar donors for the synthesis of glycoproteins and glycolipids; the turnover
of these compounds also contributes to the galactose pool. Particularly when its levels are increased,
galactose may be metabolized to galactitol by aldose reductase (AR) and to galactonic acid by galactose
dehydrogenase (GDH).
Although the liver is the major organ of galactose metabolism in humans, enzymes of the Leloir pathway
have been detected in many cell types and tissues, including red and white blood cells, fibroblasts, and
amniocytes. Indeed, deficiency of GALT and accumulation of Gal-1-P in red blood cells are the most
common points of data ascertained to confirm a diagnosis of classic galactosemia. 54, 205, 390 Gal-1-P
uridylyltransferase and GALK are also present in fetal red blood cells, liver, lung, spleen, and cardiac
muscle as early as 10 weeks’ gestation, if not sooner, and their activities are higher in the second and
third trimesters of gestation than at any time postnatally. 410 These data reaffirm the importance of
galactose metabolism in the developing fetus and raise the troubling possibility that the postnatal
abnormalities observed in patients with profound impairment of GALT may trace their origins to subtle
abnormalities of development initiated in utero.
Galactokinase
GALK has been studied genetically, biochemically, and structurally (e.g., see refs. 399, 439, and 441)
from a variety of sources ranging from microbial to human. While the gene that encodes GALK may be
unique in some genomes, in others it coexists with closely related sequences of apparently distinct
function. For example, in the yeast S. cerevisiae, GALK is encoded by the GAL1 gene, but a highly
conserved gene, GAL3, encodes a protein that is catalytically inactive and serves as a galactose sensor
and transcriptional inducer. 185 Recent structural and mutational studies 346, 441 demonstrate that a
two-residue change in Gal3p restores galactokinase activity.
In humans, the GALK sequence also exists as a gene pair, called GALK1 (GK1) and GALK2 (GK2).
GALK1 maps to chromosome 17q 21-22 115, 423 and encodes the enzyme that is most significant in terms
of GALK activity 5 and that is mutated in patients with clinical galactokinase deficiency
galactosemia. 142, 197, 375, 423, 446 The gene GK2 maps to chromosome 15 and also encodes a catalytically
active product, 256 although the preferred substrate of the GK2-encoded enzyme is N-acetylgalactosamine
rather than galactose. 256 The human GK2-encoded enzyme has 458 amino acids and a relative
molecular mass (M r ) of 50.386. 256 This GK2-encoded enzyme shows approximately 35 percent amino
acid sequence identity to the human GALK1 gene product.
The Human GALK (GK1) Gene and cDNA

The full-length human GALK1 gene consists of eight exons and spans about 7.3 kb of genomic DNA
(GenBank L76927). 29 A modified genomic sequence that includes substitutions in the 5′ UTR and in
introns 1, 2, and 5, as well as a 20-bp tandem repeat (ATTCTCCTGCCTCAGCCTCC) in three places in
intron 5, also has been identified (GenBank AF084935). 16 The human GALK1 promoter shares many
features in common to other housekeeping genes, including the absence of TATA- and CCAAT-box
motifs, a high GC content (which in part reflects potential binding sites for the Sp-1 transcription factor),
and evidence of multiple transcription initiation sites. 29 Indeed, the GALK1 mRNA is heterogeneous at the
5′ terminus, with transcription start sites that map to many points between 21 and 61 bases upstream of
the ATG translation initiation codon. 29 The organization of the human galactokinase gene is presented in
Table 72-1
Table 72-1: Exon/intron organization of the human galactokinase (GALK) gene
Exon
Number of Nucleotide number from Amino acid
Region border
nucleotides first Met (ATG) in cDNA numbers b
type a
exon
179–227 c 1–165 3/ 1–55
1
intron 1 886
exon
190 166–355 1/ 56–119
2
intron 2 459
exon
120 356–475 1/ 119–159
3
intron 3 170
exon
136 476–611 2/ 159–204
4
intron 4 127
exon
182 612–793 1/ 204–265
5
intron 5 4107
exon
151 794–944 2/ 265–315
6
intron 6 76
exon
163 945–1107 3/ 315–369
7
intron 7 82
exon
189 1108–1179 3/ 370–392
8
Total 1179 392
a The border type describes whether the exon ends after the first (1/), second (2/), or third (3/)
nucleotide of the codon.
b The amino acid number is included in two exons when the border type is 1 or 2.
c This suggests multiple starting sites.
Submitted under GenBank accession number L76927 66
The human GALK1 cDNA is approximately 1.35 kb in length and encodes a protein product of 392 amino
acids. Alignment of the predicted human GALK protein with galactokinases from other organisms 4, 77, 101
shows only low to moderate overall sequence homology; the greatest conservation, at almost 45 percent,
is with the E. coli sequence. 101 Analysis of the predicted protein sequence shows a conserved
galactokinase signature sequence in exon 1 that may be required for function and separate ATP-binding
motifs in exons 3 and 7 that are also conserved among all galactokinases. 5, 101
Structure and Catalytic Mechanism of Human GALK

The predicted protein product of GK1 is a polypeptide of 392 amino acids with a monomer molecular
mass of about 42 kDa. 423 Early work on human and other galactokinase proteins provided mixed and
sometimes conflicting results with regard to its monomeric versus dimeric nature and possible
tissue-specific or age-dependent changes in the catalytic properties or structure of the
enzyme. 47, 90, 285, 311, 410, 411, 425, 465 Early studies also demonstrated that the phosphorylation reaction
catalyzed by yeast GALK is reversible, although the equilibrium point markedly favors the production of
the product, Gal-1-P. 18 Studies in rat liver 90 and human red cells 285 further indicated that GALK is
inhibited by both its substrate and its product, which may serve to limit the accumulation of Gal-1-P in
GALT or GALE deficiency. Finally, fractionation of red blood cells from a patient with a new variant of
GALK deficiency revealed higher levels of enzyme activity in reticulocytes than in older red blood cells, 18
implying instability of the mutant protein and suggesting a mechanism to explain potential variations in
GALK levels between different tissues, at least in some patients.
Recently, a combination of biochemical and site-directed mutagenesis studies coupled with

high-resolution x-ray crystal structures for both microbial and human GALK proteins has offered detailed
insights into the substrate specificity, mechanism, and architecture of galactokinase. For example,
Thorson and colleagues performed a broad study of substrate selectivity of E. coli galactokinase 498 and
subsequently identified residues within the E. coli enzyme that mediate substrate specificity. 499 Sellick
and Reece similarly studied determinants of substrate specificity in the yeast GALK enzyme (Gal1p), 399
and Timson and Reece 447 applied a combination of biochemical studies with site-directed mutagenesis
approaches to explore the basis of substrate selectivity in human GALK. From a comparison of these
reports, it is clear that the microbial and human GALK enzymes are similar but not identical in terms of
their substrate selectivities.
Human GALK phosphorylates the C-1 hydroxyl group of α-D-galactose in an MgATP-dependent manner.
The reaction mechanism appears to be sequential, such that MgATP binds first, followed by
α-D-galactose; both substrates bind before either product (ADP and Gal-1-P) is released. 190, 446 This
apparent order of substrate binding to GALK is not conserved across all species (reviewed in ref. 190).
Structurally, human galactokinase (Fig. 72-3) is recognized as a member of the GHMP (galactokinase,
homoserine kinase, mevalonate kinase, and phosphomevalonate kinase) superfamily of small molecule
kinases 439 ; these kinases contain three common motifs and are structurally distinct from the
well-characterized P-loop-containing kinases, the actin-like ATPases, and the ATP-grasp enzymes
(reviewed in ref. 190). While the monomeric versus dimeric quaternary structure of the soluble human
enzyme remains unclear, the crystal structure suggests that it is a homodimer. In fact, a disulfide bond
appears to link the two subunits, although this bond may be an artifact of the crystallization conditions. 439
Further studies with the soluble enzyme will be required to clarify this point.
Structure of human GALK defined by x-ray crystallography. (Courtsey of H. Holden.)
Galactose-1-phosphate uridylyltransferase (GALT)

GALT, the second enzyme in the Leloir pathway, has been studied extensively in everything from
microorganisms 127, 128, 134, 245 to mammalian cells and tissues. 40, 95, 485 The human GALT monomer has
a predicted molecular mass of 44 kDa; the active enzyme is a homodimer of 88 kDa. Although the crystal
structure of the human enzyme has yet to be reported, structural data from the E. coli enzyme are
available (Protein Data Bank 1GUQ) 476, 477 and have been used to draw structural conclusions for the
human enzyme by homology modeling.
The Human GALT Gene and cDNA

In contrast to the case of GALK, all data indicate that there is only one gene encoding GALT in humans
and other organisms. 410 After some conflicting early reports regarding chromosome location, the GALT
locus was correctly assigned to chromosome 9 24, 296– 298 and localized to the short arm in the region
9p13. 93, 403, 420 The human GALT cDNA 123, 360 and gene 262 have been cloned and characterized.
The human GALT gene is arranged into 11 exons (Fig. 72-4) spanning 4.3 kb of sequence 262 ; the GALT
cDNA is 1295 bp in length. 123, 360 Comparison of the human with the E. coli, 258 yeast, 433 and mouse (N.
Leslie, GenBank Accession No. M96265) GALT amino acid sequences revealed overall identities of 46,
39, and 87 percent, respectively. 262 Some exons, notably 6, 9, and 10, have been highly conserved,
whereas others, such as exons 1, 2, 5, and 7, have been poorly conserved. Of note, the active-site
His-Pro-His triad is encoded in exon 6. The promoter region of the human GALT gene contains two
GC-rich Sp-1 sites, three AP-1 sequences, and a CCAAT sequence. There is no consensus TATA box.
Thus, despite the highly inducible nature of GALT expression in microbes (e.g., see ref. 214), the
mammalian gene appears to be expressed as a "housekeeping" gene. Indeed, studies of the mouse
GALT gene promoter 259 demonstrated no clear evidence of induction by galactose.
Structure and organization of the human GALT gene. Relative sizes of introns and exons are drawn to
scale.
A number of studies, however, have suggested evidence of regulated GALT expression or activity, in
humans and/or in rodents. For example, in one early paper Pesch and colleagues 342 reported three
prepubertal boys with classic galactosemia who were given progesterone, 10 to 20 mg/day for 6 days,
resulting in up to 10-fold enhancement of galactose oxidation, as measured by their ability to convert
[1- 14 C]galactose to 14 CO 2 . The same study showed that the formation of cataracts in young male rats
fed a high-galactose diet could be delayed by the administration of progesterone. Similarly, Rogers and
Segal 368 have shown that GALT activity can be increased by administration of folic acid. Seven-day-old
suckling rats were given 1 mg/d folic acid for the next 7 days. Liver perfusion experiments carried out after
folic acid administration showed a very rapid rate of galactose uptake for 35 minutes compared with
Copyright © The McGraw-Hill Companies, Inc. All rights reserved. - 10 - Any use is subject to the Terms of Use on www.ommbid.com.
control rats. GALT activity also was shown to be elevated in the livers of test rats. These experiments
have hinted at therapeutic possibilities, particularly in patients who have some residual GALT activity, but
they have not been pursued. Finally, more recent studies have reported evidence of tissue-specific or
developmental-specific differences in GALT expression in rodents, 174, 175 although the mechanism and
significance of these differences remain unclear.
Structure and Catalytic Mechanism of Human GALT

High-resolution x-ray crystallography of the E. coliGALT enzyme (Fig. 72-5) reveals the tight association
of subunits in the homodimer and further confirms the existence of a distinct active site in each
subunit. 476, 477 Biochemical and crystallographic studies also demonstrate that GALT is a Zn/Fe
metalloprotein in which the metal ions stabilize the structure (reviewed in ref. 127) but do not contribute to
catalysis.
Structure of E. coli GALT defined by x-ray crystallography. Blue spheres represent zinc; green spheres
represent iron. (Courtesy of H. Holden.)
The reaction catalyzed by GALT involves the transfer of a UMP group from UDP-glucose to Gal-1-P,
resulting in the formation of UDP-Gal and the release of Gal-1-P (see Fig. 72-2). The reaction
demonstrates "ping-pong," or double-displacement, kinetics, 284 and this assertion has been confirmed by
the combined research of many workers, summarized by Frey. 127 In the first half-reaction, UDP-Glc
enters the active site, a covalent bond is formed between an active-site histidine (His166 in E. coli GALT,
His186 in human GALT) and UMP, and Glc-1-P is released. In the second half-reaction, Gal-1-P enters
the active site, and UMP is transferred from covalent attachment to the enzyme to covalent attachment to
the Gal-1-P, forming UDP-Gal, which is then released.
Although the normal GALT protein appears homogeneous when viewed under denaturing conditions,
native isoelectric focusing studies of human GALT isolated from red blood cells or expressed in
recombinant form in yeast produce a very different picture; at least six bands in the pH range from 5 to 6
are seen by activity overlay stain. 130, 395 Two-dimensional gel separation studies of both wild-type and
mutant forms of human GALT revealed that at least some of this heterogeneity reflects the presence of
significant amounts of covalent UMP-bound enzyme intermediate 179 involving one or both subunits.
The subcellular localization or sequestration of GALT also was explored by Christacos and colleagues, 76
who covalently tagged endogenous yeast GALT with green fluorescent protein (GFP) and noted clear
enrichment of the protein in a small number of "spots" within the cytoplasm. Moreover, this sequestration
was observed only when all three Leloir enzymes were present and actively metabolizing galactose.
Intrinsic determinants of localization were believed to reside within the amino-terminal 134 residues of the
yeast GALT enzyme, Gal7p. Similar experiments expressing fluorescently tagged human GALT in yeast
demonstrated a similar result, indicating that at least some determinants of localization have been
conserved through evolution. These data provide compelling evidence that GALT may function in a
sequestered state in cells, perhaps explaining the relatively high substrate K m values estimated for the
purified enzyme. Clearly, there are clinical implications if some mutations at the GALT locus result not in a
reduction in enzyme activity but rather in aberrant protein interactions through impaired subcellular
localization. Such alleles might fail to function in vivo, although they might appear perfectly functional in
vitro.
Uridine Diphosphate Galactose-4′-Epimerase (GALE)

The third and final enzyme of the Leloir pathway is GALE, which functions as a homodimer to catalyze the
interconversion of UDP-Gal and UDP-Glc. 189, 442– 444 GALE is an extraordinarily important enzyme not
only because of its role in the Leloir pathway but also because it enables the endogenous synthesis of
UDP-Gal in the absence of exogenous sources of galactose. Human GALE also catalyzes the
interconversion of a pair of larger substrates, UDP-GalNAc and UDP-GlcNAc. 280, 345, 387 Notably, while
humans and some other species encode a single epimerase enzyme capable of interconverting both
UDP-Gal/UDP-Glc and UDP-GalNAc/UDP-GlcNAc, other species, including E. coli 470 and some
pathogenic microbes 64 encode these activities in separate enzymes. Like GALK and GALT, GALE is well
conserved from microbial to mammalian species. The human GALE gene 276 and cDNA 98 have been
cloned and characterized, and the human GALE enzyme has been studied both biochemically and
crystallographically. 189, 442, 443
The Human GALE Gene and cDNA

The human GALE gene was mapped to a single locus on chromosome 1 using somatic cell hybrids 25, 269 ;
fluorescent in situ hybridization subsequently localized the gene more precisely to the short arm at 1p36.
The full-length genomic GALE gene was cloned and characterized by Maceratesi and colleagues, 276 and
the sequence was submitted to GenBank under Accession No. AI022382. A corrected sequence can be
found at GenBank Accession No. DQ233667. 329 Like GALT, the human GALE gene consists of 11 exons
extending over about 4 kb of genomic DNA (Table 72-2). The first exon contains the 5′-untranslated
region, and the last exon harbors the last 54 bases of the protein-coding sequence, the termination codon,
and the 3′-untranslated sequence. Interestingly, at the boundary between the first exon and the first intron,
the splice-site donor sequence is GC and not the usual consensus GT sequence.
Table 72-2: Exon/intron organization of the human UDP-Galactose 4’-epimerase (GALE) gene
Nucleotide number Exon

Number of Amino acid
Region from first Met (ATG) in border
nucleotides numbers b
cDNA type a
exon
70 (–93)
1
intron 1 350–355
exon
121 1–121 1/ 1–41
2
intron 2 156
exon
116 122–237 3/ 41–79
3
intron 3 385
exon
114 238–351 3/ 80–117
4
intron 4 245
exon
177 352–528 3/ 118–176
5
intron 5 544
exon
114 529–642 3/ 177–214
6
intron 6 90
exon
67 643–709 1/ 215–237
7
intron 7 93
exon
86 710–795 3/ 237–265
8
intron 8 137
exon
78 796–873 3/ 266–291
9
intron 9 235
exon
120 874–993 3/ 292–331
10
intron
144
10
exon
401 994–1097 332–348
11
intron
1098–1395
11
Total 1488 c 348
a Theborder type describes whether the exon ends after the first (1/), second (2/), or third (3/)
nucleotide of the codon.
b The amino acid number is included in two exons when the border type is 1 or 2.
c Including the 5-untranslated region of exon 1.
Submitted under GenBank accession number A 1022382 104
The cDNA, originally cloned by Daude and colleagues, 98 is 1488 bp in length and predicts a protein of
348 amino acids with a molecular mass of 38 kDa. A corrected sequence for the human GALE open
reading frame can be found at GenBank Accession No. DQ233668. 329 The predicted human GALE
protein is highly conserved, demonstrating 87 percent sequence identity to rat GALE, 53 percent identity
to Kluyveromyces lactis GALE, and 51 percent sequence identity to E. coli GALE. 258
Although there is little information currently available about the regulation of GALE expression in
mammals, it would appear that red blood cell GALE concentrations are higher in newborn infants than in
adults. 28 Consistent with these results, Cohn and Segal 80 also found that GALE activities were highest in
postnatal rats before 2 days of age, with levels declining to adult values by 20 days.
Structure and Catalytic Mechanism of Human GALE

High-resolution x-ray crystallography of the human GALE enzyme 442, 443 (Fig. 72-6), coupled with
biochemical data 3, 279, 287, 308 (reviewed in refs. 127 and 189), has enabled unambiguous definition of the
mechanism of catalysis, which can be described as a three-step process: transfer of the 4′-hydroxyl
hydrogen from UDP-Gal to NAD+, rotation of the resulting 4′-ketopyranose intermediate into the active
site, and return of the hydride from NADH to reprotonate the C-4 oxygen of the sugar. 189 This mechanism
is reversible.
Structure of human GALE defined by x-ray crystallography. Stick images represent NAD+ and substrate.
(Courtesy of H. Holden.)
While the overall function and mechanism of GALE appear to be conserved across species, there are
important differences. For example, the epimerases of E. coli 486 and yeast 96 appear to bind NAD+ much
more strongly than do their mammalian counterparts. 279, 287 Comparison of available crystal structures
offers a logical explanation for this observation: There are 19 hydrogen bonds linking the dinucleotide to
E. coli GALE, whereas for the human enzyme there are only 11. 442
Of note, there are interesting developmental changes in the availability of endogenous NAD+ in human
red blood cells that have implications with regard to the assay of GALE. 28 GALE activity in hemolysates
from adults is negligible unless NAD+ is added. In contrast, GALE activity levels appear optimal in
newborn hemolysates even without exogenous cofactor. The reason is that in adults there is a stromal
nucleosidase that destroys NAD+ when the red blood cells are lysed; this nucleosidase is not present in
the red blood cells of newborn children. Other factors, such as alcohol treatment, that affect the level of
available NAD+ in cells or tissues scan imilarly influence the detection of GALE activity or galactose
oxidation. 203
The most significant difference between the human and E. coliGALE enzymes is in their substrate
specificities; UDP-GalNAc and UDP-GlcNAc are substrates for mammalian GALE, 280, 345, 387 but not for
the E. coli or yeast enzymes. 374 The importance of this aspect of GALE function in mammalian cells was
illustrated originally by M. Krieger and colleagues, 229, 230 who noted that a GALE-deficient line of Chinese
hamster ovary (CHO)–derived cells (ldlD) was unable to synthesize properly glycosylated low-density
lipoprotein (LDL) receptors unless their medium was spiked with bothgalactose and
N-acetylgalactosamine (Gal-NAc). Normal CHO cells did not require galactose or Gal-NAc
supplementation presumably because they were able to synthesize their own UDP-Gal and UDP-GalNAc
endogenously via GALE.
The structural basis for the difference in substrate specificity between the E. coli and human GALE
enzymes was identified by H. Holden and colleagues, 443 who noted that the active-site volume for the
human protein is about 15 percent larger than that observed for the bacterial epimerase. This extra space
presumably enables the larger substrate (UDP-GalNAc) to fit into the active-site cleft and/or allows the
larger 4′-ketopyranose intermediate to rotate more easily during catalysis. Holden and colleagues further
noted that the difference in active-site size between the human and E. coli GALE enzymes was largely
attributable to a single-amino-acid difference—Tyr299 in the E. coli protein is replaced in the human
enzyme with a cysteine residue (Cys307). This assertion was confirmed by a set of parallel experiments in
which substitution of cysteine for tyrosine at amino acid position 299 of the E. coli protein did confer the
ability to interconvert UDP-GalNAc/UDP-GlcNAc 443a , and similarly, substitution of tyrosine in place of
cysteine at the comparable position in human GALT (C307Y) dramatically impaired ability of the mutant
enzyme to interconvert UDP-GalNAc/UDP-GlcNAc but not UDP-Gal/UDP-Glc. 387
Accessory Pathways of Galactose Metabolism

While the Leloir pathway is clearly the predominant route for galactose metabolism in humans and other
species, it is not the only route (see Fig. 72-2). Studies in microorganisms, mice, and patients all provide
convincing evidence for the existence of routes of galactose metabolism that bypass GALK and/or GALT
and/or GALE. For example, the reduction of galactose to galactitol and the oxidation of galactose to
galactonate require none of the three Leloir enzymes. Alternatively, some other routes may bypass one of
the Leloir enzymes but not the other two. In particular, there has been a great deal of interest in routes of
galactose metabolism that bypass GALT, thereby circumventing the metabolic block that results from the
absence or impairment of GALT activity in patients with classic galactosemia.
Indeed, Berry and colleagues 37 first demonstrated the existence of a GALT-independent route of
galactose metabolism in humans when they applied an intravenous and oral [ 13 C]galactose breath test to
detect low levels of galactose oxidation in patients with classic galactosemia. Patients demonstrating no
detectable GALT activity were nonetheless able to eliminate almost 4 percent of a bolus of galactose as
expired CO 2 within a 5-hour period compared with 21 to 47 percent in controls. Similarly, an individual
completely devoid of GALT owing to homozygosity for a large gene deletion was able to metabolize 17
percent of an oral bolus of galactose within a 24-hour period, a value comparable with the 3-hour
elimination rate in a control subject. 35 More recent studies have both confirmed and extended these
conclusions. 39 The use of Epstein-Barr virus (EBV)–transformed lymphoblasts from galactosemic
subjects 35, 496 and GALT knockout mice have further confirmed the existence of GALT-independent
routes of galactose oxidation in mammals. 320, 321 The question of the mechanism of GALT-independent
galactose metabolism has been addressed by studies using yeast and patient cells in culture, as
explained below.
Reduction of Galactose to Galactitol

When galactose accumulates in the blood or tissues of patients with galactosemia, it may become a
substrate for the enzymes that catalyze the first stage of the so-called polyol pathway of carbohydrate
metabolism. 181, 326 This first reaction in this pathway, from the hexose to the corresponding hexitol, may
be catalyzed by either of two enzymes, aldose reductase (polyol NADP+ oxidoreductase, EC 1.1.1.21) 172
or L-hexonate dehydrogenase (L-glucuronate: NADP+ oxidoreductase, EC 1.1.1.19). 326 All current data,
explained below, point toward aldose reductase as the enzyme principally responsible for the direct
reduction of galactose in mammals.
Galactitol, the product of the reduction of galactose by aldose reductase (Fig. 72-7), is not a substrate for
the subsequent enzyme in the polyol pathway and, consequently, may accumulate in tissues and/or be
excreted in the urine. Indeed, abnormally high blood and/or urinary concentrations of galactitol have been
found in patients with classic galactosemia, galactokinase deficiency, and epimerase deficiency. 208, 209
Although the reduction of galactose to galactitol does not appear to offer a significant means of eliminating
galactose in the absence of GALT, it is nonetheless significant to galactosemia because of the negative
clinical consequences attributed to galactitol accumulation, at least in some tissues (discussed below).
The conversion of galactose to galactitol by a nonspecific aldose reductase and to galactonic acid by
galactose (aldehyde) dehydrogenase.
Oxidation of Galactose to Galactonate

Galactose that accumulates in the blood or tissues of patients with galactosemia also may be oxidized to
form galactonate (see Fig. 72-7). Indeed, patients with GALT deficiency given a 35 g/m 2 load of galactose
have been shown to excrete six times as much galactonate in their urine as normal subjects exposed to
the same experimental conditions. 27 Similarly, a postmortem liver sample from a patient with classic
galactosemia contained an elevated level of galactonate compared with control liver samples. 356 Further,
genetically normal rats 356 and guinea pigs 462 fed experimental diets containing large amounts of
galactose accumulated galactonate in a number of tissues and excreted increased levels of the metabolite
in urine. More recently, investigators 478 have used nuclear magnetic resonance spectroscopy 478 and gas
chromatography–mass spectroscopy 381, 497 to detect and quantitate galactonate in the urine of
galactosemic infants and adults on a galactose-restricted diet.
That galactonate forms in patients with galactosemia appears well established; however, how that
galactonate forms remains controversial. The oxidation of D-galactose to form D-galactonate is believed to
reflect the activity of a soluble NAD+-dependent galactose dehydrogenase (D-galactose: NAD+
oxidoreductase, EC 1.1.1.48) that converts D-galactose to D-galactonolactone, which then spontaneously
or enzymatically converts to galactonate. An enzyme with the anticipated activity reportedly was purified a
hundredfold from rat liver. 91, 92 However, the existence of this apparently highly active, soluble,
NAD+-dependent galactose dehydrogenase has been questioned as a possible artifact. 401, 421 Once
formed, galactonate may be futher metabolized via the reactions of the pentose phosphate pathway (Fig.
72-8).
An alternative oxidative pathway of galactose metabolism via galactonate.
The Pyrophosphorylase Pathway

In 1957, Isselbacher postulated the existence of a GALT-independent pathway of galactose metabolism in
humans. 204 The alternative pathway in humans was based on prior work performed in yeast by Kalckar
and colleagues 218 and was proposed as an explanation for the perceived ability of some patients with
classic galactosemia to tolerate more dietary galactose with age. 449 The proposed GALT-bypass pathway
(see Fig. 72-2) consists of four steps. First, α-D-galactose is phosphorylated by GALK to form Gal-1-P.
Next, a UTP-dependent glucose/galactose pyrophosphorylase (UGP), also sometimes referred to as
UTP-hexose-1-phosphate uridylyltransferase (EC 2.7.7.10), catalyzes the production of UDP-Gal (and
PP i ) from Gal-1-P and UTP. Third, UDP-Gal is converted to UDP-Glc by GALE, and finally, UDP-Glc is
recombined with PP i by the reversible UGP to generate Glc-1-P and UTP.
The existence of the pyrophosphorylase, or UGP, pathway is supported by a preponderance of recent

evidence from both yeast and mammalian cell studies. For example, Mehta and colleagues 289 and later
Lai and Elsas 246 demonstrated that overexpression of UGP in yeast deficient for GALT rescued the ability
of these cells to grow on medium containing galactose. In addition, Ross and colleagues 369 demonstrated
that whereas yeast deficient in GALT nonetheless could metabolize galactose slowly, yeast deficient in
GALK or GALE could not. This result not only demonstrated the existence of a GALT-independent
pathway of galactose metabolism in yeast but also further confirmed that both GALK and GALE are
essential components of that pathway.
A number of studies using mammalian cells in culture similarly have confirmed the existence of a
UGP-dependent GALT-bypass pathway in mammals. For example, Lai and colleagues 249 demonstrated
that GALT-deficient patient cells that were unable to grow in medium containing 0.1% galactose in place
of 0.1% glucose were "rescued" when transfected to overexpress human UGP. Schulz and colleagues 386
similarly noted that mammalian cells devoid of GALE activity were unable to deplete exogenous galactose
from their medium, illustrating the essential role of GALE in both the Leloir and GALT-independent
pathways of galactose metabolism.
The specificity and capacity of the UGP enzyme with regard to galactose metabolism have been studied
in a variety of species, tissues, and contexts. 202, 235, 408, 451 Pyrophosphorylase activity has been detected
in human adult and fetal liver, kidney, and skeletal muscle. 408 This same activity also has been detected
in the tissues of both wild-type and GALT-null mice. 260 A study of the rate of galactose
pyrophosphorylation in human liver from newborn infants and adults of all ages suggests that the
pyrophosphorylase pathway can metabolize galactose at a rate only about 1 percent of that of the Leloir
pathway. 1 Nonetheless, in the absence of a functional GALT enzyme, this relative trickle may become
significant.
GALK DEFICIENCY
Diagnosis: Acute Clinical Presentation, Enzyme Defect, and Metabolic Abnormality

Partial to complete deficiency of GALK activity (OMIM 230200) is observed as a rare autosomal recessive
condition in the live-born population (reviewed in ref. ). The reported frequencies of GALK deficiency vary
by study and location; for example, a population-based study in the United States detected a frequency of
approximately 1 in 1 million live births, 264 whereas individuals of Romani descent in Europe demonstrate
a frequency closer to 1 in 10,000 live births. 197
Acute clinical presentation

The only consistent clinical finding in GALK deficiency is cataracts. 12, 55, 82, 94, 143, 226, 266, 325, 344, 459 The
cataracts detected in GALK-deficient patients are characteristically bilateral and have been reported in
GALK-deficient children as young as 4 weeks of age. 226 Although other abnormalities have been
described in individual patients, including macular deposits, 186 mental retardation, 398 complement
deficiency, 53, 398 and seizures with neurologic deterioration, 344 these are most likely coincidental.
Pseudotumor cerebri has been reported in more than one infant. 81, 270 Since pseudotumor cerebri has
been postulated to occur by the same mechanism responsible for cataract formation, 199 it is possible that
this is a true, albeit rare, complication of GALK deficiency.
In populations under newborn surveillance for high blood galactose concentration, GALK deficiency also
may be detected presymptomatically through investigation of an abnormal blood galactose screening
result accompanied by normal GALT and GALE enzyme activities. Notably, newborn surveillance
programs that test blood galactose only, in follow-up to an abnormal GALT enzyme activity level, will not
detect GALK deficiency.
Enzyme defect
The level of GALK activity detected in patients may vary, ranging from undetectable to significant residual
activity. 340 Notably, normal hemolysate GALK activity levels also vary with age, such that normal values
for newborns range from 80 to 120 nmol/min per gram of hemoglobin, whereas for children 1 year of age
or older, the normal range is 20 to 30 nmol/min per gram of hemoglobin. 340 A carefully age-matched
control group therefore must be considered when calculating the degree of enzyme impairment in a
suspected patient. Although blood is generally the tissue investigated clinically, a whole-body galactose
oxidation study performed by Gitzelmann and colleagues 156 demonstrated that a patient with no
detectable hemolysate GALK activity also converted only 5 percent of an IV bolus of [ 14 C]galactose into
[ 14 C]carbon dioxide in a 5-hour period, leading to the conclusion that the enzyme deficiency is likely
widespread and not limited to the blood.
Studies performed in the mid-1970s further suggested the possibility of polymorphic variation in GALK
levels in the normal human population. In particular, Tedesco and colleagues 434, 435 assessed the GALK
and GALT levels in 1082 African-American and 618 Caucasian pregnant women in Philadelphia 435 and
found that the distribution of the enzyme levels was not normal; the African-American women had
significantly lower mean levels of GALK than did the Caucasian women. The existence of a low-activity
GALK variant in the African-American population was postulated. Three subjects in the study also
demonstrated GALK activity levels above the normal range, suggesting the possibility of a high-activity
variant in this population. Later studies of the so-called Philadelphia variant of GALK detected in the
African-American population suggested that the enzymatic impairment was specific to red blood cells; it
was not seen in white blood cells. 419 The clinical significance, if any, of these reported GALK variations in
the normal population remains unclear.
Metabolic abnormality
Patients with GALK deficiency who are exposed to significant levels of environmental galactose
accumulate markedly elevated levels of galactitol in the blood and urine. Gitzelmann’s initial study of a
GALK-deficient patient 144 showed that when the subject was drinking milk, large amounts of galactitol
were excreted in his urine. On a galactose intake of 360 g/day, corresponding to a milk intake of 3
liters/day, two-thirds of the load could be accounted for as galactose and galactitol in urine. The fate of the
remainder was unclear. The molar ratio of galactose to galactitol in the urine was 1:4. More recent clinical
studies of a small number of galactokinase-deficient patients demonstrated that urinary galactitol levels
can exceed 2500 mmol/mol creatinine prior to dietary restriction of galactose (C. Ficicioglu, personal
communication). For comparison, urinary galactitol levels in patients with classic galactosemia (discussed
below) range from 98 to 800 mmol/mol creatinine, and for controls, the range is from less than 2 to 78
mmol/mol creatinine (C. Ficicioglu, personal communication).
Once dietary restriction of galactose has been initiated, urinary levels of galactitol drop quickly, such that
after several weeks these values fall into the normal range (generally 3 mmol/mol creatinine or less) (C.
Ficicioglu, personal communication). As would be expected given the lack of galactokinase, no elevation
of Gal-1-P is detected in patients with GALK deficiency regardless of diet.
Prenatal diagnosis
GALK activity has been assayed in skin fibroblasts and cultured amniotic fluid cells (amniocytes) using a
method modified from the red cell enzyme assay of Beutler and colleagues. 44 The results suggested that
prenatal diagnosis of galactokinase deficiency would be possible, but the one "at risk" pregnancy
investigated turned out to have a normal outcome. 191
Treatment and Outcome

The standard treatment for patients with GALK deficiency is lifelong dietary restriction of galactose. If
treatment is initiated within the first weeks to months of life, the cataracts associated with GALK deficiency
may be prevented or reversed. Beyond a few months of age, however, permanent changes occur that are
not correctable by diet, but they may be treated by standard surgical intervention 12, 325 (reviewed in ref.
57).
Galactose restriction also has been recommended for heterozygotes, 422 but there are insufficient data to
determine the necessity or impact of this restriction. Indeed, although the relationship between
homozygosity for GALK deficiency and cataracts is well established, controversy remains as to whether
heterozygotes are at risk of developing this complication. A number of studies have supported such an
association, 109, 300, 328, 351, 413, 415, 424, 487 but others have not. 41, 278, 422, 429 Stambolian, in a review of
galactose and cataracts, concluded that in heterozygous GALK deficiency, presenile cataracts may
become manifest only if galactose intake is sufficiently high to cause an excess production of galactitol. 422
From an independent but related study, Simoons 414 suggested that even adults who are neither
homozygotes nor heterozygotes for GALK deficiency may be at greater risk of developing senile cataracts
if they consume large quantities of milk. The validity of this hypothesis remains to be tested.
Pathophysiology
The link between galactitol accumulation and cataract formation in GALK deficiency has been confirmed
unambiguously by a series of experiments involving a mouse model for the disease. Of note, aldose
reductase, the enzyme that catalyzes the formation of galactitol from galactose, is normally present in very
low levels in the mouse lens. 483 Studies reported in the mid-1990s demonstrated that transgenic mice
expressing normal levels of endogenous galactokinase but overexpressing a human aldose reductase
transgene were prone to cataract formation when fed a high-galactose diet 483a . In contrast, a GALK (GK)
knockout mouse, first reported by Ai and colleagues, 6 failed to develop cataracts even when fed a
high-galactose diet. The link between aldose reductase and GALK-deficiency cataracts was confirmed
when the GK knockout mice were crossed with the aldose reductase transgenic stock, resulting in F 1
progeny that carried both the human aldose reductase transgene and the GK knockout. These mice
developed cataracts by postnatal day 1, even in the absence of a high-galactose diet. 6 These
experiments clearly demonstrated the critical role of aldose reductase in galactosemic cataract formation.
Mutations at the GALK Locus

As of January 2008, at least 25 different mutations, including deletions, insertions, missense, and
nonsense mutations, had been characterized at the GALK loci of individuals with GALK
deficiency. 16, 196, 197, 216, 239, 328, 340, 357, 375, 423 Many of these mutations have been further characterized
with regard to functional significance either by expression of the corresponding allele in a tissue-culture
model system 340, 423 or by studies of purified recombinant protein. 375, 446
While some of the GALK mutations appear rare or even private, others appear more common. For
example, an Ala-to-Val substitution at codon 198 (A198V) is found at relatively high frequency among
Asian patients. 328 Called the Osaka variant, the A198V substitution mutation was first identified through
mass screening of newborn infants. Population studies suggested a prevalence of 4.1 percent in
Japanese and 2.8 percent in Koreans. Individuals of Taiwanese and Chinese ancestry demonstrated
lower frequencies, and the mutation was not found in African-Americans or Caucasians. Interestingly, an
unusually high frequency (7.8 percent) was found in Japanese individuals with bilateral cataracts, leading
Okano and colleagues to postulate that this mutation may be one of the genetic risk factors for cataracts
in the elderly.
Other mutations that may reflect founder effects in isolated populations include a Pro-to-Thr substitution at
codon 28 (P28T), originally identified in 6 affected Romani families from Bulgaria. 216 Subsequent studies
have detected the same mutation in Romani individuals living in different European countries, and
haplotype analyses indicate that this is almost certainly a founder mutation in the Romani population. 197
The same mutation also was found in the homozygous state in affected individuals from 5 Bosnian
families and in a single affected individual from a Turkish family. 357 These results predict that the P28T
mutation is probably one of the most common at the GALK locus in Mediterranean populations and in
those living in southeastern Europe.
A second GALK mutation whose frequency appears to reflect a founder effect is the nonsense mutation
Q382X, which has been found in 6 Costa Rican individuals, 5 of whom were homoallelic. 239
GALT DEFICIENCY (CLASSIC GALACTOSEMIA)

Classic galactosemia, which results from profound impairment of GALT activity, occurs as an autosomal
recessive disorder (OMIM 230400) affecting approximately 1 in 47,000 live-born infants 432 ; frequencies in
specific populations can vary widely (reviewed in ref. 54). A diagnosis of GALT deficiency may arise from
the investigation of acute clinical illness or may be made presymptomatically following an abnormal
newborn screening result.
Acute clinical presentation

Infants with classic galactosemia generally appear normal at birth, but within just days to weeks of
ingesting a normal milk diet, they can present with a life-threatening illness. Initial symptoms include poor
feeding with poor weight gain, vomiting and diarrhea, lethargy, and hypotonia. On physical examination,
infants are jaundiced, have hepatomegaly, may have a full fontanelle, and may show prolonged bleeding
or excessive bruising after venous or arterial sampling. The results of initial investigations indicate liver
disease (unconjugated or combined hyperbilirubinemia; raised liver transaminases; and raised plasma
amino acids, particularly phenylalanine, tyrosine, and methionine) and renal tubular disease (metabolic
acidosis, galactosuria, glycosuria, albuminuria, and aminoaciduria). In addition, hematologic abnormalities
are common, disordered clotting results from liver disease, and hemolytic anemia may occur. Septicemia
is sometimes seen with the early presentation of galactosemia (particularly E. coli) and may cause
diagnostic confusion. Cataracts usually are present but in the first weeks of life are for the most part mild
and may be detected only on slit-lamp examination.
Less frequently, patients may present with a more chronic illness marked by persistent poor feeding and
vomiting, failure to thrive, and developmental delay. This phenotypic variability may be due either to
environmental factors (e.g., those related to galactose intake) or to genetic heterogeneity with residual
GALT activity. 317
The diagnosis of suspected classic galactosemia is confirmed by GALT enzyme assay, generally of red
blood cells. The early clinical presentation of galactosemia is by no means specific, and a high index of
suspicion is necessary, particularly in countries where galactosemia is not included in a newborn
screening program. The presence of reducing substances or galactose in the urine is neither sensitive nor
specific and should not be used to either confirm or refute a diagnosis. Similarly, small quantities of
galactose are found commonly in the urine of any patient with liver disease, and galactose may disappear
rapidly from the urine of patients with galactosemia who are on IV fluids.
A fluorescent spot test (Beutler test) for GALT activity is used widely for the diagnosis of galactosemia, 43
although other enzyme testing strategies are also in use at some institutions. In the Beutler assay, whole
blood is added to a reaction mixture containing UDP-Glc, Gal-1-P, and nicotinamide adenine dinucleotide
phosphate (NADP). If GALT is active, Glc-1-P will be produced, which is then converted into Glc-6-P by
endogenous phosphoglucomutase. Finally, endogenous glucose-6-phosphate dehydrogenase oxidizes
Glc-6-P to form 6-phosphogluconate, with the concomitant reduction of NADP to NADPH. After a period of
incubation, the reaction mixture is spotted onto filter paper. If GALT activity is normal, the NADPH
produced by the coupled reaction will demonstrate fluorescence under ultraviolet light (wavelength
approximately 360 nm). Fluorescence after incubation for 1 h is normal. No fluorescence after 2 h of
incubation indicates GALT deficiency or deficiency of a coupling enzyme. False-positive results may occur
if blood is collected into EDTA tubes or stored at warm temperatures. Similarly, because of GALT activity
in transfused red cells, no enzyme assay of blood is reliable if there has been a blood transfusion within
up to 4 months.
An abnormal Beutler test must be followed by further biochemical and/or molecular confirmation of the
diagnosis. GALT activity in patients with classic galactosemia generally is undetectable or detected at less
than 1 percent normal levels. One commonly used quantitative measurement of GALT activity is based on
the consumption of UDP-Glc by transferase with measurement of residual UDP-Glc by a UDP-Glc
dehydrogenase/NAD+ assay by absorbance at 340 nm. Other assays are also available, 42 including
direct assays that follow conversion of radioactively labeled Gal-1-P to UDP-Gal 113 or direct assays that
follow conversion of unlabeled substrates, with detection and quantification of substrates and products by
high-performance liquid chromatography (HPLC). 329, 369 Although these quantitative enzyme assays are
more reliable, as with the Beutler assay, they can give false-negative results following a blood transfusion
or false-positive results if the blood sample was not stored properly, 286 although approaches have been
developed to minimize this problem. 126, 132 Measurement of hemolysate Gal-1-P may be accomplished
enzymatically, by HPLC, 369 or by a gas chromatography–mass spectrometry (GC/MS) isotope-dilution
method. 70, 382 Gal-1-P levels are invariably high in patients with classic galactosemia ingesting a normal
milk diet and fall slowly over a period of weeks to months following dietary restriction of galactose.
Alternatively, if rapid mutation analysis is available, the diagnosis often can be confirmed (but not
excluded) by DNA analysis. Finally, Barbouth and colleagues 23 recently proposed rapid analysis of
whole-body oxidation of [ 13 C]galactose to CO 2 in expired air as an additional strategy to detect
galactosemia in the neonatal period.
Newborn screening
Population newborn screening for classic galactosemia is now commonplace in many parts of the world,
allowing for presymptomatic diagnosis and intervention. Screening protocols vary, but most include some
form of GALT enzyme assay (e.g., the Beutler assay) combined with an assessment of total blood
galactose (galactose + Gal-1-P). While total blood galactose levels in patients will vary with diet, the GALT
enzyme assay result should not.
As part of the Austrian Newborn Screening Program, Item and colleagues 206 explored a two-tier system
of neonatal screening that included DNA analysis on blood spots with an initial positive biochemical
screening result. The advantage of this approach was a reduction in the false-positive rate, although there
also were certain disadvantages. A major obstacle was the extraordinary allelic heterogeneity in
galactosemia (explained below). In fact, the single confirmed case of galactosemia from this study
involved an Asian boy who had two rare mutations, one of which was novel. A second study by
Dobrowolski and colleagues 103 concluded that mutational analysis improves the sensitivity and specificity
of newborn screening for galactosemia, although cost and allelic heterogeneity continue to present
obstacles.
The necessity and long-term advantage of newborn screening for galactosemia remain points of
controversy, at least in some parts of the world. For example, although newborn screening is available in
all states within the United States and in many European nations, only a small percentage of newborns
are screened for galactosemia in the United Kingdom. In assessing the value of newborn screening for
galactosemia, a number of questions must be addressed: Does screening result in earlier diagnosis?
Does it reduce early morbidity and mortality? Does screening prevent or reduce late complications?
Unfortunately, there are no prospective studies that have answered all these questions fully. Other
relevant factors are the age at which screening is performed and the speed of turnaround relative to the
anticipated time of clinical presentation. As explained below, current data suggest that early, rapid
newborn screening for galactosemia indeed saves lives and prevents acute morbidity, 391, 393 although the
impact of newborn screening on the long-term outcomes of patients remains less clear.
In a retrospective survey conducted by Waggoner and colleagues, 463 80 percent of patients who had
newborn screening were diagnosed by 14 days, compared with only 35 percent of patients who were not
screened. Of the screened group, 20 percent were said to be without symptoms at the time of diagnosis.
By definition, none of the patients in the unscreened group were asymptomatic at the time of diagnosis. A
study by the British Paediatric Surveillance Unit ascertained new cases of classic galactosemia in the
United Kingdom and the Republic of Ireland prospectively over a 3-year period. 194 Sixty affected children
were identified, revealing an incidence of 1 in 44,000 live births in the United Kingdom as a whole and 1 in
23,500 live births in the Republic of Ireland. Because there is screening for galactosemia in Scotland and
the Republic of Ireland but not in England and Wales, it was possible to compare infants with
galactosemia born into screened populations with those born into the unscreened populations. Of 14
affected children identified by neonatal screening, all but one were diagnosed before 3 weeks of age,
compared with 41 of 45 of the infants not screened. Of the 4 unscreened children diagnosed after 3
weeks, 1 was diagnosed at 90 days and another at 120 days. In this study, 1 infant in the unscreened
group died, as opposed to no deaths in the screened group. Nevertheless, this single death occurred on
postnatal day 4 and could not have been prevented by a screening program with a turnaround time of 14
days. Furthermore, morbidity was no more frequent in the unscreened group.
One of the most serious problems in unscreened populations is that patient welfare depends entirely on
the ability of physicians to recognize and accurately diagnose the clinical disorder quickly. Not
surprisingly, especially given the rarity of galactosemia, failures do occur. For example, 2 infants identified
by newborn screening in Canada were both in the hospital with an undiagnosed acute illness at the time
the abnormal screening result was reported. 231 It is therefore very probable that there are infants with
GALT deficiency in unscreened populations who have died from gram-negative sepsis or other
complications without their underlying diagnosis being recognized. Of course, infant death in the absence
of a diagnosis means that the parents of these infants are also likely to remain unaware of their carrier
status and therefore their 1 in 4 risk of recurrence with each subsequent pregnancy.
The effectiveness of newborn screening in Ireland from 1972 to 1992 also has been reviewed. 20 Over this
20-year period, 1.2 million infants were screened, with a diagnosis of 55 classic galactosemia cases and 7
cases of the variant Duarte galactosemia (described below). Of these patients, 39 were detected by
routine newborn screening with a mean age of 8 days at diagnosis; 15 infants were diagnosed
subsequent to high-risk screening (previous sibling affected). Seven cases (11 percent) gave
false-negative results; of these infants, 5 had classic galactosemia and 2 had variant galactosemia. Three
of the 5 infants with classic galactosemia were feeding poorly at the time of screening, whereas the other
2 were on lactose-free milk. Notably, 7 older siblings of the index cases had died unexpectedly in infancy.
Six of these older siblings were born before newborn screening was introduced; these deaths were likely
due to galactosemia.
With regard to long-term outcome, the value of newborn screening remains unclear. For example,
Waggoner and colleagues 463 found no relationship between IQ at ages 6 to 9 years and the age at which
treatment was begun. Similarly, Schweitzer-Krantz 393 found no statistically significant relationship
between IQ and age at which treatment was initiated as long as that age was under 8 weeks. Although
both these studies were retrospective, each challenges the concept of newborn screening as a means of
preventing long-term complications in classic galactosemia.
Enzyme defect
Classic galactosemia results from profound impairment of the GALT enzyme. It is important to note that
while many patients with classic galactosemia have no detectable GALT activity, at least in red blood
cells, this is not universally true. Some patients demonstrate trace, albeit detectable, levels of residual
GALT activity in red blood cells or other tissues. For example, Xu and colleagues 494 applied a
high-sensitivity GALT assay to red blood cell samples, all of which were categorized by the standard
clinical assay as profoundly GALT-deficient. Of 423 samples tested, 60 demonstrated between 0.02 and
10 percent residual activity in red blood cells using the higher-sensitivity assay. Similarly, biochemical
studies of GALT enzymes expressed from individual patient alleles in a yeast model system demonstrated
a spectrum of activities ranging from true null to almost 10 percent. 365 This residual GALT activity,
expressed uniformly or in key tissues, has been proposed as one of the likely factors leading to milder
outcome in some patients. 223, 318
Variants of transferase-deficiency galactosemia

As explained earlier (and below), whereas some naturally occurring mutations in GALT are complete
nulls, others are hypomorphs. Those alleles associated with significant residual GALT activity are
sometimes said to result in variant rather than truly classic transferase-deficiency galactosemia. Some
variant forms of galactosemia have been detected as a result of newborn screening programs, whereas
others have been detected in patients who presented with neonatal signs of galactosemia but
subsequently demonstrated GALT levels above zero 494 and/or an unexpectedly high tolerance for
galactose.
The most frequent and well-studied variant of transferase-deficiency galactosemia is associated with the
so-called Duarte allele. Owing to the disproportionately large volume of information available concerning
Duarte galactosemia, this variant, along with a related allele called the Los Angeles allele, will be
described in a separate section below under "Mutations at the GALT locus." The Los Angeles and Duarte
alleles are sometimes also referred to as the Duarte1 (D1) and Duarte2 (D2) alleles, respectively.
Of the variant forms of galactosemia that have been reported, many appear to reflect mutations that
partially compromise either GALT catalytic function or GALT protein stability or both. For example, one
variant originally identified in an African-American patient by Segal and colleagues 397 demonstrated zero
GALT activity in erythrocytes but 10 percent residual activity in the liver and intestine. This variant is now
known to result from the presence of the S135L GALT allele (explained in the GALT mutations section
below), and the apparent red blood cell–specific enzyme deficiency is believed to reflect the fact that
GALT is not synthesized in circulating mature red blood cells, so a partially destabilized S135L-GALT
protein decays in those cells over time and at equilibrium falls below the threshold of detection. In liver,
intestine, and other nucleated, dividing cells, the enzyme continues to be synthesized, so its equilibrium
level remains above the threshold of detection. Other variants that demonstrated partial impairment of
activity, altered gel electrophoretic mobility, or unusual heat lability also have been described. 66, 67, 312
Metabolic abnormalities
GALT deficiency results in profound abnormalities in the accumulation of galactose and galactose
derivatives in the blood, tissues, and urine of patients. While the most extreme metabolic abnormalities
are seen among patients who consume large quantities of galactose, e.g., infants prior to diagnosis, clear
abnormalities have been documented even in utero, 9 and mild metabolic abnormalities often persist
despite prolonged dietary restriction of galactose. 70, 382, 497 The abnormalities detected in utero or in
patients on careful restriction of dietary galactose may reflect the metabolism of endogenously produced
galactose.
Metabolic features that are characteristic of classic galactosemia include the accumulation of abnormally
high levels of galactose, galactitol, galactonate, and Gal-1-P in the blood and/or tissues 382, 495 and
excretion of abnormally high levels of galactose, galactitol, and galactonate in the urine. 381, 497 There also
may be some impact on UDP-Gal and/or UDP-Glc levels and/or ratios in blood or tissues, although this
point remains controversial (explained below).
Galactose
Normal newborn infants tested after a milk meal show markedly greater increases in blood glucose than in
blood galactose. 412 Considering that milk sugar, lactose, contains equimolar amounts of glucose and
galactose, this result demonstrates the rapid conversion of galactose to glucose under normal
circumstances. In contrast, galactosemic infants show markedly greater increases in blood galactose
following a lactose or galactose load, 157, 164 and blood glucose levels in these infants can fall to
hypoglycemic levels. 157 Prior to initiation of dietary galactose restriction, hemolysate galactose levels can
rise as high as 110 mg/100 ml, 171 which corresponds to more than 6000 µmol/liter. Similarly, urinary
galactose levels can reach 227 mol/mol creatinine in untreated galactosemic patients. 211
Blood galactose levels fall quickly on initiation of dietary galactose restriction. For example, a recent study
of 41 patients on dietary restriction reported by a European group 382 demonstrated that hemolysate
galactose levels ranged from 1.6 to 8.5 µmol/liter RBCs, with a mean ± SD of 3.8 ± 1.7; the plasma
galactose level in these same patients was 3.3 ± 1.0 µmol/liter. For comparison, the hemolysate galactose
level detected in a cohort of 33 controls was 0.43 ± 0.2 µmol/liter (plasma 0.33 ± 0.06 µmol/liter). A similar
study reported by a U.S. group 322 found that the plasma galactose level detected in 15 patients with
classic galactosemia on dietary treatment was 2.72 ± 0.70 µmol/liter (mean ± SE), whereas the plasma
galactose level in 21 controls was 1.48 ± 0.32 µmol/liter.
Gal-1-P
Untreated patients with classic galactosemia accumulate excessive amounts of Gal-1-P in their red blood
cells; indeed, it was Gal-1-P that first helped to pinpoint the position of the metabolic block in this
disorder. 390 Other tissues that accumulate Gal-1-P include the liver, kidney, brain, tongue, adrenal, and
heart 389 ; the ocular lens 152 ; and cultured skin fibroblasts. 315 Gal-1-P also accumulates abnormally in
patients prior to birth; two galactosemic fetuses at 20 weeks’ gestation had levels of Gal-1-P in their livers
comparable with levels seen in newborns who died of the disorder. 9, 315 Prior to treatment, infants with
classic galactosemia may demonstrate hemolysate Gal-1-P levels as high as 2.5 to 6.5 mM. 147 Following
dietary restriction of galactose, hemolysate Gal-1-P values fall dramatically, although they may remain
somewhat elevated (e.g., 0.1 to 0.2 mM) 147, 382 ; Gal-1-P is low to undetectable by most methods in
control blood samples. Consistent with this conclusion, Chen and colleagues 70 recently applied a GC/MS
isotope-dilution assay to detect Gal-1-P values of 166 and 373 mg/liter of packed red bllod cells
(corresponds to 0.64 and 1.43 mM) in two newly diagnosed infants, 10.9 to 45 mg/liter (corresponds to
0.04 to 0.17 mM) in patients on dietary restriction of galactose, and less than 2.4 mg/liter (corresponds to
less than 0.01 mM) in normal individuals.
Galactitol
Galactitol accumulates to high levels in the blood, tissues, and urine of patients with classic
galactosemia. 353 This accumulation is particularly striking in patients prior to the onset of dietary
restriction of galactose. Following dietary intervention, galactitol levels drop, although they may never
normalize.
A number of studies have explored galactitol accumulation in galactosemia, addressing questions of

where and when accumulation occurs and whether the level of galactitol accumulation correlates with
GALT genotype. For example, both autopsy 482 and in vivo studies using proton magnetic resonance
spectroscopy 471 have confirmed that concentrations are high in the brain. Brain galactitol may be
sufficiently high in newborn infants with massive urine galactitol excretion to lead to cerebral edema. 30
Wells and Pittman 481 were the first to demonstrate the significant urinary excretion of galactitol in
galactosemia, whereas Allen and colleagues demonstrated raised levels of galactitol in amniotic fluid at 10
weeks’ (Allen personal communication) and 20 weeks’ 10 gestation.
Jakobs and colleagues 211 found that while galactitol levels in the urine of patients decreased following the
onset of dietary galactose restriction, these levels nonetheless remained 20 to 30 times higher in patients
than in age-matched control subjects, and this difference was especially marked in infants. More recently,
Schadewaldt and colleagues 382 used a stable-isotope GC/MS approach to follow galactitol levels in the
red blood cells and plasma of (treated) patients and controls. They found galactitol levels of between 5.0
and 11.2 µmol/liter in the red blood cells of patients, whereas the corresponding levels in controls ranged
between 0.25 to 1.79 µmol/liter. The galactitol level in patient plasma was 11.2 ± 1.4 µmol/liter, whereas
the corresponding level in the plasma of controls was 0.21 ± 0.06 µmol/liter.
Palmieri and colleagues 332 found that urinary galactitol was 5 to 10 times higher in patients homozygous
for the Q188R mutation (discussed below) compared with normal individuals of a similar age. Urine
galactitol also was increased in African-American patients with the S135L mutation (see below) but not in
patients with Duarte galactosemia (see below). Palmieri and colleagues 332 further demonstrated that
plasma galactitol was increased in patients with galactosemia and that, unlike urinary galactitol, the levels
did not decrease with age. Galactitol was undetectable in the plasma of normal individuals.
Berry and colleagues 32 addressed the source of the continued galactitol accumulation in patients on
galactose-restricted diets and concluded that cryptic dietary galactose could not account for the levels of
galactitol observed. Endogenous production of galactose therefore was postulated, and the production
rate required to explain the observed galactitol excretion rate was calculated. This figure agreed very
closely with that determined by more direct means in later studies. 31, 37
UDP-Gal and UDP-Glc

There has been considerable controversy over the past two decades concerning the levels and ratios of
UDP-Gal and UDP-Glc in patients with classic galactosemia; some reports claim that the levels of these
sugar nucleotides are abnormally low in patients, whereas other reports contradict those claims. Potential
perturbation of UDP-Gal and UDP-Glc levels or ratios is an important issue because both serve as key
substrates for the biosynthesis of glycoproteins and glycolipids (reviewed in refs. 11 and 370).
The confusion and controversy have been fueled by at least five factors. First, UDP-Gal and UDP-Glc
have been measured by different techniques, some of which are now recognized as unusually susceptible
to artifact. Second, studies have raised concern that dietary restriction of galactose, rather that GALT
deficiency alone, may underlie the UDP-Gal/UDP-Glc abnormality sometimes observed in patients. Third,
one study reported that abnormalities in the UDP-hexose levels of patient red blood cells were not
recapitulated in other cell types, in particular, leukocytes or fibroblasts, raising concern that red blood cells
may present a convenient but aberrant rather than representative cell type for study. Fourth, the patient
populations studied have varied. In some studies, all patients classified as affected with classic
galactosemia were included. In other studies, patients were stratified according to the level of residual
GALT activity detected by a high-sensitivity assay. This difference between patient cohorts could
potentially explain why some groups have seen differences and others have not. Finally, the conclusions
reported from recent studies of individual patient fibroblast lines cultured in vitro are not fully supported by
the data presented.
Issues of Methodology
Shin and colleagues 404 were the first to claim that there is a decreased level of UDP-Gal in the red blood
cells of galactosemic patients, and these authors postulated that such a metabolic abnormality could have
an adverse effect on the synthesis of galactosides. This concept was further developed by Ng and
colleagues, 316 who also found a UDP-Gal deficit in liver and skin fibroblasts. These early reports,
however, were based on enzymatic methods of UDP-Gal and UDP-Glc detection that were later
challenged as prone to artifact. Subsequent investigations by two groups using HPLC methods 33, 224 and
by a researcher using an enzyme technique 232, 233 supported the claim of a reduction in patient red blood
cell UDP-Gal relative to UDP-Glc, but the absolute levels of both metabolites were almost three times
lower than those reported in the initial studies.
Role of Diet
The complexity deepened when Berry, Gibson, and colleagues not only measured the levels of UDP-Gal
and UDP-Glc in patients and controls but also explored the possible impact of dietary galactose restriction
on red blood cell metabolite levels. 33, 136– 138 In particular, these researchers applied an HPLC
methodology to quantify the UDP-Gal and UDP-Glc levels in patients and controls, as well as in patients
with other (non-GALT-related) metabolic disorders requiring a milk-restricted diet. While these
rersearchers found that the mean level of UDP-Gal in the galactosemic children was 38 percent lower
than that in normal children, only 6 of the 19 galactosemic children studied demonstrated UDP-Gal levels
that were below the 95 percent confidence limit for the normal controls. Further, the mean UDP-Gal level
detected in the nongalactosemic milk-restricted children also was 38 percent lower than the control
level. 33 Similar conclusions were reached 3 years later from an expanded study. 138
The role of diet on red blood cell UDP-hexose levels was further confirmed by another study in which
milk-restricted patients with maple syrup urine disease (MSUD), a condition resulting from impaired
protein metabolism, were given galactose supplements and followed to see what impact, if any, these
supplements would have on red blood cell UDP-Gal levels. 137 As noted earlier, in the absence of
galactose supplements, these nongalactosemic individuals demonstrated abnormally low levels of
UDP-Gal in their red blood cells, comparable with those seen in milk-restricted galactosemic patients.
Strikingly, transient dietary galactose supplementation temporarily normalized the red blood cell UDP-Gal
levels in the MSUD patients, and discontinuation of the galactose supplement was accompanied by a
return to presupplementation UDP-Gal levels.
Role of Cell Type

A separate study, also conducted by Gibson and colleagues, 139 questioned whether UDP-hexose levels
detected in red blood cells accurately represent UDP-hexose levels in other tissues. Toward this end, the
authors queried UDP-Gal and UDP-Glc levels in leukocytes and cultured fibroblasts from healthy control
individuals and from patients with classic galactosemia. In contrast to the results seen and reported earlier
from studies of red blood cells, there was no statistically significant difference detected between patients
and controls with regard to UDP-Gal or UDP-Glc values detected in either leukocytes or fibroblasts.
Furthermore, there was no age-related difference in UDP-Gal levels detected in these cell types. 139
These findings seriously call into question the clinical and biological significance of any UDP-Gal
abnormalities identified in the red blood cells of patients with classic galactosemia.
Role of Residual GALT Activity

The impact of residual GALT activity on UDP-hexose accumulation in patient red blood cells was
addressed by Xu and colleagues, 493 who applied a high-sensitivity GALT assay 494 to stratify a population
of patients with classic galactosemia into those with absolutely no detectable GALT activity (designated
GG) and those with low but detectable residual activity (designated GV). HPLC studies of the
UDP-hexose levels in the red blood cells and cultured skin fibroblasts of these two cohorts demonstrated
statistically diminished levels of UDP-Gal in both cell types of the GG patients but normal UDP-Gal levels
in the GV patients. 493 Clearly, these results affect the conclusions drawn from other studies in which the
patient populations were not stratified according to residual GALT activity.
Cells in Culture
Finally, a recent study by Lai and colleagues 249 explored the levels of UDP-hexoses in two
SV40-transformed fibroblast cells lines manipulated in culture, one of which was derived from a patient
with classic galactosemia and the other of which was derived from a control individual. These authors
grew both cell lines in hexose-free medium supplemented with either 0.1% glucose or 0.1% galactose
plus 0.01% glucose and quantified the levels of UDP-Glc and UDP-Gal in each cell line under both
conditions. In contrast to the results of Gibson and colleagues, 139 who tested fibroblasts from 10 patients
and 10 control individuals and reported no statistically significant difference in UDP-hexose levels, Lai and
colleagues 249 reported that in both media the patient cells demonstrated slightly diminished levels of both
UDP-Glc and UDP-Gal relative to the control cells. It is interesting to note that when shifted from glucose
to galactose medium, however, both the control and patient cells showed similar drops in both hexoses:
UDP-Glc dropped approximately 24 percent in the control cells and approximately 30 percent in the
patient cells, and UDP-Gal dropped approximately 44 percent in the control cells and approximately 32
percent in the patient cells. Transfected to express either GALT or UGP), the patient-derived cells
demonstrated small increases in both UDP-hexoses regardless of the culture medium. A parallel
transfection was not performed on the control cells. The true significance of this work remains difficult to
interpret in part because of the small differences reported and in part because the work was conducted
using cell lines derived from single control and patient donors so that the contribution of individual
variation cannot be judged.
Prenatal diagnosis
GALT deficiency may be diagnosed accurately in the prenatal period by analysis of the GALT enzyme
activity in amniocytes or chorionic villi or by genetic analysis if familial mutations are known; as with any
prenatal assay, ensuring the absence of maternal cell contamination is essential. Impaired galactose
metabolism also may be indicated prenatally by elevated galactitol in the amniotic fluid.
GALT Enzyme Assay

The prenatal diagnosis of GALT deficiency was first achieved by the assay of amniocytes produced from
fluid obtained at 16 to 18 weeks’ gestation and from a subsequent case at 8 to 10 weeks’ gestation. The
analytical methods used 116, 299, 315 were modifications of the erythrocyte GALT assay of Ng and
colleagues. 314 Direct assay of GALT from chorionic villus biopsy samples obtained in the first trimester
also has been achieved. 234 This provides a very rapid result because cell culture is not needed, although
chorionic villus cells also may be cultured as a backup or for confirmation of the direct assay.
Galactitol
An alternate approach to the prenatal diagnosis of GALT deficiency was suggested by the work of Allen
and colleagues, 9 who demonstrated increased galactitol levels in amniotic fluids obtained from two
affected pregnancies. This technique was further developed using amniotic fluid samples obtained
between 15 and 20 weeks of gestation. 10, 212 By 1995, at least 50 prenatal cases had been studied and
successfully distinguished as affected or unaffected on the basis of amniotic fluid galactitol level. 209 The
method also has been validated in an affected pregnancy using amniotic fluid obtained at 10 weeks’
gestation (Allen, personal communication). Of course, while elevated amniotic fluid galactitol may
accurately distinguish pregnancies affected by galactosemia from unaffected pregnancies, it does not
pinpoint which of the Leloir enzymes is deficient.
Genetic Analysis
Prenatal diagnosis of GALT deficiency is also possible by DNA analysis, provided that the familial
mutations have been established previously. 192 Chorionic villus biopsy is an excellent and rapid source of
fetal DNA. Alternatively, DNA can be extracted from amniocytes, although if culture is required, this may
delay the results.
Treatment and Outcome
Prenatal management
Galactose ingestion by the pregnant mothers of galactosemic infants has not been shown to have any
harmful effects. Some clinics recommend limiting milk intake to no more than a pint per day, but there is
no evidence to suggest that this restriction is necessary. Dietary restriction of maternal lactose intake does
not prevent accumulation of galactitol in the amniotic fluid of affected fetuses. 210 In the international
survey by Waggoner and colleagues, 463 the outcome for those infants whose mothers followed
milk-restricted diets during pregnancy was no better than that of babies whose mothers followed no such
restriction.
Neonatal management
Infants suspected or confirmed to have classic galactosemia, whether symptomatic or asymptomatic at
the time of presentation, must have galactose excluded from their diet immediately. Breast-feeding and
cow’s milk formulas therefore must be stopped and replaced by a soya-based formula or an elemental
formula (discussed below).
Some infants may be seriously ill at the time of diagnosis and require considerable supportive care.
Vitamin K and fresh-frozen plasma may be necessary to correct clotting abnormalities. In order to prevent
false-negative results on subsequent diagnostic GALT enzyme analyses, blood samples should be
collected before any blood transfusion is given. Gram-negative sepsis should be assumed, and an
appropriate antibiotic should be administered intravenously. Unconjugated hyperbilirubinemia is usually

not severe enough to warrant phototherapy or exchange transfusion, but infants may be at increased risk
of kernicterus (brain damage resulting from excessive bilirubin) if albumin levels are particularly low
secondary to liver disease. If the infant is too sick to tolerate enteral feeding for more than 1 or 2 days,
then parenteral feeding should be considered but prescribed cautiously if there is significant liver disease
or thrombocytopenia. Most children at diagnosis are able to tolerate enteral feedings, although these may
have to be given initially by nasogastric tube because of poor sucking ability. Infant soya formulas are
most appropriate unless there is significant liver disease, in which case a medium-chain triglyceride
containing casein hydrolysate, such as Pregestimil, should be used until this complication has resolved. In
the absence of significant liver disease, casein hydrolysates should be avoided, however, because they
do contain small amounts of galactose (in the range of 60 to 90 mg/100 ml).
Most symptomatic infants with classic galactosemia show rapid clinical improvement following dietary
restriction of galactose. Features such as hepatomegaly, abnormal liver function, and renal tubular
disease usually resolve fully within a week or two. Infants detected presymptomatically, e.g., by newborn
screening, may be spared these potentially lethal acute symptoms by immediate dietary intervention.
Patients with classic galactosemia are at elevated risk for neonatal cataracts 392, 463 ; of 314 patients
investigated by Waggoner and colleagues, 30 percent reported cataracts. Most galactosemia-associated
cataracts self-resolve over time after galactose is removed from the diet; of the approximately 100 patients
reported in the Waggoner study, only 8 required surgical intervention. Notably, however, cataracts also
have been seen in patients at birth 104 and even in a fetus at 20 weeks’ gestation. 458
Apparently healthy newborns believed to be at risk of classic galactosemia, such as siblings of known
patients, should begin feedings with soya formula in lieu of breast milk or cow’s milk formula until a
diagnosis of galactosemia has been formally excluded. Diagnostic exclusion can be accomplished easily
by collection of cord blood for GALT enzyme analysis or, if the familial GALT mutations are known, by
DNA analysis. The results of these assays should be reported as quickly as possible so that, assuming
galactosemia has been excluded, those mothers who wish to initiate breast-feeding might do so without
undue delay.
Soya Formula versus Elemental Formula

Although most infants with classic galactosemia are moved from milk or a milk-based formula to
soya-based formulas, a number of studies have explored the potential benefits of substituting an
elemental formula, at least during the initial period of treatment. The benefit of an elemental formula is that
it contains no galactose, whereas soya formula actually contains low levels of bound galactose (about 14
mg/liter 502 ). Two recent reports 117, 502 have described cases of infants with classic galactosemia who
continued to demonstrate unacceptably elevated levels of Gal-1-P (>4 mg/dl 117 or >6.8 µmol/g Hb 332, 502 )
despite treatment with a soya-based formula. In both reports, the infants were switched to a
galactose-free elemental formula, and in all cases, patient red blood cell Gal-1-P values quickly dropped
into the acceptable range following the change. While the metabolic benefit of the elemental formula
appears clear, at least for some patients, whether this metabolic benefit translates into clinical benefits
with regard to long-term outcome remains unknown. 117, 502
PARTial GALT Deficiency

Whether infants with partial GALT deficiency also benefit from dietary intervention remains a point of
some controversy. Levy and colleagues 265 studied 10 untreated individuals, including 3 adults with Duarte
galactosemia (see below), and found no clinical abnormalities. Many centers with newborn screening do
not recommend dietary intervention for these infants, although others do, at least for the first year of life
(e.g., see ref. 118). That partial GALT deficiencies with significant residual enzyme activity are only rarely
detected in the investigation of sick infants and children in populations where there is no newborn
screening further supports the supposition that partial loss of GALT activity does not contribute to
significant morbidity, at least in childhood.
Gitzelmann and Bosshard 151 reviewed the question of treatment for partial GALT deficiency and
concluded that dietary treatment should be recommended for all infants with GALT levels of 9 µmol/h per
gram of hemoglobin or lower and/or with Gal-1-P levels greater than 10 mg/dl. According to these authors,
if at 4 months of age there is no aminoaciduria after a 1-week trial of 2 to 3 dl of cow’s milk and Gal-1-P
has fallen to less than 2 mg/dl, the infant may be put on a normal diet. However, as Gitzelmann and
Bosshard point out, because partial GALT deficiencies may be up to 10 times more common than classic
galactosemia, this approach to treatment may entail considerable work and expense, and the clinical
necessity of intervention in these cases remains unknown.
Long-term management and outcome of patients on lifelong dietary restriction of galactose

Lifelong dietary restriction of galactose remains the standard of care for treatment of classic galactosemia
despite its limited efficacy in preventing long-term complications (described below). Patients are advised
to exclude milk and milk products from their diet indefinitely because significant ingestion of galactose at
any age may be harmful. Anecdotal reports suggest that many teens or adults with galactosemia may
comply poorly, if at all, with the recommended diet; these patients are not believed to suffer acute liver or
renal disease, although more subtle complications cannot be ruled out.
Long-Term Management
Patients with classic galactosemia require careful long-term management in order to monitor
development, provide specialist dietary assessment and advice, and recognize late complications,
particularly ovarian failure. Regular assessments of development, cognitive function, and speech are
useful to provide an objective measure of outcome and to enable early palliative intervention as needed.
The U.K. Galactosemia Steering Group has recommended a plan for outpatient review and standardized
assessments. 467 Under this protocol, patients are seen by a specialist team every 3 months up to 1 year
of age, every 4 months up to 2 years of age, every 6 months to age 14 years, and annually thereafter.
More frequent assessment is necessary for girls in late childhood and adolescence to monitor pubertal
development. Specialist speech assessments are recommended at 1 and 2 years, a standardized
developmental assessment at 4 years, and intelligence assessments at 8, 14, and 18 years.
Dietary Considerations
Diet must be followed closely both to ensure maximal restriction of galactose exposure and to ensure
adequate calcium intake (discussed below). It is clear from the analysis of many different foods that a
dairy-free diet is not necessarily a galactose-free diet. These findings led Acosta and Gross 2 to conclude:
"It is certain that no patients with GALT deficiency have ever ingested a galactose-free diet." Indeed, small
amounts of free soluble galactose are present in most fresh fruits and vegetables, although in some, e.g.,
blueberries and honeydew melon, there may be clinically significant amounts. 165, 227 Larger quantities are
found in some legumes, such as peas and dried beans. 2, 227 Further, bound galactose, with glycosidic
linkages, is an important component of a number of plant compounds such as the raffinose

oligosaccharides, arabinogalactan types I and II, and galactan. In addition, galactocerebrosides and
gangliosides are found in offal. The extent to which bound galactose can be broken down in the gut is not
known, but most studies suggest that the effect of reducing dietary intake of these substances is likely to
be minimal. 150, 236, 484
A number of medications also contain lactose as an "inert" ingredient, and this should be checked before
prescribing any medications to patients with galactosemia. However, in most cases, the amount of
galactose (compared with the endogenous production) is unlikely to be significant, particularly if the
medication is given for short periods of time.
Of note, soya formula contain high concentrations of phytate and phytoestrogens (isoflavones), and
largely based on research in animals, concerns have been expressed as to its use in infants and children.
However, to date, there is no evidence to suggest that isoflavones in soya formula are harmful to human
growth or sexual development (reviewed in ref. 291). Soya infant formula therefore remains the best
choice of milk for infants with galactosemia.
Biochemical Follow-up
Regular measurement of Gal-1-P in red blood cells is the most common method used to monitor dietary
compliance in patients. As described earlier, concentrations of Gal-1-P are often very high at diagnosis
(2.5 to 6.5 mM 147 ) and fall rapidly within the first 2 to 3 months of restricting dietary galactose. 198
Acceptable values following the introduction of the diet usually are achieved within a few months, but even
with good dietary compliance, in some cases this may take up to a year. It is important to note that despite
dietary intervention, Gal-1-P levels in patients may remain elevated relative to controls. As a general rule,
the upper limit of Gal-1-P usually considered acceptable is 150 µmol/liter of red blood cells, which also
may be expressed as 0.15 mM, or 50 µg/ml packed cells, or 5 mg/dl, or 0.5 µmol/g Hb.
Whether there is any correlation between red blood cell Gal-1-P concentration and long-term outcome
remains unclear. Frequent measurement as a means of monitoring patients with classic galactosemia
consequently is difficult to justify. Poor dietary control may lead to significantly elevated levels, but some
patients have high values even with very good compliance. One proposed schedule for testing Gal-1-P
levels is every 3 months up to 1 year of age, every 6 months from 1 to 14 years, and yearly thereafter. 467
Urinary and blood galactitol also have been investigated as means of monitoring dietary
compliance. 118, 495, 497 Hutchesson and colleagues 198 found intra- and interindividual coefficients of
variation of 37 and 42 percent, respectively, for urinary galactitol, with an analytical coefficient of only 5.5
percent, indicating that the major source of variability was biological. These authors concluded that the
high intraindividual variability for galactitol and similar variability for Gal-1-P limited the utility of both
metabolites as a means of biochemical monitoring.
Additional or Alternative Treatments

In the mid-1990s, Manis and colleagues 282 undertook a clinical trial of uridine supplementation in
galactosemia. The rationale for this study derived from reports, described earlier, that levels of UDP-Gal
and/or UDP-Glc could be abnormally low in some patients. Manis and colleagues hypothesized that oral
uridine supplementation might help to alleviate this sugar nucleotide depletion and thereby alleviate
long-term complications in patients. Uridine powder at a dose of 150 mg/kg per day was given to 29
patients over a period of 2 to 5 years, and the cognitive function of these patients was compared with that
of 6 other patients who did not receive uridine. The mean study participation time of the patients was 3.9
years. Unfortunately, there was no significant difference between the treated and untreated groups,
suggesting that oral uridine is not an effective supplemental therapy for patients with classic galactosemia.
However, given the relatively small sample sizes and the broad age range of the patients studied, these
data do not completely rule out the possibility that uridine might provide some subtle benefit, at least to
certain patient cohorts. Of note, recent data from a tissue culture model system 386 suggest that uridine
treatment might benefit patients with epimerase deficiency galactosemia, as discussed below.
Cognitive Development
One of the prevalent long-term complications of classic galactosemia is impaired cognitive development,
which affects at least 30 to 50 percent of all treated patients. 463 Additionally, a number of children have
visual-perceptual dysfunction, 121, 122, 241 and there is some evidence to suggest that adults with GALT
deficiency have a particular personality trait: Individuals are described as having shy and timid
personalities, being anxious and fearful in interpersonal relationships, and developing into overdependent
adults. 120, 121, 241 Age at start of diet, except after 2 months, milk restriction during pregnancy, and
hemolysate Gal-1-P levels do not seem to correlate with DQ or IQ. 78, 392, 463
In most studies of cognitive development involving older children and adults, mean IQ for patients has
been in the range of 70 to 90, although normal or even above-average intelligence has been observed in
individual patients. 14, 106, 221, 241, 364, 392, 402, 463 Formal developmental and intelligence testing of patients
at different ages found that the mean IQ for patients was lower in older age groups, 120, 221, 241, 392, 463
suggesting that intelligence may fall with increasing age, at least in childhood. Further, the retrospective
survey conducted by Waggoner and colleagues 463 found decreases averaging 6.2 points in 42 children
who had an initial assessment at 3 to 5 years of age, repeated at 6 to 9 years, and decreases averaging
4.4 points in 46 children tested initially at 6 to 9 years, repeated at 10 to 16 years. There was not,
however, a consistent point drop for children tested with the same IQ test, and in other studies, there has
been no demonstrated decrease in cognitive ability over time, 282 suggesting that the age-related
differences reported earlier could be artifactual.
Speech and Language

Speech and language defects have been particularly frequent in patients with classic galactosemia;
delayed vocabulary and articulation problems are reported to occur in up to 90 percent of
children. 14, 366, 463, 464, 474 Developmental verbal dyspraxia (DVD), an unusual speech disorder owing to a
sensorimotor disturbance of articulation, also has been reported relatively frequently. 170, 309, 366
Individuals with DVD have difficulty with the voluntary movements required for speech, and this difficulty
may be manifest by audible or visible "groping" behaviors while attempting to form words. 187 Patients with
the most severe speech disorders or visual-perceptual difficulties generally have lower IQs 309, 463 ; it is not
clear whether these particular impairments are specific deficits or part of the global cognitive disorder. 221
Ataxia and Other Late Neurologic Complications

Progressive neurologic disease developing in treated patients after early childhood was first described in
1973. 213 Since then, there have been a number of case reports confirming this
association. 48, 49, 131, 213, 237, 392 Hypotonia; resting tremor; intention tremor; dysdiadochokinesis; ataxic
gait; dystonic, choreatic, and ballistic movements; uncoordinated facial movements; and scanning
dysarthric speech all have been described. The similar clinical presentation in these reports suggests a
specific neurologic syndrome owing primarily to cerebellar dysfunction, although extrapyramidal
disturbance also may play a part. 48 The age of onset has varied from early to middle childhood to middle
age. Early signs have included truncal ataxia and hypotonia, which then have proceeded to more florid
cerebellar disease. Only a proportion of individuals with classic galactosemia has developed this
complication. In the international survey from Waggoner and colleagues, 463 18 percent of patients over
3.5 years of age were reported to have problems with coordination or gait. In a retrospective study in
Germany, 22 percent of patients had evidence of cerebellar dysfunction; intention tremor was found in 11
of 78 patients, mild ataxia in 3, and severe ataxia in 3. 392 Kaufman and colleagues 221 reported that 12 of
45 (27 percent) patients had tremor, ataxia, or dysmetria. Early dietary intervention has not protected
individuals from late neurologic disease. 237
Brain Imaging
Abnormalities on brain imaging are common in classic galactosemia. Computed tomography (CT) and
magnetic resonance imaging (MRI) in patients with late neurologic disease have demonstrated abnormal
white matter, ventricular enlargement, diffuse cortical atrophy with basal ganglia and brain stem
involvement, and cerebellar atrophy. 48, 49, 74, 131, 237, 272 Nelson and colleagues 310 reported the results of
brain MRI in 67 transferase-deficient patients, of whom 63 had classic galactosemia. T 1 -weighted images
showed normal myelination for age in all patients. However, the reduction in peripheral white matter
attenuation on T 2 - and intermediate-weighted images that occurs normally in infancy was not present in
52 of 55 patients with classic galactosemia over the age of 1 year. This abnormality was postulated to
represent a failure of normal myelination. In 22 patients, mild lateral ventricular enlargement compatible
with cerebral atrophy was present, and in 8 patients, there was enlargement of the fourth ventricle and
cerebellar sulci, suggesting cerebellar atrophy. In 6 of these 8 patients, the cerebellar atrophy was
moderately severe. Of the 63 classic galactosemia patients, 10 were ataxic on examination, but only 2 of
these 10 had moderate cerebellar atrophy. In addition to these findings, 11 patients had two or more
discrete focal white matter lesions on T 2 -weighted images scattered through the cerebral white matter,
most commonly near the corners of the lateral ventricles. The basal ganglia nuclei and brain stem were
normal in all these patients.
Ovarian Failure
Hypergonadotrophic hypogonadism is the most common long-term complication experienced by girls and
women with classic galactosemia, reported in 80 to more than 90 percent of female patients in some
studies. 219, 372, 463 In contrast, testicular function in males appears to be normal.
The clinical manifestations of ovarian failure in galactosemia may range from delayed puberty to primary
amenorrhea, secondary amenorrhea, or oligomenorrhea. For example, in the study by Waggoner and
colleagues, 463 8 of 34 women over 17 years of age had primary amenorrhea. The mean age of menarche
in the remaining patients was 14 years (range 10 to 18 years), and the majority of these developed
secondary amenorrhea or oligomenorrhea after a relatively short period. The first biochemical evidence
for ovarian failure may be an abnormal (increased) release of follicle-stimulating hormone (FSH) and
luteinizing hormone (LH) following a luteinizing hormone–releasing hormone (LRH) stimulation test,
although most patients also have raised basal FSH and LH levels. Estradiol concentration may appear
normal initially with high gonadotropin levels, indicating continued follicular development, but these levels
fall as ovarian failure progresses. 219
As with other late complications, ovarian failure appears to be independent of dietary compliance, as well
as independent of the age at which treatment was initiated. The etiology of ovarian failure remains
unknown, although damage to ovaries both prenatally and postnatally is implicated. Prenatal damage is
suggested by reduced oocyte numbers in rats subjected to prenatal exposure to high levels of
galactose, 71 and streak ovaries have been reported on laparoscopy. 188, 302 However, postnatal damage
is supported by a report of normal ovarian tissue in a girl who died from E. coli sepsis at the age of 5
days 302a and by the finding that in some patients gonadal function may be normal initially but can become
abnormal with time. 219
Ovarian histology in galactosemic women with ovarian failure has been variable. 135 Oocyte numbers
generally have been absent or low, with the stroma reported as fibrous or streaked in appearance.
Possible mechanisms for the ovarian dysfunction include the production of abnormal FSH, 352 abnormal
receptor activity, or a direct toxic effect of Gal-1-P or other metabolites on ovarian tissue. Considering that
women with GALK deficiency do not experience ovarian dysfunction, it is unlikely that galactose,
galactitol, or galactonate alone account for this aspect of pathophysiology.
The intimate involvement of GALT enzyme and FSH in the anterior pituitary was explored by Daude and
colleagues, 99 who studied rats at different stages of the estrous cycle. Coexpression of FSH and GALT
found in specific cell fractions, especially during proestrus and estrus, indicated that GALT expression is
strongly associated with gonadotropin-producing cells. Although the origin of this variation in GALT
expression remains unknown, it could be related to varying physiological requirements for galactose,
especially since FSH (as well as LH and thyrotropes) contains galactose-rich oligosaccharide side chains.
Inadequate provision of UDP-Gal at specific stages of the estrus cycle therefore may contribute to ovarian
dysfunction.
Prestoz and colleagues 352 have shown that the sera of galactosemic women contain quantitatively
abnormal isoforms of FSH. Using isoelectric focusing and immunodetection of FSH isoforms, this group
demonstrated that in addition to the apparently normal acidic isoforms, sera from galactosemic females
displayed isoforms migrating close to neutral pH. These altered isoforms were explained by a deficiency
of galactose in the carbohydrate moiety of FSH to which sialic acid is linked. This observation is
particularly striking given a subsequent report, by Menezo and colleagues, 290 of successful pregnancy
and delivery in a galactosemic woman with hypergonadotropic hypogonadism following treatment with
recombinant FSH.
A retrospective cross-sectional study of 53 females with classic galactosemia led Guerrero and
colleagues 167 to propose that the risk of premature ovarian failure can be predicted from three areas of
pathology: the patient’s molecular genotype for GALT, alternative pathways for galactose metabolism
(with a galactose oxidation level of less than 5 percent increasing the risk), and the patient’s environment
at diagnosis and during treatment (with a mean erythrocyte Gal-1-P level of greater than 3.5 mg/dl
increasing the risk). Thus the overall risk of premature ovarian failure may relate to several factors in
combination rather than to any one independently. Gonadal dysfunction in galactosemia and its various
possible etiologies have been reviewed by a number of authors. 124, 135, 271
Considering the prevalence of primary or premature ovarian failure among girls and women with classic
galactosemia, all female patients should be followed by an endocrinologist or other appropriate specialist,
especially during late childhood and puberty. Raised gonadotropin levels in infancy may be an indication
of early ovarian failure, but their interpretation is not straightforward, and measurement thereafter in early
childhood is not helpful. Recommendations of the ages at which endocrine assessment should be
undertaken have been published. Walter and colleagues 467 suggest measurement of FSH, LH, and
estradiol at 6 months and then again at 10 and 12 years and, if necessary, yearly thereafter, with referral
to a pediatric endocrinologist by age 10 years. The age when hormone therapy should begin is debatable.
Some groups have recommended treatment from the age of 16, 149 whereas others suggest starting
earlier. The recommendation from the U.K. Galactosemia Steering Group and the British Society for
Paediatric Endocrinology and Diabetes is that hormone therapy be given to patients with elevated basal
gonadotrophin levels and low estradiol starting at between 11 to 13 years. 467 A suggested regimen is
given in Table 72-3. All oral hormone preparations contain small quantities of lactose, but the total dose is
very small in comparison with endogenous production.
Table 72-3: Recommendations for the treatment of hypergonadotropic hypogonadism in females
with galactosemia. From Walter et al. 467 See text for details
Year of
Treatment
treatment
Ethinyl estradiol at a dose of 2 mg/day. If pubertal advance is particularly slow, then the
1
time at this dose should be reduced to 6 months.
2 Ethinyl estradiol at a dose of 5 mg/d.
Ethinyl estradiol 10 mg/d. The patient and parents should be informed that vaginal
3 spotting or withdrawal bleeds may occur during this year. If vaginal spotting or
withdrawal bleeds occur, a progestogen should be added at the start of each month.
Loestrin (an oral contraceptive preparation containing 20 mg ethinyl estradiol and a

progestogen) given in the usual way, ie, daily for 21 days followed by 7 days off). This
4 and preparation should be adequate for most young women. If there is breakthrough
thereafter bleeding or any other symptoms of estrogen deficiency, then a preparation containing
30 mg of ethinyl estradiol can be used. (Natural estrogens would be an acceptable
alternative to synthetic estrogen.)
Outpatient review should be every 4 months during the first 2 years of hormone
Notes
treatment and every 6 months thereafter.
Monitor
• Physical development and growth throughout these years of pubertal induction.
• Bone age yearly from the start of treatment.
• Pelvic ultrasound (for uterine dimensions) at the start of treatment; and repeat after 2
and 4 years.
• Blood pressure.
Special care and advice should be given to those with a family history or past history of
venous thrombosis and to those who are overweight or who smoke.
Pregnancy in Women with Galactosemia

Although most female patients with classic galactosemia are infertile by the time they reach childbearing
age, a small number have given birth. 60, 100, 184, 228, 324, 327, 376, 463 In addition to unassisted conception, a
single pregnancy has been reported after oocyte donation, 379 and as cited earlier, an additional
pregnancy was reported after treatment with recombinant FSH. 290 The majority of these pregnancies
resulted in normal infants. 60, 290, 367, 376, 463 Further, Gal-1-P levels do not appear to be elevated in the
cord blood of infants born to galactosemic mothers, even if maternal levels are high. 367, 376, 377 Indeed,
maternal Gal-1-P levels have been reported to rise toward the end of pregnancy and after delivery
possibly because of endogenous lactose production and subsequent catabolism to glucose and
galactose. 61, 327, 377 However, there have been no apparent negative consequences of this transient
increase in maternal Gal-1-P flux.
Female Reproductive Health in GALT Heterozygotes

There is limited evidence that ovarian function may be abnormal in women who are heterozygous for
mutations linked to classic galactosemia. For example, women heterozygous for mutations or
polymorphisms associated with low GALT activity have been found to have significantly higher basal FSH
levels. 88 Nevertheless, a retrospective study of 103 obligate carriers (mothers of galactosemic children)
demonstrated that time to pregnancy and menstrual cycle patterns were no different from those of 116
control women (mothers of children with phenylketonuria). 390 In this study, however, it was not possible to
determine whether menopause occurred earlier in the galactosemia carriers because the number of
women over the age of 45 in the study was small. A recent study of Indian women with premature ovarian
failure also demonstrated no evidence of a greater frequency of GALT mutations. 244 Finally, Knauff and
colleagues recently tested the serum levels of anti-M?an hormone (AMH) in carrier women and also found
no evidence of ovarian failure. 244a
GALT carrier status also has been linked anecdotally to an increased incidence of Mullerian fusion
defects, although this association remains controversial. For example, Mullerian aplasia and streaklike
ovaries were reported in one girl whose mother also had a Mullerian fusion anomaly. 19 The girl, her
mother, and her maternal grandmother were all heterozygous for mutations associated with classic
galactosemia. Another study reported that heterozygosity for a recognized GALT variant allele (N314D)
was more common in a series of 33 women with endometriosis than in control individuals (30 percent
compared with 14 percent), as well as in 13 females with vaginal agenesis (46 percent compared with 14
percent in control individuals). 85, 87 However, subsequent studies have not supported an association
between GALT carrier status and endometriosis, 169, 173, 301, 427 and there have been no reports of
Mullerian fusion defects in infants born to mothers homozygous for the N314D allele.
Possible Association of Galactose Metabolism with Increased Ovarian Cancer Risk

Impaired galactose metabolism, coupled with high levels of lactose consumption, also has been linked
with a higher risk for ovarian cancer in some studies but not in others. For example, in one U.K. study, a
statistically significant increase in the frequency of the N314D substitution was observed in women with
serous and undifferentiated histologic subtypes of ovarian cancer but not with mucinous, endometrioid, or
clear cell subtypes. 301 Further, an independent study reported an increased frequency of homozygosity
for N314D in ovarian cancer cases, but in contrast to the U.K. study, these were most evident in
endometrioid and clear cell subtypes. 86 In these patients, GALE activity also was significantly lower in
ovarian cancer patients compared with control individuals. In a third study, only abnormally low GALT
activity was linked, albeit not significantly, with the risk of ovarian cancer. 475 Most recently, two large
studies from the United States reported no effect of N314D genotype or of lactose/galactose intake or of
reduced galactose metabolism with ovarian cancer. 84, 158
Growth, Bone Density, and Body Composition

A number of studies have reported abnormalities of growth, bone density, and/or body composition in
patients with classic galactosemia. There is no clear evidence of prenatal growth restriction, 336 but
postnatal growth rates may be delayed. Two early studies reported that growth is often delayed in
childhood and early adolescence but that growth also often continues beyond the age of 18 and into the
early 20s, 377, 463 so final stature may be within the normal range. Delayed growth in females may be
related to ovarian dysfunction, although it also occurs in male patients, 336, 377 challenging the notion that
ovarian failure alone accounts for the pattern. Whether this pattern is a consequence of inadequate
nutrient intake in childhood or a feature of galactosemia itself is also unclear.
Decreased bone density has been reported to occur in both children and adults with classic
galactosemia. 220, 335, 371 Whether this complication is secondary to dietary insufficiency of calcium, 373
related to ovarian failure, 220 or rather reflects a true downstream effect of GALT deficiency 220, 371 has
long been a point of contention. A recent study of 40 children with classic galactosemia 335 revealed that
mean areal bone mineral density (BMD) Z-scores of the lumbar spine and of the femoral neck were
decreased. Mean volumetric BMD of the femoral neck also was significantly lower in galactosemics (P <
0.001). Of note, the recommended dietary allowances (RDAs) for calcium, magnesium, zinc, vitamin D,
and protein were met in all patients, and mean serum levels of calcium, phosphate, magnesium, zinc,
1,25-dihydroxy vitamin D (1,25OHD), parathormone (PTH), 17β-estradiol, bone alkaline phosphatase
(BAP), and undercarboxylated osteocalcin (ucOC) were normal, suggesting that dietary insufficiency
cannot account for the problem. Further, serum levels of insulin-like growth factor 1 (IGF-1), carboxylated
osteocalcin (cOC), N-terminal telopeptide (NTX), and C-terminal telopeptide (CTX) also were significantly
lower in galactosemics than in control subjects. Together these data support the conclusion that the
decreased bone density observed in patients with classic galactosemia reflects decreased bone
metabolism, 335 although the underlying mechanism remains unclear.
With regard to body composition, height Z-scores, mean fat mass, lean tissue mass, and IGF-1 all were
reported to be decreased in a group of Dutch children and adolescents with classic galactosemia. 334
Notably, lean tissue mass (adjusted for height) but not fat mass correlated with IGF-1 Z-score. Panis and
colleagues 334 hypothesized that the lower IGF-1 might be related to abnormal glycosylation secondary to
GALT deficiency.
Prevention of Osteoporosis
Soya-based infant formulas normally provide adequate calcium in infancy; however, by 1 year of age, the
volume of soya formula taken begins to fall, so it becomes necessary either to substitute a commercially
available calcium-enriched soya milk or to give adequate calcium supplements. In a recent study of 19
patients, only 5 met the reference nutrient intake for calcium despite the use of calcium-supplemented
milk and regular dietetic counseling. 373 Panis and colleagues 338 recently reported the results of a 2-year
randomized, double-blind, placebo-controlled clinical trial in which 40 children with classic galactosemia
(13 males and 27 females aged 3 to 17 years) were chosen to receive either 750 mg calcium, 1.0 mg
vitamin K(1), and 10.0 µg vitamin D(3) or placebo daily. Bone mineral content (BMC) of the femoral neck,
lumbar spine, and total body, as well as body composition data, was determined by dual energy x-ray
absorptiometry (DXA) at baseline and after 1 and 2 years. Diet was assessed, and biochemical
measurements also were determined at these times. In the children receiving treatment, cOC
concentration increased significantly (P < 0.001) and ucOC concentration decreased significantly (P =
0.001) compared with the children receiving placebo. BMC of the lumbar spine also increased (P = 0.001),
as did lean tissue mass (LTM; P = 0.016) and fat mass (FM; P = 0.014) in the treatment group compared
with the placebo group. The significant increase in cOC and decrease in ucOC concentrations were
observed in both prepubertal and pubertal children in the treatment group, although the significant
increase in BMC of the lumbar spine was seen only in prepubertal children in the treatment group.
In a follow-up to this study, Panis and colleagues 337 have proposed further measures toward the
prevention of early osteoporosis in galactosemia. Specifically, these authors proposed that DXA should be
used for bone mass assessment beginning at age 4 years. In children and adolescents, total-body BMC,
as well as LTM, also should be measured and compared with reference values for healthy control subjects
to allow for correction of the confounding effect of bone size. DXA should be repeated every 2 years in
cases of normal BMC. If BMC is low, lifestyle factors such as physical activity, intake of calcium and
vitamins K and D, and estrogen supplementation (in girls) should be optimized. DXA should be repeated
yearly in the event of a BMC below 0 SD in order to identify deteriorations and improvements early.
Pathophysiology
General considerations
The pathophysiology of classic galactosemia remains largely unexplained, although a number of
interesting hypotheses have been put forward, as described below. In considering these hypotheses,
several key questions must be addressed: Do the acute and long-term sequelae of classic galactosemia
reflect the same underlying mechanism? When does the pathology leading to long-term complications
begin: in utero, after exposure to dietary galactose, or after prolonged exposure to endogenously
produced galactose? Are the different organ systems (e.g., brain, ovary) affected by different forms of
pathophysiology, or is the underlying pathophysiology the same?
Timing
A number of studies have addressed the issue of timing with regard to pathophysiology and patient
outcome. For example, Waggoner and colleagues reported that IQ did not correlate with the age when
treatment was begun. 463 This negative result was emphasized in a study of affected siblings in which the
older siblings, treated beginning 1 to 11 months after birth, performed as well in psychometric tests as did
the younger siblings treated from birth. 105
The failure to correlate long-term outcome with the length of exposure to dietary galactose suggests the
possibility that damage might begin in utero. This idea is supported by evidence that the pathways of
galactose metabolism normally develop around the tenth week of gestation 410 and that the galactosemic
fetus has abnormal levels of metabolites from about that time. 9, 315 Several studies have further indicated
that clinical features of galactosemia also may have their origins in utero. For example, cataracts have
been seen in patients at birth 104 and even in a fetus at 20 weeks’ gestation. 458 Further, histologic studies
of the livers of infants who died of galactosemia have concluded that cirrhosis originated in utero, 240 and
there is a possibility that ovarian failure may be determined prenatally. 135
Galactitol
Galactitol accumulation in the ocular lens likely accounts for the galactose-dependent cataracts observed
in patients with classic galactosemia because it clearly accounts for cataracts in patients with GALK
deficiency (explained earlier). However, since none of the other acute or long-term clinical complications
of classic galactosemia are seen in GALK deficiency, it seems unlikely that galactitol accumulation in
other tissues is the culprit behind the other sequelae.
Gal-1-P
As explained earlier, Gal-1-P accumulates to abnormally high levels in the blood and tissues of patients
with classic galactosemia; dietary restriction of galactose mitigates this abnormality but does not fully
resolve it. Elevated Gal-1-P has been proposed as potentially toxic for a variety of reasons (reviewed in
ref. 261). For example, Mayes and Miller 288 suggested that accumulation of Gal-1-P could lead to futile
cycles of phosphorylation in which ATP is trapped in Gal-1-P, which then may be dephosphorylated to
metabolites that have no energy-generating capability. There is some support for this hypothesis, which
has a parallel in the metabolism of fructose in hereditary fructose intolerance, from animal model systems
in which liver necrosis was induced by galactose analogues. 426 Indirect support also comes from two
other groups 125, 238 that each reported that a galactose load given to adult galactosemic patients induced
hyperuricemia, ostensibly reflecting increased catabolism and depletion of hepatic nucleotides.
An alternate mechanism for Gal-1-P toxicity was proposed by Gitzelmann, 147 who speculated that
elevated levels of Gal-1-P might inhibit essential glucose-metabolizing enzymes, e.g.,
glucose-6-phosphatase, glucose-6-phosphate dehydrogenase, phosphoglucomutase, glycogen
phosphorylase, and/or UDP-Glc-pyrophosphorylase. Of note, Roth and colleagues also have reported that
high concentrations of Gal-1-P inhibit galactosyl transferase, 370 which could at least in part explain the
glycosylation defects associated with classic galactosemia (described below). More recently, other
possible roles have been proposed for Gal-1-P in the pathophysiology of galactosemia. For example,
each of several groups have noted that Gal-1-P may inhibit the function of inositol
monophosphatase, 46, 289, 341, 417 thereby potentially affecting inositol production or metabolism in brain
and other organs. Slepak and colleagues 416, 417 have further proposed that Gal-1-P accumulation in
GALT-deficient cells exposed to galactose may lead to a cellular stress response.
Surprisingly, attempts to correlate long-term outcome in GALT deficiency (as measured by IQ) with
average Gal-1-P levels on treatment have been unsuccessful. 78, 463 This lack of correlation could imply
that while Gal-1-P may play an important role in the pathophysiology of acute symptoms, it does not
contribute to long-term complications. Alternatively, the absence of a correlation between blood Gal-1-P
and cognitive outcome in patients simply may reflect a timing issue; the principle window of Gal-1-P
toxicity may occur early in utero.
UDP-Gal and UDP-Glc

As described earlier, some authors have reported abnormal levels or ratios of UDP-Gal and/or UDP-Glc in
cells or blood samples from patients with classic galactosemia. If true, this metabolic abnormality
potentially could contribute to the aberrant biosynthesis of glycoproteins and glycolipids reported in
patients 316, 404 (discussed below). Indeed, Lebea and Pretorius 254 have further postulated that
abnormally low levels of UDP-Gal in galactosemics could impede the function of cerebroside galactosyl
transferase, which is the enzyme responsible for galactosylation of cerebrosides in the brain.
Glycosylation defects
A literature trail reaching back more than 35 years documents evidence of abnormal glycosylation in
patients with classic galactosemia. For example, Haberland and colleagues 168 found an abnormality in
the brain glycoproteins of a galactosemic child at autopsy and suggested that the cause could be a defect
in galactose incorporation. Witting and colleagues 488 observed a similar phenomenon. Petry and
colleagues 343 reported a deficiency of normal galactose and N-acetyl galactosamine–containing
glycolipids and an accumulation of their precursors in the brains and lymphocytes of galactosemic
patients. A deficiency of galactosylceramide (Gal-Cer), one of the characteristic brain glycolipids and a
marker for myelinating cells, was regarded as an indicator of impaired myelin development. The authors
also observed an inexplicable increase in glucosyl ceramide (Glc-Cer). Notably, in a ceramide galactosyl
transferase knockout mouse that makes no Gal-Cer, replacement of Gal-Cer by Glc-Cer causes a more
compact myelin that results in changes in MRI signal. 58
Since UDP-Gal is the substrate for the galactosylation of glycolipids and glycoproteins, Daude and
colleagues 97 explored the differential expression of GALT in rats during periods of myelination in vivo and
in vitro. GALT expression appeared weak in brain and peripheral nerve tissue during late embryonic
development, and expression was highest in the postnatal period that correlates with the time of maximal
brain myelination in rodents. The authors concluded that the altered myelin signal observed in
galactosemic patients might reflect, at least in part, inappropriately low levels of GALT activity during
critical periods of brain maturation.
Further evidence of abnormal glycosylation in patients with classic galactosemia was reported by Dobbie
and colleagues 102 and Ornstein and colleagues, 330 who by quite different approaches demonstrated that
there is a low uptake of galactose into the glycoproteins of cultured skin fibroblasts from GALT-deficient
patients compared with control subjects. Kadhom and colleagues 215 similarly demonstrated low efficiency
of [ 14 C]galactose incorporation into GALT-deficient cells, and Wolfrom and colleagues 491 reported
impaired glucose uptake by galactosemic fibroblasts in cultured cells relative to control cells. As
mentioned earlier, Prestoz and colleagues 352 demonstrated abnormal isoforms of FSH in classic
galactosemic patients. These authors concluded that N-acetyl galactosamine sites on the abnormal FSH
were insufficiently galactosylated and further postulated that this glycosylation defect in the pituitary might
underlie the ovarian dysfunction seen in female galactosemia patients.
Jaeken and colleagues 207 suggested that the pathogenetic mechanisms in classic galactosemia may be
similar to those found in the congenital disorders of glycosylation (CDGs), which comprise a major group
of abnormalities involving the nervous system and ovarian failure in some forms. The similarities observed
included decreased thyroid-binding globulin and abnormalities in the isoelectric focusing pattern of the
sialo transferrins. Since abnormal lysosomal enzymes have been found in CDGs, the authors went on to
study two newborn galactosemic babies and found that prior to the onset of treatment, both infants had
abnormal isoelectric focusing (IEF) patterns of serum β-N-acetylhexosaminidase and α-fucosidase; the
abnormal IEF patterns normalized after a week of treatment. The authors postulated that in classic
galactosemia, increased Gal-1-P levels interfere with carbohydrate processing at the Golgi
galactosyltransferase step, leading to a partial deficiency of the terminal disaccharide (sialic acid +
galactose). Consistent with this hypothesis, Bonham and colleagues 51 demonstrated a correlation
between urinary transferrin:creatinine ratio and blood Gal-1-P levels in galactosemic patients on
treatment.
Finally, three further reports by different groups further define the nature and mechanism of abnormal
glycosylation in galactosemia. Stibler and colleagues 430 found the presence of carbohydrate-deficient
isoforms of serum transferrin in the blood of patients with classic galactosemia. In particular, these authors
found abnormal asialo- and/or disialotransferrin in samples derived from untreated patients; these
carbohydrate-deficient isoforms were observed rarely in samples from patients on dietary restriction of
galactose. Charlwood and colleagues 69 corroborated those results and then extended them by applying
HPLC analysis of transferrin-derived N-linked glycans to explore in greater detail the structural basis of the
glycosylation defects in patients. These authors found that the transferrin glycans from untreated patients
were more heterogeneous than were their normal counterparts; they contained four major truncated
glycans in addition to a smaller amount of the disialylated biantennary complex type. 69 These authors
concluded that the truncated glycans were deficient in sialic acid and galactose and that their structures
suggested inadequate galactosyltransferase activity in the biosynthetic tissue, assumed to be liver. Of

note, while the serum transferrin glycans largely normalized in patients on dietary restriction of galactose,
they did not become completely normal in all patients, even after prolonged treatment. More recent
studies of serum transferrin from untreated patients with classic galactosemia 431 again corroborate the
earlier conclusions but also extend those conclusions to demonstrate that the glycosylation defects fall
into two distinct categories, representing defects in both N-glycan assembly and processing.
GALT Deficiency in the Absence of Significant Disease
Mice
More than a decade ago, Leslie and colleagues 263 applied gene targeting to generate a GALT-null
mouse. These mice demonstrated liver GALT levels that were undetectable and therefore comparable
with tissue levels seen in humans homozygous for the common Q188R mutation. Furthermore, tissue and
erythrocyte levels of Gal-1-P and galactose were extremely high in these mice. Surprisingly, however, the
GALT-deficient mice remained healthy and fertile, even when fed a high-galactose diet.
The disparity between the mouse and human responses to GALT loss leave open the question of why
GALT deficiency results in such negative outcomes for humans. One hypothesis raised by Segal 394 is
that the early toxic signs of classic galactosemia may reflect a combined effect of Gal-1-P and galactitol
accumulation. Mice expressing only weak aldose reductase activity accumulate less galactitol and hence
would escape the early toxic presentation. 394 Further investigations in suckling and 7-week-old
GALT-deficient mice demonstrated that these animals are able to oxidize some galactose, 320 presumably
via the UGP pathway explained earlier. 260 Combined, these data suggest one of two alternative
hypotheses: Either high Gal-1-P levels alone may be insufficient to cause galactose toxicity, or
alternatively, whatever factors sensitize humans to high levels of Gal-1-P may not be recapitulated in
mice.
Humans
Of note, there also have been reports of two patients with classic galactosemia who each went off dietary
restriction of galactose as young children and yet who only experienced mild outcomes in
adulthood. 255, 333, 384 These reports raise the controversial question of whether lifelong dietary restriction
of galactose is truly necessary for all patients. Further, it remains unclear whether these two patients
represent a significant fraction of the galactosemia community or may be outliers.
Mutations at the GALT Locus

As of December 2007, more than 200 different mutations had been identified in the GALT alleles of
patients with transferase-deficiency galactosemia (reviewed in refs. 63, 452, and 455); subsets of these
mutations are catalogued in each of two current online databases (
www.arup.utah.edu/database/galactosemia/GALT_welcome.php and
www.hgmd.cf.ac.uk/ac/gene.php?gene=GALT). Most of these mutations are missense point mutations;
silent, nonsense, and noncoding changes also have been reported, as has one large (~5 kb)
deletion 79, 455 and a number of small deletions. Several apparent polymorphisms also have been
reported 63 (e.g., HapMap, www.hapmap.org/). Further reports of novel mutations continue to appear. As
is the case for many other genetic disorders, the frequencies of specific GALT alleles appear to vary by
population or racial group, likely reflecting founder effects or other genetic forces (discussed below).
Specific mutations and population allele frequencies

As is the case for many other genetic disorders, such as cystic fibrosis (see Chap. 201) and
phenylketonuria (see Chap. 77), a small number of mutations at the GALT locus are frequent, but most
are rare. More than half the total base changes have been documented in only one study, and except for
the base changes in linkage disequilibrium (LD) with N314D (see "N314D and the Duarte Alleles" below),
fewer than 10 percent of mutations have been found in multiple human groups or geographic
populations. 454 Among the most frequently observed mutations are Q188R, K285N, and S135L, of which
only S135L is at a CpG hypermutable site. N314D is perhaps the most common variant allele observed in
humans from many populations, but as explained below, this change appears to be a GALT polymorphism
that can exist in LD with causal mutations rather than a causal mutation itself.
Mutations of European origin (Q188R and K285N)

Q188R is the most frequently occurring GALT mutation associated with classic galactosemia in European
populations and in people of predominantly European descent.The Glu-to-Arg substitution at amino acid
position 188 of the GALT protein results from an A→G transition at nucleotide c.563 in exon 6. First
reported in 1991 by Reichardt and colleagues, 358 this mutation is located in a domain highly conserved
across species and is in close proximity to the putative enzyme catalytic site, which is a
histidine-proline-histidine triad. 119
Although Q188R originally was thought to account for as few as 25 percent of mutant alleles in North
American patients, 358 subsequent studies of larger patient cohorts revealed an allele frequency of closer
to 60 to 70 percent. 56, 110, 112, 262, 274, 361 Ng and colleagues 317 and Wong and colleagues 492 recorded
the mutation’s relative frequency in North American patients more precisely by differentiating between
Q188R alleles in individuals of principally European descent (63.5 percent) and Hispanics of Mexican
descent (50 to 58 percent). In fact, taking Europe as a whole, the overall frequency of Q188R on more
than 700 mutant chromosomes studied was 64 percent, but there was considerable variation in the
relative frequency in individual European populations. The highest frequencies are seen in the Republic of
Ireland, 305, 306 in both the Traveller and non-Traveller populations, 306 and in Great Britain 454 ; the
frequency decreases in continental Europe with movement through populations in an eastern direction.
The allele frequency of Q188R in galactosemic patients from Spain and Portugal is between 50 and 60
percent. 159 Hungary has the lowest frequency reported in a European population at only 33.3 percent. 195
The relative frequency of Q188R in North American Hispanics has been estimated to be 50 to 58
percent, 317, 492 which is similar to the frequency seen on the Iberian peninsula 159, 454 (Boleda, personal
communication). A study of patients of Caucasian, Hispanic, and African-American heritage in Texas 500
established that Q188R accounts for 37 percent of the combined galactosemia alleles in this mixed
population. Of note, Q188R is not seen in patients of Asian descent. 17, 317
The high relative frequency of Q188R in populations of European descent and its absence in Asians
indicate that this mutation arose after the migration out of Africa and also after the subsequent divergence
between Caucasian and Asian populations. Although a single ancestral origin has yet to be proved, the
high relative frequency of Q188R in Ireland and the British Isles suggests that at least one point of origin
traces to northwestern Europe, possibly Ireland, and that the observed gradient seen in Europe has
occurred through migration and admixture.
Patients homozygous for the Q188R mutation demonstrate essentially no residual GALT activity in red
blood cells 317 or transformed lymphoblasts, 129 and homozygosity for the Q188R GALT allele has been
associated with a poorer cognitive outcome 402 and increased risk for verbal dyspraxia compared with
other genotypes. 366
A second mutation that is common among patients of European origin is K285N, which occurs with an
overall frequency in Europe of about 8 percent. As is the case with Q188R, there are large differences in
the relative frequency of K285N in different populations. In many areas it is the second most common
disease-causing GALTmutation, and in some countries of eastern and central Europe, it accounts for 25
to 35 percent of mutant chromosomes. 163, 242, 273, 501 This allele’s frequency distribution suggests that the
center of diffusion for K285N lies farther to the east than does that for Q188R.
More recently, the absolute allele frequencies of the four most common mutations were determined by
screening almost 4800 adults. 432 The allele frequency of Q188R was found to be 1 in 344 in whites
(thereby 1 person in 172 carrying this mutation), with just half this frequency seen in individuals of
Hispanic origin. K285N occurred on 1 in 1612 alleles and was found only in individuals of western and
southern European origin. Using an "index mutation method" originally described by Beutler and Gelbart 45
as applied to the prevalence of pyruvate kinase deficiency, Suzuki and colleagues 432 calculated that the
overall frequency of galactosemia in whites is about 1 in 47,000.
Mutations of African Origin (S135L and F171S)

One mutation, a Ser-to-Leu substitution at codon 135, designated S135L, is almost exclusive to black
Africans and African-Americans. The mutation was reported originally by Reichardt and colleagues, 362
and additional studies 250, 492 have established that S135L accounts for almost 50 percent of mutant GALT
alleles in African-Americans. A second mutation, a Phe-to-Ser substitution at codon 171, designated
F171S, also first reported by Reichardt and colleagues, 362 is much rarer, accounting for only about 4
percent of mutant alleles in African-American patients. 492
In one study of African patients, Manga and colleagues 281 reported that S135L accounted for 91 percent
of mutant chromosomes in galactosemic patients in South Africa. The frequency of this allele in 600
healthy black South Africans was 1 in 150 (the carrier frequency being 1 in 75). A similar study led
Henderson and colleagues 177 to conclude that the carrier rate for S135L in the black population of Cape
Town, South Africa, was 1 in 60. Although the sample number was low for other ethnic groups, the
mutation was not found in the Khoisan population of other southern African countries, including Angola,
Botswana, and Namibia. 281 Two of 292 individuals screened from countries in central Africa (including
Cameroon and the Central African Republic) and 1 of 199 individuals from western Africa (including
Guinea, Sierra Leone, Ivory Coast, Ghana, and Senegal) carried the S135L mutation. Applying the index
mutation method to the data of Manga and colleagues, 281 Suzuki and colleagues 432 estimated that the
frequency of galactosemia in individuals of African descent is about three times higher than in Europeans.
Of note, the Q188R mutation, found in three white galactosemics in South Africa, was not detected among
the black population. 281 In African-Americans, however, this mutation is found at relative frequencies
varying from 12 to 21 percent, 112, 250, 317, 492 which approximates the expected frequency resulting from
European admixture. 68 Approximately 30 to 35 percent of African-American mutant GALT chromosomes
remain uncharacterized. It is probable that, as with other major human groups and other disorders, these
mutations will prove to be multiple and varied.
The almost exclusive occurrence of S135L in individuals of African origin suggests that this mutation is
relatively new and that it arose after the first wave of migration of Homo sapiens out of Africa more than
60,000 years ago. The C→T transition resulting in S135L is at a hypermutable site. In African blacks, it is
found at varying frequencies, and it appears to be rare or absent in southwestern Africa. 281 The ancestors
of present-day African-Americans originated from a number of geographic locales in western Africa. 68
The high relative frequency of S135L in southern African blacks suggests either that they could have been
a source of slaves in North America or that they have close affinities with the populations from which most
of the slaves were taken. 281 Genetic drift and subsequent range expansion also could have contributed to
the high relative frequency of S135L in African-Americans.
The S135L mutation is also noteworthy because of the low-level residual activity associated with this
allele 252, 479 (discussed below). This allele is often associated with milder clinical outcomes in patients,
and this allele is also believed to account for the so-called Negro variant of GALT deficiency originally
described by Baker and colleagues. 21
Mutations of Asian Origin

Classic galactosemia is extremely rare in Asian populations, with an incidence estimated at 1 in 400,000
among Chinese 72 and 1 in a million among Japanese. 15 A Thai infant diagnosed with classic
galactosemia also has been reported, although mutational analysis was not performed. 52 The common
European mutations, Q188R and K285N, are rare in Asian galactosemics. 17, 317 Efforts at GALT
genotyping of Japanese patients with classic galactosemia have revealed notable allelic
heterogeneity. 17, 184 For example, a study of 23 mutant alleles revealed 14 different mutations, 10 of
which had not been seen in Caucasians. 184
Mutations in Other Populations

In the Traveller population of the Republic of Ireland, where the incidence of transferase-deficient
galactosemia is estimated to be 1 in 550, Q188R is the sole mutation found. 306 The relative frequency of
the Q188R allele in this group is an order of magnitude higher than that in the non-Traveller population
(0.046 compared with 0.005). Evidence does not support population inbreeding as a cause of this high
frequency in the Travellers. Rather, founder effects combined with demographic factors similar to those
that have acted on the Saguenay population 182 have been proposed as a possible explanation.
In individuals of Indian and Pakistani ancestry, Q188R is very rare, 317 and in the British Isles, only
"private" mutations have been found in affected individuals of Arabic or Indian ancestry. 453 A recent study
of GALT mutations in Iranian galactosemic patients found that Q188R accounted for approximately 57
percent of mutant alleles; S135L, Y209S, and A320T each accounted for close to 7 percent; and K285N
accounted for 3.57 percent of mutant alleles. 292
The only large deletion allele of GALT has been reported in individuals of Ashkenazi Jewish
origin 22, 56, 79, 304 (Tyfield, personal communication). This deletion removes the majority of the GALT
locus and may be missed by many conventional strategies of mutation detection, potentially leading to
aberrant genotyping conclusions. 22
In Turkey, 8 different base changes have been characterized on 32 mutant alleles. 400 Q188R is by far the
most common mutation, accounting for 57 percent of mutant chromosomes, 83 a relative frequency
considerably higher than in Italy 277 and even higher still than in the Czech and Slovak Republics. 242 Four
of the remaining 7 mutations have been reported in German patients, and 1 mutation, a Phe-to-Tyr
substitution at codon 294, designated F294Y, is novel. 400, 454 Whether this mutation is unique to the
Turkish population is not known. Nevertheless, it is certain that admixture between Turkish and Germanic
populations is responsible for the appearance of rare mutations in both populations.
Functional impact of specific mutations

At least five different strategies have been applied to assess the functional impact of patient mutations in
GALT. These include (1) whole-body galactose oxidation studies in patients, 23, 34, 36, 38, 250 (2) enzymatic
or metabolic studies of patient samples or cells in culture, 129, 167, 248, 313, 317, 318 (3) studies of nonhuman
mammalian cells or yeast cells engineered to express patient alleles of GALT, 17, 129, 184, 358, 362, 365 (4)
studies of recombinant proteins purified from yeast 89 or E. coli hosts, 251 and (5) in silico predictions
based on sequence homology lineups and the known properties of amino acids. Each method has its
strengths and limitations, as explained below. While none of these methods is perfect, when used in
appropriate combinations, they have yielded a wealth of information concerning the functional significance
of patient alleles of GALT.
Whole-Body Galactose Oxidation Studies

Whole-body galactose oxidation studies are conducted in vivo, which is a clear strength, but the results
are not specific to GALT function; variation in other Leloir enzymes or bypass pathways can influence
outcome. Further, subtle tissue-specific differences in GALT function may be obscured, and for
heteroallelic patients, it may be difficult to distinguish the functional significance of individual mutations.
A number of groups have applied whole-body galactose oxidation studies to estimate the level of GALT
function in patients of known genotype. For example, Berry and colleagues 34 used a galactose breath test
that quantified [1- 13 C]galactose conversion to 13 CO 2 2 hours after administration of a bolus of galactose.
Individuals homoallelic for Q188R and with no detectable erythrocyte GALT activity eliminated less than 2
percent of the administered galactose compared with 8 to 28 percent in control subjects. This level of
oxidation defined a severe metabolic phenotype. Indeed, patients homoallelic for Q188R metabolized
levels of galactose indistinguishable from those metabolized by patients homoallelic for almost complete
deletion of the GALT locus. 35, 496 In a similar manner, other mutations, including L195P, E308K, V151A,
M142K or Q344K, and K285N, also were defined as resulting in a severe metabolic phenotype. Notably,
however, for some mutations, whole-body galactose oxidative capacity at 2 hours did not necessarily
predict the clinical outcome. S135L, for example, did not impair total-body galactose oxidation whether in
a homoallelic state or in combination with a severe mutation, 34 although lymphoblasts from these patients
demonstrated clear metabolic abnormality, 496 and these patients exhibited signs of clinical disease,
especially when on a high-galactose (milk) diet.
Enzymatic or Metabolic Studies of Patient Samples or Cells

Most patients with classic galactosemia are diagnosed on the basis of impaired GALT activity and/or
elevated galactose metabolites (e.g., Gal-1-P) in blood. Studies of these data therefore are clearly
valuable, although they remain subject to a number of serious limitations. For example, red blood cells are
not always representative of enzyme activity or metabolic profile seen in other tissues. 139 Interpretation of
data from heteroallelic patients also can be complex.
A number of groups have studied EBV–transformed lymphoblasts from genotyped patients to ascertain
the functional effects of specific GALT mutations. 35, 129, 496 For example, Fridovich-Keil and
Jinks-Robertson 129 demonstrated that extracts of lymphoblasts homoallelic for Q188R exhibited only
about 1 percent of the GALT activity detected in extracts from control lymphoblasts. Yager and
colleagues 496 found that lymphoblasts homoallelic for Q188R were able to oxidize galactose at about 3
percent of control values after 2.5 h, whereas cells homoallelic for S135L could oxidize substantially more
galactose (about 16 percent of controls). 496
Mammalian Cell (Chinese Hamster Ovary, CHO) and Yeast Expression

Studies of Human GALT Mammalian cell expression studies of human GALT allow for the evaluation of
individual alleles; expression of human GALT in a null-background strain of yeast also enables studies of
patient alleles, either singly or in combination, but only missense or nonsense coding mutations may be
studied, and subtle issues involving message or protein stability, in theory, may not be faithfully
recapitulated. Nonetheless, for all cases reported to date, the results from yeast have correlated well with
results from corresponding patient samples; this is not the case for the COS (CV-1 in Origin expressing
SV40 genetic material) cell system. For example, Reichardt and colleagues 358 reported COS cell studies
demonstrating that the Q188R mutation is associated with approximately 10 percent residual activity, a
result that is clearly in conflict with enzyme data from patient blood or cell studies, 129, 317, 350, 418
galactose oxidation studies, 34 and yeast expression studies. 129, 365 A similar problem arose for S135L;
data from COS cells 362 demonstrated near-normal activity, whereas other studies revealed that in patient
cells, 252, 496 yeast, 365 and purified recombinant protein, 247, 480 the true activity of S135L GALT is closer
to 5 to 10 percent wild-type. The reason for the disparity in outcome for these mutations between the COS
and other systems remains unclear but may result from interactions between endogenous hamster GALT
subunits and the expressed human GALT subunits. Notably, for some alleles, for example N314D, both
the COS 363 and yeast expression systems 130 gave consistent results.
One interesting application of the yeast expression system reported by Henderson and colleagues 179
addressed the covalent heterogeneity of human GALT (hGALT). These authors performed 2D western
blot analyses of soluble extracts of yeast expressing substituted human GALT alleles to demonstrate that
the Q188R, S135L, and N314D substitutions each had no effect on the formation of uridylylated
intermediate. This is consistent with earlier reports that Q188R hGALT or the E. coli equivalent Q168R
eGALT remained competent to form an uridylylated intermediate. 134, 251
The most comprehensive study of the relationship between GALT genotype, activity, and galactose
sensitivity using in vitro functional assays was published by Riehman and colleagues. 365 A yeast
expression system was used to explore the impact of 16 different mutations (R67C, S135L, L139P,
V151A, F171S, P183T, Q188R, R201H, R231H, R259W, K285N, E291K, N314D, Y323D, R333W, and
T350A) on the activities and abundance of human GALT proteins. Functional analyses for nine of the
mutations had not been reported previously. The results demonstrated convincingly that patient GALT
alleles represent a spectrum of activity and that not all are completely null.
This study also explored the impact of galactose exposure on strains of yeast expressing each mutant
allele. These studies extended the original observation of Douglas and Hawthorne 107 that GALT-null
yeast arrest growth in medium containing even trace quantities of galactose (e.g., 0.05 percent) despite
the presence of another metabolizable carbon source (e.g., glycerol/ethanol). With few exceptions, yeast
expressing the lowest-activity GALT proteins demonstrated the most significant growth impairment in the
presence of galactose, as well as the most pronounced and prolonged accumulation of Gal-1-P.
For most mutations, particularly those demonstrating null or very low activity, the yeast expression system
substantiated activities reported in humans carrying the corresponding mutations. Furthermore, the
galactose-sensitive growth of yeast expressing these alleles generally corresponded inversely with the
degree of residual GALT activity associated with the allele in vitro. In particular, strains expressing alleles
with approximately 10 percent or greater residual activity grew indistinguishably from strains expressing
wild-type GALT; strains expressing alleles with between 1 and 5 percent residual GALT activity
demonstrated intermediate growth impairment; and strains expressing alleles with no detectable GALT
activity completely arrested growth on exposure to galactose. Only one mutation, R67C, violated this
pattern, demonstrating more than 2 percent residual activity in vitro yet failing to restore any detectable
growth in the yeast. Reasons for this disparity are under investigation.
Purified Recombinant Protein

Studies of purified recombinant hGALT proteins have enabled precise kinetic evaluation of the functional
consequence of specific mutations. For example, Wells and Fridovich-Keil 480 applied a recombinant
approach to demonstrate that while the specific activity of S135L hGALT is 10-fold lower than wild-type,
the K m value differs by under 2-fold.
Crews and colleagues 89 addressed the structure/function relationships of codon 171 of human GALT.
The naturally occurring mutation, F171S, results in a total loss of enzyme activity. Crews and colleagues
systematically altered the physical properties of residue 171 by substituting each of three other amino
acids selected for their size, charge, or hydrophobicity in place of the normal phenylalanine and then
tested to see which substitutions resulted in a functional GALT protein. Four of the substituted alleles
(F171S, F171L, F171Y, and F171W) were studied in the yeast expression system, and the human
sequence was homology-modeled onto the published structure of the uridylylated E. coli enzyme. 440
These authors concluded that the altered catalytic efficiency associated with mutations at position 171 is
brought about by steric hindrance of Gln-188, thus preventing it from participating in catalysis.
A similar but more extensive analysis was applied by Quimby and colleagues, 355 who addressed the
importance of the proline (P185) in the highly conserved GALT active-site His-Pro-His triad. The crystal
structure of the E. coli enzyme suggests that this proline plays a role in positioning the active-site histidine
near the substrate. To examine the role of this proline in the human sequence, these authors performed
saturating mutagenesis at Pro-185 within human GALT and characterized each resulting mutant enzyme
using a null-background yeast expression system. Activity analyses in crude lysates indicated that only
proline at position 185 produced wild-type levels of activity, although five other amino acids, namely, Ala,
Gly, Ser, Gln, and Glu, all produced partially active enzymes. Western blot analyses of the GALT proteins
in these lysates demonstrated that abundance varied from 9 to 118 percent of wild-type and was largely
independent of activity. All five active mutant proteins were purified and characterized with regard to
specific activity, apparent K m for both substrates, and temperature dependence of activity. The results
demonstrated that position 185 of human GALT is important for the rate of reaction of the enzyme with the
first substrate, UDP-Glc, but less so for the reaction of the uridyl enzyme intermediate with the second
substrate, Gal-1-P. The properties of the substituted enzymes suggested that the substituting amino acids
could be clustered into five groups: Pro (wild-type), Ser and Ala, Gly, Glu and Gln, and all others. P185S
and P185A were each approximately 15-fold slower than wild-type at forming uridyl enzyme and had a
temperature optimum of about 37°C. The rate of uridylation of P185G was inhibited approximately
30-fold, with a temperature optimum also about 37°C. P185E and P185Q were approximately 80-fold
slower at uridyl enzyme formation and exhibited temperature optima of about 42°C. Finally, no other
mutants showed detectable enzymatic activity. These results were fully consistent with the accepted
catalytic model for this enzyme and also were consistent with the published x-ray structure of the E. coli
enzyme.
Most recently, Lai and Elsas 247 applied a recombinant approach to study the structure/function
relationships of residue 135 in human GALT. These authors overexpressed either wild-type or mutant
alleles of human GALT in a GALT-null strain of E. coli and then studied the resulting recombinant proteins
either in the context of host cell lysate or in purified form. The mutant alleles studied included A135, C135,
H135, L135, S132-H135, T135, and Y135; of these, only the threonine-substituted protein (S135T) had
more than 10 percent wild-type enzyme activity in lysates. To further characterize the S135L hGALT
protein, these authors performed kinetic studies of purified wild-type and mutant GALT proteins.
Consistent with the prior results of Wells and colleagues, 480 Lai and Elsas 247 found that the apparent
V max of the purified S135L hGALT protein was significantly reduced from 80 ± 5.9 to 5.8 ± 1.8 µmol
Glc-1-P released per minute per milligram of hGALT protein, with no increase in K m for Gal-1-P for the
second displacement.
In Silico Prediction
Finally, in silico approaches have been applied to the question of functional significance of patient
mutations. For example, the SIFT (Sorting Intolerant From Tolerant, http://blocks.fhcrc.org/sift/SIFT.html)
program uses sequence-homology alignments of related sequences, coupled with information about the
characteristics of specific amino acids, to predict amino acid substitutions that should be well tolerated
versus those that should not. Because it is fast and inexpensive, this methodology has become a
commonplace tool for predicting the functional significance of novel GALT mutations identified in some
clinical laboratories; however, the full reliability of those predictions for human GALT remains to be
confirmed.
Heteroallelic Subunit Interactions

Adding further complexity to GALT structure and function is the fact that many, if not most, patients are
heteroallelic at their GALT loci. Cells expressing two distinct alleles of GALT also express two distinct
subunits, assuming that both alleles encode stable proteins. Studies of human GALT expressed in yeast
have demonstrated that patient mutations can affect both subunit association and dimer function, resulting
in a significant impact on holoenzyme formation and function.
For example, while many coexpressed hGALT subunits assort independently, leading to the expected
1:2:1 ratio of dimer configurations, some do not. 75 Of note, one of the alleles demonstrating a skewed
ratio of subunit association was Q188R, the most common mutation found in Caucasian patients
(described earlier). Further evidence of subunit interaction was reported by Elsevier and Fridovich-Keil, 114
who found that Q188R/wild-type hGALT heterodimers exhibited only approximately 15 percent of the
activity seen in wild-type homodimers. Importantly, in these same experiments, R333W/wild-type
heterodimers exhibited almost exactly the expected 50 percent activity. Both Q188R and R333W hGALT
homodimers exhibited no detectable activity. 114 Data consistent with the conclusion of a "partial
dominant-negative" effect of the Q188R allele in heterozygotes were later reported from studies of patient
samples. 469 Of note, although GALT Q188R/wild-type heterodimers may exhibit only approximately 15
percent normal activity, given that 25 percent of the GALT dimers in heterozygous cells are
wild-type/wild-type, and therefore exhibit normal activity, the combined level of GALT activity in
Q188R/wild-type cells will be greater than 30 percent. While the full explanation for the impact of Q188R
on GALT heterodimer formation and function remains uncertain, recent homology modeling studies
performed by Marabotti and Facchiano 283 have offered an interesting possibility.
It is worth noting that an early paper by Nadler and colleagues 307 also reported evidence of "positive"
allelic or subunit interaction in GALT. These authors performed pairwise fusions of patient cells that were
themselves devoid of detectable GALT activity and in some cases were able to reconstitute at least partial
GALT activity. This apparent complementation has yet to be recapitulated under defined conditions
involving human GALT alleles that carry known mutations.
N314D and the Duarte alleles

One of the most common and best studied of sequence variants in human GALT is the substitution
N314D, 109, 262, 268 which current data suggest is a polymorphism with an allele frequency of
approximately 0.133 in populations of European ancestry (CEPH) and approximately 0.033 in populations
of Chinese ancestry, among others ( www.hapmap.org/index.html.en). N314D is not found with any
significant frequency in individuals of African descent, although owing to admixture, it is found in
African-Americans. This distribution suggests that the N314D variant arose after the first human
migrations out of Africa but before the divergence and radiation east and west. 17 Of note, cross-species
sequence alignments also demonstrate that the residue corresponding to N314 in the GALT enzymes of
mouse, rat, and dog are all D rather than N 130 (B. Coffee, personal communication).
The N314D allele has received significant attention in past decades in part because it is common, in part
because the encoded protein has an altered charge and therefore migrates aberrantly on native starch or
isoelectric focusing gels, 44, 225, 408 and in part because, in some contexts, it is associated with partial
impairment of GALT activity. While it is now clear that N314D itself accounts for the altered P i and
therefore altered native gel migration of the substituted GALT protein, 130 it is also now clear that other
mutations, occurring in cis with N314D, account for the altered expression and/or altered GALT activity
originally attributed by some investigators to N314D. In brief, N314D exists in LD with a number of other
base changes, some of which are common and others of which are rare (see below). The common
sequence variations that exist in LD with N314D fall into two sets. The first set includes N314D and the
silent base substitution L218L originally identified by Gathof and colleagues 133 and confirmed by
others. 162, 253, 349 Alleles that carry both N314D and L218L are referred to as Los Angeles (LA) or
Duarte1 (D1) alleles. These LA/D1 alleles exhibit normal or even above-normal GALT activity. 26 The
second set includes N314D together with three base changes in introns (IVS4nt-27g to c and IVS5nt-24g
to a, also called the Sac 1 polymorphism, 267 and IVS5nt+62g to a 242, 406 ) and a 4-bp deletion in the 5′
untranslated region 116 to 119 bases upstream from the first methionine codon. 243 These GALT alleles
are called Duarte2 (D2) alleles; D2 alleles demonstrate about 50 percent wild-type levels of GALT activity.
Of note, both D1 and D2 alleles are also reported to have an extended sequence of 28 adenine
nucleotides in intron 10 in place of 17 adenine nucleotides in the corresponding sequence of the normal
allele. 406
The functional significance of the N314D substitution and the D1 and D2 alleles was a point of intense
controversy in the early to middle 1990s, and the literature on this topic from that period is complex and in
some cases internally contradictory. For example, Andersen and colleagues 13 applied immunochemical
studies to demonstrate that the difference in net GALT activity associated with the D1 and D2 alleles in
patient blood reflected differences in GALT protein abundance rather than differences in specific activity.
Consistent with that conclusion, both COS cell 363 and yeast 130 expression studies demonstrated
essentially wild-type levels of activity from N314D GALT. The yeast studies further demonstrated that
while N314D did not impair the activity of GALT, it was sufficient to cause the characteristic isoelectric
focusing gel mobility shift noted earlier.
Podskarbi and colleagues, 349 who first reported the additional base changes in association with N314D
on the D1 and D2 alleles, proposed that the intron base changes on D2 may be involved in regulation of
GALT gene expression. However, Lai and colleagues reported that the RNA levels associated with both
D1 and D2 alleles were indistinguishable, proposing instead that the N314D substitution reduced stability
of the GALT protein. 248 These authors further proposed that the L218L silent mutation found on D1 but
not D2 alleles might overcome the inherent instability of N314D GALT by increasing translation of the
protein. 253
The dust began to settle when Kozak and colleagues 243 proposed that the 4-bp (GTCA) deletion in the 5′
region of the D2 allele might be functionally significant. Using a computer search, 176 they showed that two
binding factors (activator proteins AP1 Q2 and AP1 Q4) would lose their binding motifs in the D2 alleles.
Two years later, functional analyses carried out by Trbusek and colleagues 450 and Elsas and
colleagues 111 confirmed that alleles carrying the deleted promoter region are transcribed less efficiently
than the wild-type allele. These results strongly support the hypothesis that the 5′ deletion mutation at nt
−119/−116 is the major factor contributing to the diminished expression and activity of the D2 variant of
galactosemia.
Patients who carry a D2 allele on one chromosome 9 and carry a "classic galactosemia" allele (e.g.,
Q188R) on their other chromosome 9 are said to have Duarte galactosemia (D/G); these individuals
exhibit about 25 percent normal levels of GALT activity and are often picked up by newborn screening.
There is no substantial evidence to demonstrate that these individuals suffer any negative clinical
consequences, either acute or long term. Further, the work of Berry and colleagues 38 has shown that
individuals who are compound heterozygotes for Q188R and N314D or homoallelic for N314D
demonstrate normal in vivo oxidation of galactose. However, recent studies by Ficicioglu and
colleagues 118 demonstrated that at least some D/G infants exhibit elevated levels of red blood cell
galactitol, galactonate, and Gal-1-P on a normal diet but not on a galactose-restricted diet, raising concern
that these children may not be fully metabolically normal. Some centers currently recommend temporary
and/or partial dietary restriction of galactose for patients with D/G galactosemia, 151 whereas others do
not.
Allele Frequencies of the D1 and D2 Alleles in Different Populations

The prevalence of N314D in the nongalactosemic population is about 5 to 6 percent in North
America. 109, 268 In Germany, Podskarbi and colleagues 348, 349 differentiated between D1 and D2 alleles
and reported a relative frequency of 5.4 percent for D1 and 9.5 percent for D2 alleles. In the Czech and
Slovak Republics, the relative frequencies are slightly lower, with the D1 mutations occurring on 2.8
percent of 1008 control chromosomes and D2 mutations occurring on 5.4 percent. 243 In Japan, the
prevalence of the Duarte allele is 2 percent. 200
These findings were corroborated in a study of adult subjects reported by Suzuki and colleagues, 432 who
found significant differences in the frequency of both D1 and D2 alleles between different populations and
major human groups. D2 alleles occurred on 6 percent of chromosomes in non-Hispanic whites and on
4.5 percent of chromosomes in Hispanics. The frequency in African-Americans and Asians was
significantly lower, at 1.3 and 1.5 percent, respectively. Interestingly, Native Americans had the highest
frequency of D2 alleles, at 11.8 percent of chromosomes, and this was significantly different from other
human groups. A geographic difference also was found in European populations, the highest being in
northern Europe (7.9 percent) and the lowest in southern Europe (2.0 percent).
In some human groups, D1 alleles (N314D with L218L) occurred less frequently than D2 alleles, whereas
in others, they occurred with the same or at even higher frequencies. In African-Americans and Pacific
Islanders, for example, D1 and D2 alleles were found on virtually the same numbers of chromosomes. In
non-Hispanic whites, Hispanics, and Native Americans, on the other hand, the frequency of D1 was about
half that of D2, and so, once again, the frequency was highest by a factor of 2 in Native Americans.
Although this pattern also was observed in northern, western, and eastern Europe, it was reversed in
southern Europe (Greece, Italy, Macedonia, Portugal, and Spain), where D1 was found on twice the
number of chromosomes as D2.
Compound Effects of Mutations on D1 and D2 Alleles

N314D also has been identified in cis with ostensibly profound mutations in GALT. Individually, these
alleles are rare; nevertheless, biochemical and clinical phenotypes that have been reported in affected
individuals carrying multiple mutations are worthy of mention. For most of the mutations, biochemical
effects in association with D1 or D2 alleles have not been investigated in vitro.
Three nonsense mutations have been reported on either D1 or D2 alleles. R204X, found in association
with D2, affects sequences upstream from the N314D mutation and is a nonsense mutation that results in
a complete absence of accumulated GALT protein or activity when expressed in yeast. 73 The phenotype
of individuals carrying this mutation would be expected to be consistent with one of hemizygosity for the
mutation on the other allele. In fact, four children from the British Isles who are heteroallelic for R204X and
either Q188R or L195P all demonstrate negligible erythrocyte GALT activity, and one infant
(R204X/L195P) died in the neonatal period with severe sepsis 453 (Holton, personal communication).
E340X, on the other hand, is downstream of N314D and occurs on D1 alleles that also carry the L218L
mutation. One Turkish patient was reported who is homoallelic for E340X on D1 alleles, and although
erythrocyte GALT activity in this patient was negligible, there was no evidence of developmental or
neurologic impairment at the age of 8. 133 On the other hand, of two siblings homoallelic for E340X on D1
alleles, one suffered considerable neurologic damage. 388 The third nonsense mutation reported in cis with
N314D is W316X, which is located 2 amino acids downstream from N314D on D2 alleles. Some residual
enzyme activity (1.5 percent of normal) was detected in the erythrocytes of the affected individual in whom
this allele was first reported. 418 This patient also had R67C on the other allele, so the impact of each
mutation on phenotype is difficult to ascertain.
When N314D occurs in cis with a missense mutation, there may be some anomalies in the compound
effect in different individuals. E203K, for example, has been reported on chromosomes both with and
without N314D in cis. 113 It is clear that the mutation on its own effects a reduction in enzyme activity, but
in cis with N314D, the overall effects on biochemical phenotype are more abstruse and may be influenced
by the genotype of the other allele. Lai and colleagues proposed that E203K acts as a "revertant" to
N314D in cis, restoring thermal stability and enzyme activity to normal. 248 However, an affected individual
heterozygous for the E203K/N314D allele and a splice-site mutation (IVS3nt-2 a to c) also was reported to
have virtually undetectable erythrocyte GALT activity, 113 thereby challenging the "reversion" concept.
R148W, a mutation usually reported as the only mutation on galactosemia alleles, was found once on a
D2 allele in an affected individual living in Australia. The fact that both R148W and R204X occur at CpG
hypermutable sites would suggest that the mutations could have arisen independently on a number of
haplotype backgrounds. A few other mutations also have been reported on D2 alleles. These include
T23A and Q207X, 500 M279R, and a compound deletion/insertion mutation at codon 14 designated
A14delCinsTT. 456 Except for M279R, found in a homoallelic state in an individual who was the offspring of
a consanguineous marriage, the mutations occurred on single alleles. The situation is clearly complex,
and factors beyond GALT genotype that vary between individuals also may influence phenotypic outcome.
Attempts at genotype-phenotype correlation

Dietary restriction of galactose results in rapid clinical improvement in the neonate. However, the
neurologic and intellectual prognosis of galactosemic patients is often poor, and ovarian failure is virtually
ubiquitous among female patients. Although only two or three GALT mutations may account for 70 to 80
percent of mutant chromosomes in some populations, the extraordinary allelic heterogeneity of classic
galactosemia suggests that some degree of phenotypic severity may reflect GALT genotype. A number of
studies have attempted to address the question of genotype-phenotype correlation in classic
galactosemia; the results of these studies have been mixed and in some cases fully contradictory
(discussed below), perhaps reflecting the relatively small sample sizes of the cohorts studied and in part
likely reflecting the fact that many factors beyond GALT genotype influence patient outcome.
The most common mutation, Q188R, generally is associated with a severe biochemical phenotype. One
study suggested that homozygosity for Q188R alone was a significant determinant of poor clinical
outcome, 110 whereas another study reported no statistically significant difference in long-term outcome for
galactosemic individuals who were Q188R homoallelic, Q188R heteroallelic, or Q188R negative. 222
Although homozygosity for Q188R is the most common mutant genotype in most European populations or
populations of predominantly European descent, it is now clear that many other mutations are also
associated with long-term clinical complications. Twenty-two percent of base changes at the GALT gene
are nonsense mutations, splice-site mutations, and frameshift deletions or insertions. 454 Although they
are individually rare, when one of these occurs in a homoallelic or heteroallelic state with Q188R, a more
severe clinical phenotype would be anticipated and indeed has been reported. 133, 347 Another missense
mutation, K285N, that is relatively common in some European populations is invariably associated with a
total loss in erythrocyte GALT activity, 347, 418 and these patients also demonstrate a severe
phenotype. 347, 453
The phenotype associated with a Leu-to-Pro mutation at codon 195, designated L195P, may be variable.
Although the results of expression analysis have shown an almost complete loss of enzyme activity, 359
patients homozygous for L195P are reported to manifest a milder clinical course, 409 whereas others who
are heteroallelic for Q188R and L195P have a more severe phenotype. 453 In some populations, these
mutations account for 5 to 30 percent of mutant alleles. 163, 242, 277, 348, 453 Thus, assuming
Hardy-Weinberg equilibrium, homozygosity for these and heterozygosity for Q188R would be expected to
account for 30 to 40 percent of affected individuals.
Other mutations have been found associated with a milder biochemical phenotype 347, 405, 406, 409, 418 and
clinical outcome, such as the S135L mutation described earlier. Other examples include individuals
homoallelic for S329F or IVS8nt+13, in whom erythrocyte GALT activity between 1.5 and 2 percent of
normal is found. Trace residual GALT activity is also detectable in patients heteroallelic for T138M,
T350A, R259W, or A330V and a null mutation. 409 Plasma Gal-1-P concentrations are reported to
decrease much more quickly following dietary restriction of galactose in these patients than in
galactosemic patients with two null alleles, and late-onset complications have not been observed even
among adult patients carrying these mutations. 409 Other mutations are associated with erythrocyte GALT
activity of between 15 and 25 percent of normal when they occur in combination with a null allele, and
these include IVS3nt+29g to c (Schonstadt), IVS8nt+58g to t (Munich 2), R333G (Marburg), 409 and
R333Q. 409 After the neonatal period, individuals carrying the first three of these mutations required no
treatment for normal development. 409 In vitro expression analysis of R333Q has revealed residual activity
similar to that of the D2 allele. 17, 183
Together these reports suggest that GALT genotypes associated with trace-to-significant residual GALT
activity are more likely to result in mild outcomes than are genotypes that leave no residual activity.
However, it is also clear that GALT genotype is but one of many factors influencing long-term outcome.
The most obvious indicator of this reality is the observation that families with more than one affected child
may see disparate outcomes among their own children who nonetheless share identical GALT genotypes.
GALE DEFICIENCY

Partial to complete deficiency of GALE activity (OMIM 230350; reviewed in ref. 448) occurs as an
autosomal recessive trait in the live-born population. Originally identified as a biochemical oddity found in
the circulating red and white blood cells of clinically well individuals, 146 GALE deficiency was long
considered a benign condition. Despite profound impairment in the blood, normal or near-normal GALE
activity was detected in the fibroblasts, liver, phytohemaglutinin (PHA)-stimulated leukocytes, and even
EBV-transformed lymphoblasts from affected individuals. 146, 154, 294 This apparently innocuous form of
GALE deficiency therefore was termed peripheral epimerase deficiency. 59, 155, 201, 293, 331
Individuals with peripheral GALE deficiency who have been studied have shown apparently normal growth
and development despite raised red blood cell Gal-1-P levels, and there has been no evidence of renal
disease or cataracts. Indeed, most patients with peripheral GALE deficiency likely go undiagnosed
because only newborn screening protocols that test total blood galactose inparallel with GALT enzymatic
level, rather than in follow-up to an abnormal GALT result, are likely to detect GALE deficiency. The
frequency of GALE deficiency therefore is difficult to gauge with certainty. Nonetheless, one estimate
derived from newborn screening data predicted that mild to intermediate GALE deficiency may affect as
many as 1 in 6700 African-American newborns and approximately 1 in 70,000 Caucasians. 8
Heterogeneity of GALE deficiency

One of the perplexing realities of GALE deficiency, as it is currently understood, is its heterogeneity. There
is allelic heterogeneity, variability in the degree of enzyme defect, variability in which substrate pair is most
affected by the enzyme defect, and variability in which tissues are most affected by the defect. That there
ashould lso be extraordinary variability in clinical outcome is therefore not surprising. While future studies
will be required to explain the mechanistic bases of the observed variability, several relationships are
already clear. First, the degree of GALE defect in the red blood cells does not correlate with the degree of
GALE defect seen in other tissues. For example, both clinically benign 154 and clinically severe 193, 378
GALE deficiencies may be identified by a complete absence of enzyme activity in red blood cells. On the
other hand, phytohemagglutinin treatment of leukocytes from a patient with the benign form of the
condition resulted in the production of near-normal epimerase activity, and a long-term lymphoblast cell
line had control levels of the enzyme, whereas comparable cells from a severely affected patient did
not. 294 Recent studies reported by Openo and colleagues 329 further demonstrate the lack of correlation
between red blood cell GALE and enzyme levels detected in other cells (lymphoblasts).
In contrast with peripheral GALE deficiency, there is also a very rare but clinically severe generalized form
of GALE deficiency that was first described by Holton and colleagues. 193 Patients with generalized
epimerase deficiency demonstrate marked loss of GALE activity in all tissues tested. Generalized GALE
deficiency appears to be an exceptionally rare disorder, with only seven patients from two extended
families described 180, 193, 378, 468 (Walter, personal communication). Both families affected by generalized
epimerase deficiency are of Pakistani origin, and in each case the parents of the affected individuals are
consanguineous.
Clinical presentation of generalized GALE deficiency

The first child described with generalized GALE deficiency was born in 1980 and presented soon after
birth with jaundice, weight loss, hypotonia, and vomiting. She had a generalized aminoaciduria and
galactosuria. Her symptoms improved on a low-lactose formula, and she started to gain weight, but she
had persistent hypotonia, developmental delay, and severe sensorineural deafness. She also had some
facial dysmorphism, with an inner canthal distance below the third centile, palpebral fissure length greater
than the 97th centile, posteriorly rotated ears, and a short philtrum. On follow-up evaluation, this patient
has been noted to have normal pubertal development with no evidence of ovarian failure 468 (Walter,
personal communication).
A second affected child was born to this same family in 1990. This infant was diagnosed prenatally and
started on a low-galactose diet from birth. A galactose challenge at 12 months of age led to evidence of
liver damage, with increased liver transaminases and a raised bilirubin. Unlike her sister, this child is not
significantly dysmorphic and does not have sensorineural deafness, although she does have poor growth
and moderate learning difficulties.
The first affected child born into the second family was born in 1984. She fed poorly, became irritable, and
developed corneal opacities and liver disease within the first week of life. She improved on dietary
restriction of galactose but subsequently showed moderate learning difficulties and severe sensorineural
deafness. She has some mild facial dysmorphism, with micrognathia, ligamentous laxity, and posteriorly
rotated ears; persistent femoral anteversion; internal tibial torsion; and short stature and poor growth. Her
pubertal development and ovarian function have been normal.
A brother, born in 1995, also was found to have epimerase deficiency. This child has more severely
dysmorphic features, with marked micrognathia and fixed flexion deformities of his fingers. Although he
has had recurrent middle ear disease, he does not have sensorineural deafness. He is hypotonic and has
moderate developmental delay. His weight gain and growth have been poor. Epimerase deficiency awas
lso found in a first cousin of these children, born in 1994. She has similar dysmorphic features,
sensorineural deafness, poor growth, and global developmental delay. A further cousin from this family
also had severe congenital deformities but did not have epimerase deficiency.
Two additional siblings with GALE deficiency ahave lso been born into this extended family. The first was
born in 2003. He had the dysmorphic features reminiscent of the other affected children. Despite dietary
treatment from birth and good compliance at the age of 7 weeks, he developed severe liver disease and
died at age 2 months. No other disorder could be identified as the cause of the fatal liver failure in this
child. A sister born in 2005 also has dysmorphic features, but she has not developed liver disease.
In view of the high degree of consanguinity over many generations in the affected families, it is probable
that there may be two or more recessive disorders contributing to the negative outcomes observed. While
the clinical response to dietary restriction of galactose confirms that the acute and potentially lethal
symptoms result from GALE deficiency, it is difficult to determine precisely which long-term complications
are due to epimerase deficiency. All six surviving affected children from both families have learning
difficulties, developmental delay, hypotonia, and poor growth, and four have sensorineural deafness. In
contrast to the clinical picture in GALT deficiency, ovarian failure has not occurred in the two girls who are
old enough to have entered puberty.
Intermediate GALE deficiency

In addition to patients with ostensibly benign peripheral epimerase deficiency and severe generalized
epimerase deficiency, there appear to be patients who fall between these two extremes; these patients
are said to have intermediate epimerase deficiency. Of these patients, some have been diagnosed
clinically, whereas more recently others have been diagnosed biochemically in follow-up to an abnormal
newborn screen. The first patient, reported by Boleda and colleagues, 50 presented with severe but
transient disease that was said to be related to galactose ingestion. This infant presented at the age of 48
h with seizures, vomiting, and hypoglycemia. He improved following withdrawal of lactose from the diet but
became ill again on two occasions following its reintroduction. While on a galactose-free diet, his Gal-1-P
level in blood was elevated at 0.81 mg/dl (controls < 0.6 mg/dl), and his UDP-Gal was slightly increased.
Erythrocyte GALT and GALK activities were normal, but GALE activity was just below the normal range at
9 months of age. The child remained well following reintroduction of a normal diet at the age of 1 year.
Boleda and colleagues 50 postulated that this patient’s transient but severe galactose-dependent disease
was due to low-normal epimerase activity.
A second patient, reported by Quimby and colleagues 354 and Alano and colleagues, 7 was a male of
mixed Pakistani-European ancestry who was diagnosed in the neonatal period following an abnormal
newborn screen but who remained clinically well on a lactose-containing diet throughout the first year of
life. However, developmental delay became apparent during the second year of life, and by the age of 5
years, the patient exhibited global delays in language and cognitive abilities. Biochemical studies revealed
that this patient exhibited only about 15 percent normal GALE activity in transformed lymphoblasts, 354 in
contrast with the normal GALE activity seen in the lymphoblasts of patients with truly peripheral GALE
deficiency.
A number of other patients with apparently intermediate GALE deficiency also have been reported. For
example, each of three groups reported patients with juvenile cataracts who had a considerable (10 to 20
percent of normal values) or moderate (40 to 60 percent of normal values) reduction in red blood cell
GALE activity. 385, 407, 460 In the study reported by Shin and colleagues, 407 the fibroblast activity in some
patients ranged from normal to a significant decrease, but there was no correlation with the activity in red
blood cells. Most significant was the fact that GALE activity in the lens was greatly decreased in the
patients studied.
These results underscore the limitations of red blood cell GALE activity alone as a means of distinguishing
peripheral from generalized or intermediate GALE deficiency. Indeed, a recent study by Openo and
colleagues 329 demonstrated a spectrum of GALE activities ranging from nearly normal to approximately
15 percent of normal in the lymphoblasts of patients identified as infants on the basis of reduced red blood
cell GALE. These authors found no correlation between red blood cell GALE and lymphoblast GALE.
Clear correlations were identified, however, between the degree of lymphoblast GALE impairment and the
degree of metabolic abnormality observed on exposure of the cells to trace quantities of galactose.
Available medical data further demonstrated that some but not all of these patients accumulated abnormal
levels of Gal-1-P when ingesting a normal diet. The potential clinical implications of these observations
remain unclear. Nonetheless, the point that epimerase-deficiency galactosemia is a spectrum rather than
a binary disorder is now firmly established.
Enzyme defect
There have been no reports to date of patients being completely GALE deficient; even the most severely
affected patients demonstrate low but nonzero levels (e.g., approximately 5 to 10 percent) of residual
GALE activity. 468 This observation stands in stark contrast to both GALK deficiency and GALT deficiency,
in which patients are found among the live-born population with no residual activity whatsoever. This
observation is consistent with the prediction of Kalckar, 217 who postulated that complete loss of GALE
activity in a mammal would be devastating and likely incompatible with life.
The findings in all forms of GALE deficiency are consistent with the idea that clinical severity is a function
of the degree of enzyme impairment in tissues other than red blood cells and that the apparent
tissue-specific GALE impairment observed likely reflects a combination of tissue differences in GALE
expression coupled with differences in the stability of the various mutant GALE enzymes. Although
possible differences in expression remain speculative, differences in stability have been implied or
demonstrated in several ways. For example, GALE purified from fibroblast and lymphocyte cell lines of the
benign form was shown to have reduced heat stability and a greater requirement for the NAD+ coenzyme,
although the pH optimum and K m for UDP-Gal were the same as those found in cells with the normal
enzyme. 153 The liver enzyme from a patient with the severe form of the disorder also had a normal pH
optimum and K m , but the enzyme activity was not enhanced by increased NAD+ concentrations. 140
Working with human GALE expressed in yeast, Wohlers and colleagues 489, 490 demonstrated altered
thermal stability and/or increased sensitivity to proteolysis of GALE proteins carrying substitutions found in
the GALE alleles of patients with ostensibly peripheral epimerase deficiency. Timson 445 obtained a similar
result from his work with recombinant human GALE isolated from E. coli. The implication is that long-lived
cells, such as erythrocytes, that no longer synthesize new enzyme would lose their GALE activity through
turnover, whereas cells continuously synthesizing fresh GALE (e.g., lymphoblasts or fibroblasts) would
not.
Metabolic abnormality
Patients with GALE deficiency, whether peripheral, intermediate, or generalized, accumulate abnormally
high levels of erythrocyte Gal-1-P while receiving dietary galactose. 148, 295, 378 Erythrocyte UDP-Gal levels
also accumulate to abnormally high levels in these patients. 180, 295 Erythrocyte UDP-Gal levels also
accumulate to abnormally high levels in these patients, as do plasma levels of galactitol. 180, 208, 295 In onw
patient with generalized GALE deficiency, UDP-Gal levels rose rapidly to a plateau level of around 180
µmol/liter on galactose intakes as little as 1 g/day (Fig. 72-9). This value compares well with the
erythrocyte UDP-Gal levels detected in four Japanese neonates with ostensibly peripheral GALE
deficiency; the levels detected in those patients were 7.23 ± 3.14 mg/dl of packed red blood cells, 295
which corresponds to about 278 µmol/liter. For comparison, the erythrocyte UDP-Gal levels of 18 normal
neonates in the same study averaged 1.00 ± 0.45 mg/dl of packed red blood cells. 295 Henderson and
colleagues 180 suggested that the level of UDP-Gal in red blood cells was limited in this patient by the
availability of UDP-Glc from the pyrophosphorylase pathway, which would be required for operation of the
transferase reaction. These authors postulated that once the necessary supply of UDP-Glc became
exhausted, GALT would cease to function effectively, and Gal-1-P would begin to accumulate. 180
Consistent with this hypothesis, Mizoguchi and colleagues 295 found that the erythrocyte UDP-Glc levels in
GALE-deficient patients were only about 20 percent of the normal level.
Accumulation of galactose metabolites in a child with the severe form of epimerase deficiency given
various galactose supplements. Open circles = levels of Gal-1-P in packed red cells; − = mean levels of
Gal-1-P in blood samples collected when the galactose intake was the same; open squares = levels of
UDP-Gal in packed red cells; … = levels of UDP-Gal in blood samples collected when the galactose
intake was the same.
Recent data from yeast, 369 Chinese hamster ovary cells, 386 and patient lymphoblasts 329 further confirm
that cells with even partial impairment of GALE accumulate elevated levels of both UDP-Gal and Gal-1-P.
That Gal-1-P accumulates in GALE-impaired cells despite the presence of normal GALT enzyme implies,
as postulated by Henderson and colleagues, 180 that blocking the Leloir pathway at the step of GALE
"backs up" the system, presumably through inhibition of GALT function. As predicted, in GALE-deficient
cells exposed to galactose, UDP-Glc indeed became deficient. 386 A similar observation was reported from
studies of patient blood. 329 Of note, the basis of UDP-Glc depletion in GALE-null cells may involve
depletion of the uridine pool. Consistent with this hypothesis, Schulz and colleagues 386 observed that
uridine supplementation of GALE-deficient (ldlD) Chinese hamster ovary cells restored their UDP-Glc
levels and also complemented their growth impairment in the presence of galactose. The potential clinical
implications of this result are discussed below.
Prenatal diagnosis
One early paper described the prenatal diagnosis of GALE deficiency by assay of the enzyme in
amniocytes. 141 A prenatal diagnosis also was obtained in the second pregnancy of a woman whose first
child suffered from the severe form of GALE deficiency. The result suggested that the fetus was
heterozygous for the deficiency, and postnatal erythrocyte epimerase levels at 14 days of age supported
this conclusion. The baby suffered none of the neonatal problems associated with epimerase deficiency
and has developed normally. Given awareness of the familial mutation (V94M 490 ) associated with
generalized GALE deficiency in the two known affected families, genetic testing is also now an option for
prenatal diagnosis.
Treatment and outcome

No treatment appears to be required for peripheral epimerase deficiency, although ascertainment bias and
lack of long-term follow-up for most patients mean that this conclusion must remain speculative.
Treatment for children with generalized epimerase deficiency, as for those with GALT deficiency, consists
of restriction of dietary galactose. If epimerase deficiency is absolute, then there should, in theory, be no
production of galactoproteins and galactolipids in the absence of exogenous sources of galactose. It has
been suggested, therefore, that a small amount of exogenous galactose be prescribed in the diet to allow
for the manufacture of these essential compounds. However, to date, there have been no reports of
humans who are truly devoid of GALE activity. The most severely affected patients known are all
homozygous for the V94M substitution, which is associated with 5 percent residual activity with regard to
UDP-Gal and 25 percent residual activity with regard to UDP-GalNAc. 489, 490 Indeed, production of
low-density lipoproteins, which are themselves galactoproteins, was normal in fibroblasts from one of
these patients. 230
One of the most important unanswered questions with regard to treatment of GALE deficiency stems from
the fact that a significant fraction of patients with red blood cell GALE deficiency who are clinically well as
infants nonetheless demonstrate intermediate GALE deficiency in other tissues and accumulate abnormal
levels of Gal-1-P in their blood when ingesting a normal diet. The questions are whether these patients
should be treated, and if so, how and for how long? The answers are not yet known.
A final point with regard to treatment of GALE-deficiency galactosemia revisits the question of uridine
supplementation. While uridine supplementation was tried and found to be ineffective for patients with
classic transferase-deficiency galactosemia, 282 recent data as described earlier, 386 suggest that uridine
supplementation might offer some benefit to patients with GALE deficiency.
Pathophysiology
The basis of the pathophysiology in generalized GALE-deficiency galactosemia remains unknown; further,
it remains unclear how much overlap there may be between pathophysiology in GALE deficiency and in
GALT deficiency. In large part the uncertainty derives from two realizations. First, profound impairment of
GALT compromises the ability to metabolize galactose, but profound impairment of GALE compromises
both the ability to metabolize galactose and the ability to synthesize endogenous galactose. This
difference may be especially important during early development in utero because exogenous sources of
galactose would be limited at best. The second difference between GALT and GALE deficiency that may
affect pathophysiology derives from the fact that patients with classic galactosemia accumulate
predominantly Gal-1-P and galactitol in cells and tissues, whereas patients with generalized GALE
deficiency also accumulate dramatically elevated levels of UDP-Gal. If UDP-Gal levels are abnormal at all
in patients with classic galactosemia, which is a point of dispute, they are low, not elevated.
Finally, recent studies from yeast provide compelling evidence that the galactose sensitivities of GALT-null
versus GALE-null cells are distinct. For example, Ross and colleagues 369 demonstrated that GALE-null
yeast arrest growth on exposure to levels of galactose that are 10-fold lower than the levels needed to
growth-arrest GALT-null yeast. Wasilenko and colleagues 472 created and applied a
doxycycline-repressible allele of GALE to test the quantitative relationship between GALE activity and
galactose sensitivity in yeast. These authors observed a smooth linear relationship between galactose
metabolism and GALE activity over a range from 0 to approximately 5 percent but a steep threshold
relationship between growth rate and GALE activity over the same range. The relationship between
abnormal accumulation of metabolites and GALE activity also was linear through this range, suggesting
that if the abnormal accumulation of metabolites underlies galactose-dependent growth arrest in
GALE-impaired yeast, either the impact of two or more metabolites must be synergistic and/or the
threshold of sensitivity to a single metabolite must be very steep. Consistent with this conclusion, Mumma
and colleagues 303 demonstrated from their studies of yeast expressing a doxycycline-repressible allele of
GALK that accumulation of Gal-1-P alone could account for the galactose sensitivity of GALT-null yeast,
but Gal-1-P alone could not account for the galactose sensitivity of GALE-null yeast. Together these
results suggest that the pathophysiology of generalized epimerase deficiency and classic galactosemia
may overlap, but they are not fully identical.
Mutations at the GALE locus

As of December 2007, at least 20 mutations at the human GALEgene locus of patients with GALE
impairment had been reported. As with GALT, most GALE mutations appear to be individually rare, having
been reported only on single alleles or in a homoallelic state in one individual. 275, 276, 329, 339 Exceptions
include K257R and G319E, each of which has been identified in multiple patients, 473 and V94M, identified
by Wohlers and colleagues 489, 490 in the Pakistani patient first reported by Holton and colleagues 193 and
also found in the other affected family 468 that was not known to be related to the first. Although a
polymerase chain reaction (PCR) method has been developed for detecting nine known mutations in
GALE, 178 a large number of mutations have yet to be characterized. For example, the five mutations
characterized by Maceratesi and colleagues in 35 patients with a biochemical diagnosis of epimerase
enzyme deficiency accounted for fewer than 13 percent of mutant chromosomes in that cohort. 275, 276
Structural abnormalities, such as large deletions or rearrangements in the GALE gene, have not been
detected as a significant cause of GALE deficiency, 98 although a growing number of noncoding mutations
have added to the list of missense mutations. 329 A total of 16 missense mutations and 1 nonsense
mutation in human GALE are currently listed in the Human Gene Mutation Database
(www.hgmd.cf.ac.uk/ac/gene.php?gene=GALE).
Functional Impact of GALE Mutations

As with GALT, a number of different approaches have been applied to ascertain the functional
significance of patient mutations in GALE. Those methods reported include studies of patient cells or
samples, 329, 468 studies of patient alleles expressed in a null-background strain of yeast, 354, 473 studies of
recombinant human GALE purified from a bacterial host 73, 445 and in silico predictions of mutation
impact 276 based on degree of sequence conservation and other factors. As with GALT mutations, these
methodologies demonstrate that while some GALE mutations result in profound loss of function, others do
not. Notably, while some noncoding sequence changes have been identified in patient alleles, contrary to
early prediction, 276 point missense rather than noncoding regulatory mutations appear to account for
many cases of both peripheral and generalized GALE deficiency. 490 The difference in severity appears to
depend on the combined degree of functional impairment of the alleles in question and perhaps also on
whether that impairment results, at least in part, from protein instability.
A null-background yeast expression system for the human GALE (hGALE) enzyme was developed by
Quimby and colleagues to enable structural and functional studies of wild-type and mutant alleles. 354 In
sequential series of experiments, Quimby and colleagues 354 and Wohlers and colleagues 490 investigated
six mutations using this system, and although the numbers were small, their results suggested that some
naturally occurring GALE substitution mutations result in severe catalytic impairment, whereas others
have little effect on enzyme function. Evidence of impaired stability also was detected in some of the
mutant GALE proteins. 489, 490
With respect to UDP-Gal as a substrate, the mutations associated with an almost complete loss of
catalytic activity were V94M, G90E, and L183P, the first being the mutation found in a homoallelic state in
the Pakistani children with generalized epimerase deficiency. D103G, L313M, and N34S, on the other
hand, demonstrated near-normal activity under standard assay conditions, although the N34S hGALE
protein was significantly impaired relative to the wild-type protein under conditions of limiting NAD+.
Coexpression studies suggest that allelic interaction, either as partial dominant negative or partial
dominant positive, may be allele-specific. 354, 490 More recently, Timson 445 has applied an E. coli
expression system to study the catalytic impact of an expanded list of mutations in human GALE.
Substrate Specificity
As mentioned earlier, hGALE catalyzes both the interconversion of UDP-Gal/UDP-Glc and
UDP-GalNAc/UDP-GlcNAc. Wohlers and colleagues 490 also tested UDP-GalNAc as a substrate in
studies of a small set of patient mutations and found that some patient mutations (G90E and D103G)
disrupt GALE activity equally with respect to both substrates; others do not. V94M, in particular, exhibited
considerably higher activity with respect to UDP-GalNAc (24 percent) than to UDP-Gal (5 percent).
UDP-GalNAc and UDP-GlcNAc play key roles in the assembly of complex polysaccharides and other
glycosylated macromolecules, especially in the brain, and Wohlers and colleagues 490 have speculated
that the impact of patient mutations on this reaction may be as relevant as the interruption of the Leloir
pathway in the pathogenesis of epimerase-deficiency galactosemia. Recently, Schulz and colleagues, 387
Openo and colleagues, 329 and Waslienko and Fridovich-Keil 473 have further explored the differential
impact of mutations on the two GALE substrate pairs.
Detailed kinetic studies by Wohlers and Fridovich-Keil 489 on the impact of the V94M mutation on enzyme
activity demonstrated that the mutant protein is impaired at the level of V max rather than K m . The x-ray
crystallographic studies of Thoden and colleagues 444 gave added insight into the effects of the mutation
on the 3D structure of the enzyme. These structural studies showed that in the wild-type enzyme, the
hydrophobic side chain of Val94 packs near the aromatic group of the catalytic Tyr157. This acts as a
molecular "fence," or boundary, to limit the rotation of the glycosyl portions of the UDP-sugar substrates
within the active site. The net effect of the valine-to-methionine substitution is to open up the
Ala93-to-Glu96 surface loop, thereby allowing free rotation of the sugar moiety in the active site and/or
nonproductive substrate binding. The fact that the overall effect is to limit the time the ligand is bound in a
productive mode near the Ser132 and Tyr157 residues substantiates the major impact of the substitution
on V max rather than on K m .
ACKNOWLEDGMENTS
Drs. Fridovich-Keil and Walter acknowledge most sincerely the efforts and contributions made by Drs.
Linda Tyfield and John Holton in previous versions of this chapter and by Drs. Stanton Segal and Gerard
Berry to earlier versions. We are also indebted to our many colleagues throughout the world who have
assisted us in preparing this updated version, in particular Dr. Hazel Holden, who contributed the
structural images of GALK, GALT, and GALE, and Ms. Cheryl Strauss for her careful editorial assistance.
REFERENCES
1. Abraham H, Howell R: Human hepatic uridine diphosphate galactose pyrophosphorylase Its

characterization and activity during development. J Biol Chem. 244:545, 1969. PMID 5768855
2. Acosta P, Gross K: Hidden sources of galactose in the environment. Eur J Pediatr. 154:S87, 1995.
3. Adair WJ, et al: 4-Uloses as intermediates in enzyme-nicotinamide adenine dinucleotide-mediated

oxidoreductase mechanisms. I. Uridine diphosphate-galactose-4-epimerase. J Biol Chem. 248:4635,
1973. PMID 7671974
4. Adams C, et al: Gene organization and structure of the Streptomyces lividans gal operon. J
Bacteriol. 170:203, 1988. PMID 3335481
5. Ai Y, et al: Isolation, characterization, and location on chromosome 11. Genome Res. 5:53, 1995.
PMID 8717055
6. Ai Y, et al: A mouse model of galactose-induced cataracts. Hum Mol Genet. 9:1821, 2000. PMID
10915771
7. Alano A, et al: Molecular characterization of a unique patient with epimerase-deficiency

galactosaemia. J Inherit Metab Dis. 21:341, 1998. PMID 9700591
8. Alano A, et al: Evidence for multiple alleles. J Invest Med. 45:191A, 1997.
9. Allen J, et al: Evidence for galactosaemia in utero. Lancet. 1:603, 1980. PMID 6102329
10. Allen J, Holton J, Gillett M: Gas-liquid chromatographic determination of galactitol in amniotic fluid
for possible use in prenatal diagnosis of galactosaemia. Clin Chim Acta. 110:59, 1981. PMID 7214715
11. Amado M, et al: Galactosyltransferases for all functions. Biochim Biophys Acta. 1473:35, 1999.
PMID 10580128
12. Anand V, Sindhi J: Kinase defect in galactosaemia A case report. Indian Pediatr. 16:379, 1979.
PMID 511315
13. Andersen MW, et al: Immunochemical studies of the Duarte and Los Angeles variants. Hum
Genet. 65:287, 1984. PMID 6321325
14. Antshel K, Epstein I, Waisbren S: Cognitive strengths and weaknesses in children and adolescents
homozygous for the galactosemia Q188R mutation A descriptive study Neuropsychology. 18:658,
2004. PMID 15506833
15. Aoki K, Wada Y: Outcome of the patients detected by newborn screening in Japan Acta Paediatr
Jpn. 30:429, 1988. PMID 3150232
16. Asada M, et al: Molecular characterization of galactokinase deficiency in Japanese patients. J

Hum Genet. 44:377, 1999. PMID 10570908
17. Ashino J, et al: Molecular characterization of galactosemia (type 1) mutations in Japanese. Hum
Mutat. 6:36, 1995. PMID 7550229
18. Atkinson M, Burton R, Morton R: Equilibrium constant of phosphoryl transfer from adenosine
triphosphate to galactose in the presence of galactokinase. Biochem J. 78:813, 1961. PMID 13684980
19. Aughton D: Müllerian duct abnormalities and galactosaemia heterozygosity Report of a family. Clin
Dysmorphol. 2:55, 1993. PMID 8298739
20. Badawi N, et al: A controversial disorder. Screening and outcome. Ireland 1972–1992. Ir Med J.
89:16, 1996. PMID 8984074
21. Baker L, et al: Symptomatic and asymptomatic homozygotes in one Negro sibship. J Pediatr.
68:551, 1966.
22. Barbouth D, et al: Prevention of a molecular misdiagnosis in galactosemia. Genet Med. 8:178,
2006. PMID 16540753
23. Barbouth D, et al: Screening newborns for galactosemia using total body galactose oxidation to
CO 2 in expired air. Pediatr Res. 62:720, 2007. PMID 17957157
24. Benn P, et al: Confirmation of the assignment of the gene for galactose-1-phosphate
uridylyltransferase (E.C. 2.7.7.12) to human chromosome 9. Cytogenet Cell Genet. 24:37, 1979. PMID
456040
25. Benn P, et al: Assignment of a gene for uridine diphosphate galactose-4-epimerase to human
chromosome 1 by somatic cell hybridization, with evidence for a regional assignment to 1pter yields
1p21. Cytogenet Cell Genet. 24:138, 1979. PMID 477411
26. Bergren W, Donnell G: A new variant of galactose-1-phosphate uridyltransferase in man The Los
Angeles variant. Ann Hum Genet. 37:1, 1973. PMID 4759900
27. Bergren W, et al: Identification in the urine. Science. 176:683, 1972.
28. Bergren WR, Ng WG, Donnell GN: Uridine diphosphate galastose-4-epimerase in human and
other mammalian hemolysates. Biochim Biophys Acta. 315:464, 1973.
29. Bergsma D, et al: Fine structure of the human galactokinase GALK1 gene. Genome Res. 6:980,
1996. PMID 8908517
30. Berry G, et al: In vivo evidence of brain galactitol accumulation in an infant with galactosemia and
encephalopathy. J Pediatr. 138:260, 2001. PMID 11174626
31. Berry G, et al: The rate of de novo galactose synthesis in patients with galactose-1-phosphate
uridyltransferase deficiency. Mol Genet Metab. 81:22, 2004. PMID 14728988
32. Berry G, et al: The effect of dietary fruits and vegetables on urinary galactitol excretion in
galactose-1-phosphate uridyltransferase deficiency. J Inherit Metab Dis. 16:91, 1993. PMID 8487507
33. Berry G, et al: Red blood cell uridine sugar nucleotide levels in patients with classic galactosemia
and other metabolic disorders. Metabolism. 41:783, 1992. PMID 1619998
34. Berry G, et al: Galactose breath testing distinguishes variant and severe galactose-1-phosphate
uridyltransferase genotypes. Pediatr Res. 48:323, 2000. PMID 10960497
35. Berry GT, et al: Evidence for alternate galactose oxidation in a patient with deletion of the
galactose-1-phosphate uridyltransferase gene. Mol Genet Metab. 72:316, 2001. PMID 11286505
36. Berry GT, et al: Quantitative assessment of whole body galactose metabolism in galactosemic
patients. Eur J Pediatr. 156:S43, 1997. PMID 9266215
37. Berry GT, et al: Endogenous synthesis of galactose in normal men and patients with hereditary
galactosemia. Lancet. 346:1073, 1995. PMID 7564790
38. Berry GT, et al: In vivo oxidation of [ 13 C]galactose in patients with galactose-1-phosphate
uridyltransferase deficiency. Biochem Mol Med. 56:158, 1995. PMID 8825079
39. Berry GT, et al: Extended [ 13 C]galactose oxidation studies in patients with galactosemia. Mol
Genet Metab. 82:130, 2004. PMID 15172000
40. Bertoli D, Segal S: Developmental aspects and some characteristics of mammalian

galactose-1-phosphate uridyl transferase. J Biol Chem. 241:4023, 1966. PMID 5920808
41. Beutler E: Screening for galactosemia Studies of the gene frequencies for galactosemia and the
Duarte variant. Isr J Med Sci. 9:1323, 1973. PMID 4149512
42. Beutler E: Galactosemia Screening and diagnosis. Clin Biochem. 24:293, 1991. PMID 1959220
43. Beutler E, Baluda M: A simple spot screening test for galactosemia. J Lab Clin Med. 68:137, 1966.
PMID 4380286
44. Beutler E, et al: A new genetic abnormality resulting in galactose-1-phosphate uridyltransferase

deficiency. Lancet. 1:353, 1965. PMID 14246708
45. Beutler E, Gelbart T: Estimating the prevalence of pyruvate kinase deficiency from the gene
frequency in the general white population. Blood. 95:3585, 2000. PMID 10828047
46. Bhat PJ: Galactose-1-phosphate is a regulator of inositol monophosphatase A fact or a fiction?

Med Hypotheses. 60:123, 2003. PMID 12450779
47. Blume K, Beutler E: Purification and properties of galactokinase from human red blood cells. J Biol
Chem. 246:6507, 1971. PMID 4943674
48. Bohles H, Wenzel D, Shin Y: Progressive cerebellar and extrapyramidal motor disturbances in
galactosaemic twins. Eur J Pediatr. 145:413, 1986. PMID 3792389
49. Bohu P, et al: Study of 3 cases]. Rev Neurol (Paris). 151:136, 1995.
50. Boleda M, et al: Severe neonatal galactose-dependent disease with low-normal epimerase activity.
J Inherit Metab Dis. 18:88, 1995. PMID 7623454
51. Bonham J, et al: A possible indicator of control. Society for the Study of Inborn Errors of
Metabolism Annual Symposium,[[AU. : Location?]] 1994, p 136.
52. Boonyawat B, Kamolsilp M, Phavichitr N: Galactosemia in Thai patient at Phramongkutklao

Hospital A case report. J Med Assoc Thai. 88:S275, 2005. PMID 16858969
53. Borzy M, et al: Recurrent sepsis with deficiencies of C 2 and galactokinase. Am J Dis Child.
138:186, 1984. PMID 6695877
54. Bosch A: Classical galactosaemia revisited. J Inherit Metab Dis. 29:516, 2006. PMID 16838075
55. Bosch A, et al: A review of the literature. J Inherit Metab Dis. 25:629, 2002. PMID 12705493
56. Bosch A, et al: Identification of novel mutations in classical galactosemia. Hum Mutat. 25:502,
2005. PMID 15841485
57. Bosch AM, Van Gennip AH, van Kempen JV, Wanders RJ, Wijburg FA: Clinical features of
galactokinase deficiency A review of the literature. J Inherit Metab Dis. 25:629, 2002. PMID 12705493
58. Bosio A, et al: Composition and biophysical properties of myelin lipid define the neurological
defects in galactocerebroside- and sulfatide-deficient mice. J Neurochem. 70:308, 1998. PMID
9422376
59. Bowling F, et al: A case of uridine diphosphate galactose-4-epimerase deficiency detected by

neonatal screening for galactosaemia. Med J Aust. 144:150, 1986. PMID 3945203
60. Briones P, Giros M, Martinez V: Second spontaneous pregnancy in a galactosaemic woman

homozygous for the Q188R mutation. J Inherit Metab Dis. 24:79, 2001. PMID 11286387
61. Brivet M, et al: 343, 1989. [PMID. : 2512439]
62. Bunn H, Higgins P: Reaction of monosaccharides with proteins Possible evolutionary significance.
Science. 213:222, 1981. PMID 12192669
63. Calderon F, et al: Mutation database for the galactose-1-phosphate uridyltransferase (GALT)
gene. Hum Mutat. 28:939, 2007. PMID 17486650
64. Canals R, et al: The UDP N-acetylgalactosamine 4-epimerase gene is essential for mesophilic
Aeromonas hydrophila serotype O34 virulence. Infect Immun. 74:537, 2006. PMID 16369010
65. Caputto R, et al: Isolation of the coenzyme of the galactose phosphate–glucose phosphate
transformation. J Biol Chem. 184:333, 1950. PMID 15422002
66. Chacko C, Christian J, Nadler H: Unstable galactose-1-phosphate uridyl transferase A new variant
of galactosemia. J Pediatr. 78:454, 1971. PMID 5544155
67. Chacko C, et al: The Chicago variant of clinical galactosemia. Hum Genet. 37:261, 1977. PMID
885545
68. Chakraborty R, et al: New data. Am J Hum Genet. 50:145, 1992. PMID 1729885
69. Charlwood J, et al: Defective galactosylation of serum transferrin in galactosemia. Glycobiology.

8:351, 1998.
70. Chen J, et al: Erythrocyte galactose-1-phosphate quantified by isotope-dilution gas

chromatography–mass spectrometry. Clin Chem. 48:604, 2002. PMID 11901058
71. Chen Y, et al: Reduction in oocyte number following prenatal exposure to a diet high in galactose.
Science. 214:1145, 1981.
72. Cheung K, et al: A case report and review of disease incidence. J Paediatr Child Health. 35:399,
1999. PMID 10457302
73. Chhay J: , et al: A yeast model reflects biochemical severity among members of a family
segregating three variant alleles of galactose-1-P-uridylyltransferase. J Inherit Metab Dis (in press).
PMID 10457302
74. Choulot J, et al: Default of myelisation caused by deficient synthesis of UDP-galactose?]. Arch Fr
Pediatr. 48:267, 1991. PMID 2069479
75. Christacos N, Fridovich-Keil J: Impact of patient mutations on heterodimer formation and function
in human galactose-1-P-uridylyltransferase. Mol Genet Metab. 76:319, 2002. PMID 12208137
76. Christacos NC, et al: Subcellular localization of the enzyme galactose-1-phosphate

uridylyltransferase in the yeast Saccharomyces cerevisiae. Mol Genet Metab. 70:272, 2000. PMID
10993714
77. Citron B, Donelson J: Sequence of the Saccharomyces GAL region and its transcription in vivo. J
Bacteriol. 158:269, 1984. PMID 6715281
78. Cleary M, et al: Relationship of IQ to biochemical control and genotype. J Inherit Metab Dis.
18:151, 1995. PMID 7564232
79. Coffee B, et al: Characterization of an unusual deletion of the galactose-1-phosphate uridyl

transferase (GALT) gene. Genet Med. 8:635, 2006. PMID 1707988
80. Cohn R, Segal S: Some characteristics and developmental aspects of rat uridine
diphosphogalactose-4-epimerase. Biochem Biophys Acta. 171:333, 1969. PMID 4304332
81. Colin J, et al: [Cataract due to galactokinase deficiency in a premature infant]. Arch Fr Pediatr.
33:77, 1976. PMID 185981
82. Cook J, Don N, Mann T: Hereditary galactokinase deficiency. Arch Dis Child. 46:465, 1971. PMID
5109408
83. Coskun T, et al: Mutational analysis of Turkish galactosaemia patients. J Inherit Metab Dis.
18:368, 1995. PMID 7474913
84. Cozen W, et al: Galactose-1-phosphate uridyl transferase (GALT) genotype and phenotype,
galactose consumption, and the risk of borderline and invasive ovarian cancer (United States). Cancer
Causes Control. 13:113, 2002. PMID 11936817
85. Cramer D, et al: Vaginal agenesis (Mayer-Rokitansky-Kuster-Hauser syndrome) associated with

the N314D mutation of galactose-1-phosphate uridyl transferase (GALT). Mol Hum Reprod. 2:145,
1996. PMID 9238673
86. Cramer D, et al: A case-control study of galactose consumption and metabolism in relation to
ovarian cancer. Cancer Epidemiol Biomark Prev. 9:95, 2000.
87. Cramer D, et al: Endometriosis associated with the N314D mutation of galactose-1-phosphate
uridyl transferase (GALT). Mol Hum Reprod. 2:149, 1996. PMID 9238674
88. Cramer DW, et al: Determinants of basal follicle-stimulating hormone levels in premenopausal
women. J Clin Endocrinol Metab. 79:1105, 1994. PMID 7962282
89. Crews C, et al: Functional consequence of substitutions at residue 171 in human galactose-1-P
uridylyltransferase. J Biol Chem. 275:22847, 2000. PMID 10811638
90. Cuatrecasas P, Segal S: Mammalian galactokinase Developmental and adaptive characteristics in

the rat liver. J Biol Chem. 240:2382, 1965. PMID 14304841
91. Cuatrecasas P, Segal S: Mammalian galactose dehydrogenase I. Identification and purification in

rat liver. J Biol Chem. 241:5904, 1966. PMID 5954367
92. Cuatrecasas P, Segal S: Mammalian galactose dehydrogenase II. Properties, substrate specificity,
and developmental changes. J Biol Chem. 241:5910, 1966. PMID 4380933
93. Dagna Bricarelli F, et al: Evidence for assignment of the GALT locus to the 9p21 band. Hum
Genet. 59:112, 1981. PMID 6276288
94. Dahlqvist A, Gamstorp I, Madsen H: A patient with hereditary galactokinase deficiency. Acta
Paediatr Scand. 59:669, 1970. PMID 5482251
95. Dale G, Popjak G: Purification of normal and inactive galactosemic galactose-1-phosphate

uridyltransferase from human red cells. J Biol Chem. 251:1057, 1976. PMID 814122
96. Darrow RA, Rodstrom R: Purification and properties of uridine diphosphate galactose 4-epimerase
from yeast. Biochemistry. 7:1645, 1968. PMID 5650372
97. Daude N, et al: In vivo and in vitro expression of rat galactose-1-phosphate uridyltransferase
(GALT) in the developing central and peripheral nervous system. Brain Res Dev Brain Res. 94:190,
1996. PMID 8836578
98. Daude N, et al: Molecular cloning, chracterization, and mapping of a full-length cDNA encoding
human UDP-galactose 4′-epimerase. Biochem Mol Med. 56:1, 1995. PMID 8593531
99. Daude N, et al: Expression of galactose-1-phosphate uridyltransferase in the anterior pituitary of

rat during the estrous cycle. Neuroendocrinology. 64:42, 1996.
100. de Jongh S, et al: Spontaneous pregnancy in a patient with classical galactosaemia. J Inherit
Metab Dis. 22:754, 1999. PMID 10472536
101. Debouck C, et al: Structure of the galactokinase gene of Escherichia coli, the last (?) gene of the
gal operon. Nucl Acids Res. 13:1841, 1985. PMID 3158881
102. Dobbie JA, Holton JB, Clamp JR: Defective galactosylation of proteins in cultured skin fibroblasts
from galactosaemic patients. Ann Clin Biochem. 27:274, 1990. PMID 2382964
103. Dobrowolski S, et al: New assays to increase the sensitivity and specificity of newborn screening
for galactosemia. J Mol Diagn. 5:42, 2003. PMID 12552079
104. Donnell G, et al: Observations on results of management of galactosemic patients, in Hsia D (ed).
: Galactosemia. Springfield, IL, Charles C Thomas, 1969, p 247. PMID 12552079
105. Donnell G, et al: Clinical aspects of galactosaemia, in Burman D, Holton J, Pennock C (eds). :
Inherited Disorders of Carbohydrate Metabolism. Lancaster, UK, MTP Press, 1980, p 114. PMID
12552079
106. Donnell G, Collado M, Koch R: Growth and development of children with galactosemia. J Pediatr.
58:836, 1961. PMID 13723756
107. Douglas HC, Hawthorne DC: Enzymatic expression and genetic linkage of genes controlling
galactose utilization in Saccharomyces. Genetics. 49:837, 1964. PMID 14158615
108. Elman M, Miller M, Matalon R: Galactokinase activity in patients with idiopathic cataracts.
Ophthalmology. 93:210, 1986. PMID 3951827
109. Elsas LJ, et al: A common mutation associated with the Duarte galactosemia allele. Am J Hum
Genet. 54:1030, 1994. PMID 8198125
110. Elsas LJ, Fridovich-Keil JL, Leslie ND: Galactosemia A molecular approach to the enigma. Int
Paediatr. 8:101, 1993.
111. Elsas LJ, et al: Functional analysis of the human galactose-1-phosphate uridyltransferase
promoter in Duarte and LA variant galactosemia. Mol Genet Metab. 72:297, 2001. PMID 11286503
112. Elsas LJ, et al: A molecular approach to galactosemia. Eur J Pediatr. 154:S21, 1995. PMID
7671959
113. Elsas LJ, et al: A strategy to identify new biochemical phenotypes and molecular genotypes. Am
J Hum Genet. 56:630, 1995. PMID 7887461
114. Elsevier JP, Fridovich-Keil JL: The Q188R mutation in human galactose-1-phosphate
uridylyltransferase acts as a partial dominant negative. J Biol Chem. 271:32002, 1996. PMID 8943248
115. Elsevier S, et al: Assignment of the gene for galactokinase to human chromosome 17 and its
regional localisation to band q21-22. Nature. 251:633, 1974.
116. Fensom A, Benson P, Blunt S: Prenatal diagnosis of galactosaemia. Br Med J. 4:386, 1974.
PMID 4154122
117. Ficicioglu C: , et al: Effect of galactose free formula on galactose-1-phosphate in two infants with
classical galactosemia. Eur J Pediatr (in press). PMID 4154122
118. Ficicioglu C, Yager C, Segal S: Galactitol and galactonate in red blood cells of children with the
Duarte/galactosemia genotype. Mol Genet Metab. 84:152, 2005. PMID 15670721
119. Field TL, Reznikoff WS, Frey PA: Galactose-1-phosphate uridylyltransferase Identification of
histidine-164 and histidine-166 as critical residues by site-directed mutagenesis. Biochemistry.
28:2094, 1989. PMID 2541773
120. Fishler K, et al: Intellectual and personality development in children with galactosemia. Pediatrics.
50:412, 1972.
121. Fishler K, et al: Psychological correlates in galactosemia. Am J Ment Defic. 71:116, 1966. PMID
5964014
122. Fishler K, et al: Developmental aspects of galactosemia from infancy to childhood. Clin Pediatr
(Phila). 19:38, 1980.
123. Flach JE, Reichardt JKV, Elsas LS: Sequence of a cDNA encoding human
galactose-1-phosphate uridyl transferase. Mol Biol Med. 7:365, 1990. PMID 2233247
124. Forges T, et al: Pathophysiology of impaired ovarian function in galactosaemia. Hum Reprod
Update. 12:573, 2006. PMID 16835432
125. Forster J, et al: A possible role of nucleotide deficiency in galactosemic liver injury. Klin
Wochenschr. 53:1169, 1975. [PMID1214453]
126. Frazier D, Clemons E, Kirkman H: Minimizing false-positive diagnoses in newborn screening for
galactosemia. Biochem Med Metab Biol. 48:199, 1992. PMID 1476788
127. Frey PA: The Leloir pathway A mechanistic imperative for three enzymes to change the
sterochemical configuration of a single carbon in galactose. FASEB J. 10:461, 1996. PMID 8647345
128. Frey PA, et al: Detection, isolation, and characterization of the uridylyl enzyme. Methods
Enzymol. 87:20, 1982. PMID 6294449
129. Fridovich-Keil JL, Jinks-Robertson S: A yeast expression system for human

galactose-1-phosphate uridylyltransferase. Proc Natl Acad Sci USA. 90:398, 1993. PMID 8421669
130. Fridovich-Keil JL, Quimby BB, Wells L, Mazur LA, Elsevier JP: Characterization of the N314D
allele of human galactose-1-phosphate uridylyltransferase using a yeast expression system. Biochem
Mol Med. 56:121, 1995. PMID 8825075
131. Friedman J, Levy H, Boustany R: Late onset of distinct neurologic syndromes in galactosemic
siblings. Neurology. 39:741, 1989. PMID 2710367
132. Fujimoto A, et al: Quantitative Beutler test for newborn mass screening of galactosemia using a
fluorometric microplate reader. Clin Chem. 46:806, 2000. PMID 10839768
133. Gathof B, et al: Characterization of two stop codon mutations in the galactose-1-phosphate
uridyltransferase gene of three male galactosemic patients with severe clinical manifestation. Hum
Genet. 96:721, 1995. PMID 8522334
134. Geeganage S, Frey PA: Transient kinetics of formation and reaction of the uridylyl-enzyme form
of galactose-1-P uridylyltransferase and its Q168R-variant Insight into the molecular basis of
galactosemia. Biochemistry. 37:14500, 1998. PMID 9772178
135. Gibson J: Gonadal function in galactosemics and in galactose-intoxicated animals. Eur J Pediatr.
154:S14, 1995. PMID 7671957
136. Gibson J, et al: Effect of glucose and galactose loading in normal subjects on red and white blood
cell uridine diphosphate sugars. Biochem Mol Med. 55:8, 1995. PMID 7551832
137. Gibson J, et al: Effect of galactose supplementation. Am J Clin Nutr. 63:704, 1996. PMID
8615352
138. Gibson JB, et al: Comparison of erythrocyte uridine sugar nucleotide levels in normals, classic
galactosemics, and patients with other metabolic disorders. Metabolism. 44:597, 1995. PMID 7752907
139. Gibson JB, et al: Uridine diphosphate hexoses in leukocytes and fibroblasts of classic
galactosemics and patients with other metabolic diseases. Pediatr Res. 36:613, 1994. PMID 7877880
140. Gillett M: , et al: Investigation of human uridine 5′-diphosphate galactose 4′-epimerase. Bath, UK,
University of Bath, 1985. PMID 7877880
141. Gillett M, Holton J, MacFaul R: Prenatal determination of uridine diphosphate

galactose-4-epimerase activity. Prenat Diagn. 3:57, 1983. PMID 6844265
142. Gir M, et al: [Mutation P28T in gene GK1 as the cause of a familial galactokinase deficiency].
Arch Soc Esp Oftalmol. 78:111, 2003. PMID 12647253
143. Gitzelmann R: Deficiency of erythrocyte galactokinase in a patient with galactose diabetes.

Lancet. 2:670, 1965. PMID 4158218
144. Gitzelmann R: Hereditary galactokinase deficiency, a newly recognized cause of juvenile

cataracts. Pediatr Res. 1:14, 1967.
145. Gitzelmann R: Formation of galactose-1-phosphate from uridine diphosphate galactose in

erythrocytes from patients with galactosemia. Pediatr Res. 3:279, 1969. PMID 5807057
146. Gitzelmann R: Deficiency of uridine diphosphate galactose 4-epimerase in blood cells of an

apparently healthy infant. Helv Paediatr Acta. 27:125, 1972. PMID 4644860
147. Gitzelmann R: Galactose-1-phosphate in the pathophysiology of galactosemia Eur J Pediatr.

154:S45, 1995. PMID 7671964
148. Gitzelmann R, et al: Galactose metabolism, hereditary defects and their clinical significance, in
Burman D, Holton J, Pennock C (eds). : Inherited Disorders of Carbohydrate Metabolism. Lancaster,
UK, MTP Press, 1980, p 92. PMID 7671964
149. Gitzelmann R, et al: Disorders of galactose metabolism, in Fernandes J, Saudubray J, Tada K

(eds). : Inborn Metabolic Diseases. Berlin, Springer-Verlag, 1990, p 95. PMID 7671964
150. Gitzelmann R, Auricchio S: The handling of soya alpha-galactosides by a normal and a

galactosemic child. Pediatrics. 36:231, 1965.
151. Gitzelmann R, Bosshard NU: Partial deficiency of galactose-1-phosphate uridyltransferase. Eur J

Pediatr. 154:S40, 1995. PMID 7671963
152. Gitzelmann R, Curtius H, Schneller I: Galactitol and galactose-1-phosphate in the lens of a

galactosemic infant. Exp Eye Res. 6:1, 1967. PMID 6019477
153. Gitzelmann R, Haigis E: Appearance of active UDP-galactose 4′-epimerase in cells cultured from
epimerase-deficient persons. J Inherit Metab Dis. 1:41, 1978. PMID 117236
154. Gitzelmann R, Steimann B: Uridine diphosphate galactose 4-epimerase deficiency. Helv Paediatr
Acta. 28:497, 1973. PMID 4785150
155. Gitzelmann R, et al: Uridine diphosphate galactose 4′-epimerase deficiency. Helv Paediatr Acta.
31:441, 1976.
156. Gitzelmann R, Wells H, Segal S: Galactose metabolism in a patient with hereditary galactokinase
deficiency. Eur J Clin Invest. 4:79, 1974. PMID 4365005
157. Golberg L, et al: A possible explanation of the nature of the biochemical lesion. Arch Dis Child.
31:254, 1956. PMID 13363463
158. Goodman M, et al: Association of galactose-1-phosphate uridyltransferase activity and N314D

genotype with the risk of ovarian cancer. Am J Epidemiol. 156:693, 2002. PMID 12370157
159. Gort L, et al: Mutational spectrum of classical galactosaemia in Spain and Portugal. J Inherit
Metab Dis. 29:739, 2006. PMID 17041746
160. Gray G: Carbohydrate digestion and absorption Role of the small intestine. N Engl J Med.
292:1225, 1975. PMID 1093023
161. Gray G, Santiago N: Intestinal βgalactosidases I. Separation and characterization of three

enzymes in normal human intestine. J Clin Invest. 48:716, 1969. PMID 5774109
162. Greber S, et al: Mutations in classical and Duarte2 galactosemia. Pediatr Res. 38:434, 1995.
163. Greber-Platzer S, et al: Mutation analysis by denaturing gradient gel electrophoresis. Hum Mutat.
10:49, 1997. PMID 9222760
164. Greenman L, Rathbun J: Galactose studies in an infant with idiopathic galactose intolerance.
Pediatrics. 2:666, 1948.
165. Gropper S, et al: more foods to consider for the galactose-restricted diet. J Am Diet Assoc.
100:573, 2000. PMID 10812385
166. Gross K, Acosta P: Fruits and vegetables are a source of galactose implications in planning the
diets of patients with galactosaemia. J Inherit Metab Dis. 14:253, 1991. PMID 1886408
167. Guerrero NV, et al: Risk factors for premature ovarian failure in females with galactosemia. J
Pediatr. 137:833, 2000. PMID 11113841
168. Haberland C, et al: A histopathological and biochemical study. J Neuropathol Exp Neurol. 30:431,
1971. PMID 4105426
169. Hadfield R, et al: Absence of a relationship between endometriosis and the N314D polymorphism
of galactose-1-phosphate uridyl transferase in a UK population. Mol Hum Reprod. 5:990, 1999. PMID
10508229
170. Hansen T, et al: Neuropsychological and linguistic follow-up studies of children with
galactosaemia from an unscreened population. Acta Paediatr. 85:1197, 1996. PMID 8922083
171. Hartmann A, Grunwaldt E, James D: Blood galactose in infants and children. J Pediatr. 43:1,
1953. PMID 13062066
172. Hayman S, Kinoshita J: Isolation and properties of lens aldose reductase. J Biol Chem. 240:877,
1965. PMID 14275148
173. He C, et al: No association of endometriosis with galactose-1-phosphate uridyl transferase

mutations in a Chinese population. Environ Mol Mutagen. 47:307, 2006. PMID 16416427
174. Heidenreich R, et al: Developmental and tissue-specific modulation of rat galactose-1-phosphate

uridyltransferase steady state messenger RNA and specific activity levels. Pediatr Res. 34:416, 1993.
PMID 8255669
175. Heidenreich RA: Regulation of galactose-1-phosphate uridyltransferase gene expression Eur J

Pediatr. 154:S28, 1995. PMID 7671960
176. Heinemeyer T, et al: TRANSFAC, TRRD and COMPEL. Nucleic Acids Res. 26:362, 1998. PMID
9399875
177. Henderson H, et al: The clinical and molecular spectrum of galactosemia in patients from the
Cape Town region of South Africa. BMC Pediatr. 2:7, 2002. PMID 12350230
178. Henderson JM: , et al: A PCR-based method for detecting known mutations in the human UDP
galactose-4′-epimerase gene associated with epimerase-deficiency galactosemia. Clin Genet (in
press). PMID 12350230
179. Henderson JM, Wells L, Fridovich-Keil JL: Covalent heterogeneity of the human enzyme
galactose-1-phosphate uridylyltransferase. J Biol Chem. 275:30088, 2000. PMID 10884393
180. Henderson MJ, Holton JB: Further observations in a case of uridine diposphate
galactose-4-epimerase deficiency with a severe clinical presentation. J Inherit Metab Dis. 6:17, 1983.
PMID 6408303
181. Hers H: L’Aldose reductase. Biochim Biophys Acta. 37:120, 1960. PMID 14401390
182. Heyer E: Genetic consequences of differential demographic behaviour in the Saguenay region,
Qu补c Am J Phys Anthropol. 98:1, 1995. PMID 8579187
183. Hirokawa H, et al: Molecular characterization of galactosemia mutations (type 1) in Japanese.

International Congress of Inborn Errors of Metabolism,[[AU. : Location?]] 1997.
184. Hirokawa H, et al: prediction of clinical phenotype from genotype in Japanese patients. Eur J
Hum Genet. 7:757, 1999. PMID 10573007
185. Hittinger CT, Carroll SB: Gene duplication and the adaptive evolution of a classic genetic switch.
Nature. 449:677, 2007. PMID 17928853
186. Hodes B, et al: Macular deposits in galactokinase deficiency. Metab Pediatr Syst Ophthalmol.
8:39, 1985. PMID 3870945
187. Hodge M: Assessment of children with developmental apraxia of speech A rationale. Clin
Commun Disord. 4:91, 1994. PMID 7522738
188. Hoefnagel D, Wurster-Hill D, Child E: Ovarian failure in galactosaemia. Lancet. 2:1197, 1979.
PMID 91932
189. Holden HM, Rayment I, Thoden JB: Structure and function of enzymes of the Leloir pathway for
galactose metabolism. J Biol Chem. 278:43885, 2003. PMID 12923184
190. Holden HM, Timson DJ, Reece RJ: Galactokinase Structure, function and role in type II
galactosemia. Cell Mol Life Sci. 61:2471, 2004. PMID 15526155
191. Holton J, Allen J, Gillett G: Prenatal diagnosis of disorders of galactose metabolism. J Inherit
Metab Dis. 12:202, 1989. PMID 2509807
192. Holton J, Tyfield L: Prenatal diagnosis of inherited metabolic diseases, in Williams D, Marks V
(eds). : Scientific Foundations of Biochemistry in Clinical Practice. Oxford, UK, Butterworth
Heinemann, 1994, p 237. PMID 2509807
193. Holton JB, et al: A new severe variant due to uridine diphosphate galactose-4-epimerase
deficiency. Arch Dis Child. 56:885, 1981. PMID 7305435
194. Honeyman M, et al: Results of the British Paediatric Surveillance Unit Study, 1988Œ1990. Arch
Dis Child. 69:339, 1993. PMID 8215542
195. Horv೨ A, et al: Distribution of Q188R and N314D mutations in the Hungarian galactosemic
population. Hum Mutat. 16:91, 2000. PMID 10874317
196. Hunter M, et al: Novel mutations in the GALK1 gene in patients with galactokinase deficiency.
Hum Mutat. 17:77, 2001. PMID 11139256
197. Hunter M, et al: The P28T mutation in the GALK1 gene accounts for galactokinase deficiency in
Roma (Gypsy) patients across Europe. Pediatr Res. 51:602, 2002. PMID 11978884
198. Hutchesson A, et al: Biological variability in metabolite concentrations. J Inherit Metab Dis.
22:139, 1999. PMID 10234609
199. Huttenlocher P, Hillman R, Hsia Y: Pseudotumor cerebri in galactosemia. J Pediatr. 76:902,

1970. PMID 5444582
200. Ichiba Y, et al: The variant of galactosemia. Jpn J Pediatr. 42:579, 1989.
201. Ichiba Y, Namba N, Misumi H: Uridine diphosphate galactose 4-epimerase deficiency. Am J Dis
Child. 134:995, 1980. PMID 7424864
202. Isselbacher K: A mammalian uridinediphosphate galactose pyrophosphorylase. J Biol Chem.

232:429, 1958. PMID 13549431
203. Isselbacher K, Krane S: Studies on the mechanism of the inhibition of galactose oxidation by
ethanol. J Biol Chem. 236:2394, 1961. PMID 13718082
204. Isselbacher KJ: Evidence for an accessory pathway of galactose metabolism in mammalian liver.
Science. 126:652, 1957.
205. Isselbacher KJ, et al: Congenital galactosemia, a single enzyme block in galactose metabolism.
Science. 123:635, 1956.
206. Item C, et al: Mutations at the galactose-1-P uridyltransferase gene in infants with a positive
galactosemia newborn screening test. Pediatr Res. 51:511, 2002. PMID 11919338
207. Jaeken J, Kint J, Spaapen L: Serum lysosomal enzyme abnormalities in galactosaemia. Lancet.
340:1472, 1992. PMID 1360589
208. Jakobs C, et al: Plasma polyol levels in patients with cataract. J Inherit Metab Dis. 13:517, 1990.
PMID 2122118
209. Jakobs C, et al: Prenatal diagnosis of galactosemia. Eur J Pediatr. 154:S33, 1995. PMID
7671961
210. Jakobs C, et al: Dietary restriction of maternal lactose intake does not prevent accumulation of
galactitol in the amniotic fluid of fetuses affected with galactosaemia. Prenat Diagn. 8:641, 1988. PMID
3211853
211. Jakobs C, Schweitzer S, Dorland B: Galactitol in galactosemia. Eur J Pediatr. 154:S50, 1995.
PMID 7671965
212. Jakobs C, et al: An accurate approach to the prenatal diagnosis of galactosemia. Pediatr Res.
18:714, 1984. PMID 6433315
213. Jan J, Wilson R: Unusual late neurological sequelae in galactosaemia. Dev Med Child Neurol.
15:72, 1973. PMID 4729415
214. Johnston M: A model fungal gene regulatory mechanism The GAL genes of Saccharomyces
cerevisiae. Microbiol Rev. 51:458, 1987. PMID 2830478
215. Kadhom N, et al: Relationship with neurological sequelae. Biochem Med Metab Biol. 52:140,
1994. PMID 7993662
216. Kalaydjieva L, et al: A founder mutation in the GK1 gene is responsible for galactokinase
deficiency in Roma (Gypsies). Am J Hum Genet. 65:1299, 1999. PMID 10521295
217. Kalckar HM: Galactose metabolism and cell šsociology. Š Science. 150:305, 1965.
218. Kalckar HM, Braganca B, Munch-Petersen A: Uridyltransferase and the formation of uridine
diphosphate galactose. Nature. 172:1039, 1953.
219. Kaufman F, et al: Gonadal function in patients with galactosaemia. J Inherit Metab Dis. 9:140,
1986. PMID 3091920
220. Kaufman F, et al: Effect of hypogonadism and deficient calcium intake on bone density in patients
with galactosemia. J Pediatr. 123:365, 1993. PMID 8355111
221. Kaufman F, et al: Cognitive functioning, neurologic status and brain imaging in classical
galactosemia. Eur J Pediatr. 154:S2, 1995. PMID 7671958
222. Kaufman FR, et al: Correlation of cognitive, neurologic, and ovarian outcome with the Q188R
mutation of the galactose-1-phosphate uridyltransferase gene. J Pediatr. 125:225, 1994. PMID
8040766
223. Kaufman FR, et al: Correlation of ovarian function with galactose-1-phosphate uridyl transferase
levels in galactosemia. J Pediatr. 112:754, 1988. PMID 2834527
224. Keevill N, Holton J, Allen J: The investigation of UDPglucose and UDPgalactose concentration in
red blood cells of patients with classical galactosaemia. Clin Chim Acta. 221:135, 1993. PMID
8149631
225. Kelley RI, Harris H, Mellman WJ: Characterization of normal and abnormal variants of
galactose-1-phosphate uridyltransferase (EC 2.7.7.12) by isoelectric focusing. Hum Genet. 63:274,
1983. PMID 6303941
226. Kerr M, et al: Galactokinase deficiency in a newborn infant. Arch Dis Child. 46:864, 1971. PMID
5129188
227. Kim H-O, Hartnett C, Scaman CH: Free galactose content in selected fresh fruits and vegetables
and soy beverages. J Agric Food Chem. 55:8133, 2007. PMID 17844990
228. Kimonis V: Increased fertility in a woman with classic galactosaemia. J Inherit Metab Dis 24 607,
2001. [PMID. : 11757592]
229. Kingsley D, et al: Reversible defects in O-linked glycosylation and LDL receptor expression in a
UDP-Gal/UDP-GalNAc 4-epimerase deficient mutant. Cell. 44:749, 1986. PMID 3948246
230. Kingsley DM, Krieger M, Holton JB: Structure and function of low-density lipoprotein receptors in
epimerase-deficient galactosemia. N Engl J Med. 314:1257, 1986. PMID 3702926
231. Kirby L, et al: Screening of newborn infants for galactosemia in British Columbia. Can Med Assoc
J. 132:1033, 1985. PMID 3986728
232. Kirkman H: Uridine diphosphate glucose and uridine diphosphate galactose in galactosemia. J
Pediatr. 117:838, 1990. PMID 2231224
233. Kirkman H: Measurements of uridine diphosphate glucose and uridine diphosphate galactose An
appraisal. Eur J Pediatr. 152:S75, 1995. PMID 7671969
234. Kleijer W, et al: First-trimester diagnosis of galactosaemia. Lancet. 1:748, 1986. PMID 2870264
235. Knop J, Hansen R: Uridine diphosphate glucose pyrophosphorylase IV. Crystallization and
properties of the enzyme from human liver. J Biol Chem. 245:2499, 1970. PMID 5427280
236. Koch R, et al: Nutrition in the treatment of galactosemia. J Am Diet Assoc. 43:216, 1963. PMID
14043973
237. Koch T, et al: Neurologic complications in galactosemia. Pediatr Neurol. 8:217, 1992. PMID
1622520
238. Kogut M, et al: Observations in normal children and in patients with hereditary fructose
intolerance and galactosemia. Pediatr Res. 9:774, 1975. PMID 1187240
239. Kolosha V, et al: Novel mutations in 13 probands with galactokinase deficiency. Hum Mutat.
15:447, 2000. PMID 10790206
240. Komrower G, et al: The experience of a generation. J Inherit Metab Dis. 5:96, 1982.
241. Komrower G, Lee D: Long-term follow-up of galactosaemia. Arch Dis Child. 45:367, 1970. PMID
5427852
242. Kozak L: Mutation analysis of GALT gene in Czech and Slovak galactosemia patients.
International Congress of Inborn Errors of Metabolism,[[AU. : Location?]] 1997.
243. Kozak L, Francova H: Presence of a deletion in the 5′ upstream region of the GALT gene in
Duarte (D2) alleles. J Med Genet. 36:576, 1999. PMID 10424825
244. Kumar K, et al: Screening of the galactose-1-phosphate uridyltransferase gene in Indian women
with ovarian failure. Reprod Biomed Online. 11:444, 2005. PMID 16274605
244a. Knauff E, Richardus R, Eijkemans M, Broekmans F, de Jong F, Fauser B, Bosch A:

Heterozygosity for the classical galactosemia mutation does not affect ovarian reserve and
menopausal age. Reprod Sci. 14:780, 2007 PMID 18089596
245. Kurahashi K, Sugimura A: Purification and properties of galactose-1-phosphate uridyl transferase

from Escherichia coli. J Biol Chem. 235:940, 1960. PMID 14412847
246. Lai K, Elsas L: Overexpression of human UDP-glucose pyrophosphorylase rescues

galactose-1-phosphate uridyltransferase-deficient yeast. Biochem Biophys Res Commun. 271:392,
2000. PMID 10799308
247. Lai K, Elsas LJ: Structure-function analyses of a common mutation in blacks with
transferase-deficiency galactosemia. Mol Genet Metab. 74:264, 2001. PMID 11592823
248. Lai K, et al: The Duarte allele impairs biostability of galactose-1-phosphate uridyltransferase in
human lymphoblasts. Hum Mutat. 11:28, 1998. PMID 9450900
249. Lai K, et al: GALT deficiency causes UDP-hexose deficit in human galactosemic cells.
Glycobiology. 13:285, 2003.
250. Lai K, et al: A prevalent mutation for galactosemia among black Americans. J Pediatr. 128:89,
1996. PMID 8551426
251. Lai K, Willis A, Elsas L: The biochemical role of glutamine 188 in human galactose-1-phosphate
uridyltransferase. J Biol Chem. 274:6559, 1999. PMID 10037750
252. Landt M, et al: Correlation of activity and immunoreactive protein in erythrocytes and leukocytes.
J Pediatr. 130:972, 1997. PMID 9202622
253. Langley SD, et al: Molecular basis for Duarte and Los Angeles variant galactosemia. Am J Hum
Genet. 60:366, 1997. PMID 9012409
254. Lebea P, Pretorius P: The molecular relationship between deficient UDP-galactose uridyl
transferase (GALT) and ceramide galactosyltransferase (CGT) enzyme function A possible cause for
poor long-term prognosis in classic galactosemia. Med Hypotheses. 65:1051, 2005. PMID 16125333
255. Lee PJ, et al: A woman with untreated galactosaemia. Lancet. 362:446, 2003.
256. Lee RT, et al: Cloning of a human galactokinase gene (GK2) on chromosome 15 by
complementation in yeast. Proc Natl Acad Sci USA. 89:10887, 1992. PMID 1438294
257. Leloir L: The enzymatic transformation of uridine diphosphate glucose into a galactose derivative.
Arch Biochem. 33:186, 1951. PMID 14885999
258. Lemaire H, Meuller-Hill B: Nucleotide sequence of the galE and galT gene of E. coli. Nucleic
Acids Res. 14:7705, 1986. PMID 3022202
259. Leslie N, Bai S: Functional analysis of the mouse galactose-1-phosphate uridyl transferase
(GALT) promoter. Mol Genet Metab. 72:31, 2001. PMID 11161826
260. Leslie N, et al: UDP-galactose pyrophosphorylase in mice with galactose-1-phosphate

uridyltransferase deficiency. Mol Genet Metab. 85:21, 2005. PMID 15862277
261. Leslie ND: Insights into the pathogenesis of galactosemia. Annu Rev Nutr. 23:59, 2003. PMID
12704219
262. Leslie ND, et al: The human galactose-1-phosphate uridyl transferase gene. Genomics. 14:474,
1992.
263. Leslie ND, et al: A mouse model of galactose-1-phosphate uridyl transferase deficiency. Biochem
Mol Med. 59:7, 1996. PMID 8902187
264. Levy H: Screening for galactosemia, in Burman D, Holton J, Pennock C (eds). : Inherited
Disorders of Carbohydrate Metabolism. Lancaster, UK, MTP Press, 1980, p 133. PMID 8902187
265. Levy H, et al: Clinical and biochemical studies. J Pediatr. 93:390, 1978.
266. Levy N, Krill A, Beutler E: Galactokinase deficiency and cataracts. Am J Ophthalmol. 74:41,
1972. PMID 5036447
267. Lin H, Reichardt J: Linkage disequilibrium between a SacI restriction fragment length
polymorphism and two galactosemia mutations. Hum Genet. 95:353, 1995. PMID 7868133
268. Lin H-C, et al: On the molecular nature of the Duarte variant of galactose-1-phosphate uridyl
transferase (GALT). Hum Genet. 93:167, 1994. PMID 8112740
269. Lin M, et al: Regional mapping of the gene for human UDPGal 4-epimerase on chromosome 1 in
mouse-human hybrids. Cytogenet Cell Genet. 24:217, 1979. PMID 509992
270. Litman N, Kanter A, Finberg L: Galactokinase deficiency presenting as pseudotumor cerebri. J

Pediatr. 86:410, 1975. PMID 163306
271. Liu G, Hale G, Hughes C: Galactose metabolism and ovarian toxicity. Reprod Toxicol. 14:377,
2000. PMID 11020650
272. Lo W, et al: Curious neurologic sequelae in galactosemia. Pediatrics. 73:309, 1984.
273. Lukac-Bajalo J, et al: Higher frequency of the galactose-1-phosphate uridyl transferase gene
K285N mutation in the Slovenian population. Clin Biochem. 40:414, 2007. PMID 17303100
274. Lukac-Bajalo J, et al: Frequencies of Q188R and N314D mutations and IVS5-24g>a intron
variation in the galactose-1-phosphate uridyl transferase (GALT) gene in the Slovenian population.
Clin Chem Lab Med. 40:1109, 2002. PMID 12521227
275. Maceratesi P, et al: Missense mutations in Japanese patients with epimerase deficiency
galactosemia. Am J Hum Genet. 59:A204, 1996.
276. Maceratesi P, et al: Human UDP-galactose 4′-epimerase (GALE) gene and identification of five
missense mutations in patients with epimerase-deficiency galactosemia. Mol Genet Metab. 63:26,
1998. PMID 9538513
277. Maceratesi P, et al: Three new mutations (P183T, V150L, 528insG) and eleven sequence
polymorphisms in Italian patients with galactose-1-phosphate uridyltransferase (GALT) deficiency.
Hum Mutat. 8:369, 1996. PMID 8956044
278. Magnani M, et al: Red blood cell galactokinase activity and presenile cataracts. Enzyme. 29:58,
1983. PMID 6840050
279. Maitra U, Ankel H: Uridine diphosphate-4-keto-glucose, an intermediate in the uridine

diphosphate-galactose-4-epimerase reaction. Proc Natl Acad Sci USA. 68:2660, 1971. PMID 4941982
280. Maley F, Maley GF: The enzymatic conversion of glucosamine to galactosamine. Biochim
Biophys Acta. 31:577, 1959. PMID 13628700
281. Manga N, et al: The molecular basis of transferase galactosaemia in South African negroids. J
Inherit Metab Dis. 22:37, 1999. PMID 10070616
282. Manis F, et al: A longitudinal study of cognitive functioning in patients with classical
galactosaemia, including a cohort treated with oral uridine. J Inherit Metab Dis. 20:549, 1997. PMID
9266391
283. Marabotti A, Facchiano AM: Homology modeling studies on human galactose-1-phosphate

uridylyltransferase and on its galactosemia-related mutant Q188R provide an explanation of molecular
effects of the mutation on homo- and heterodimers. J Med Chem. 48:773, 2005. PMID 15689161
284. Markus J, et al: Evidence for a uridylyl-enzyme intermediate by kinetic and exchange reaction
studies. J Biol Chem. 252:5363, 1977. PMID 889611
285. Mathai C, Beutler E: Biochemical characteristics of galactokinase from adult and fetal human red
cells. Enzymologia. 33:224, 1967. PMID 5621948
286. Mathias D, Bickel H: Follow-up study of 16 years of neonatal screening for inborn errors of
metabolism in West Germany. Eur J Pediatr. 145:310, 1986. PMID 3770001
287. Maxwell ES: The enzymic interconversion of uridine diphosphogalactose and uridine
diphosphoglucose. J Biol Chem. 229:139, 1957. PMID 13491567
288. Mayes JS, Miller LR: The metabolism of galactose by galactosemia fibroblasts. Biochim Biophys
Acta. 313:9, 1973. PMID 4745684
289. Mehta DV, Kabir A, Bhat PJ: Expression of human inositol monophosphatase suppresses
galactose toxicity in Saccharomyces cerevisiae Possible implications for galactosemia. Biochem
Biophys Acta. 1454:217, 1999. PMID 10452956
290. Menezo Y, et al: Case report. J Assist Reprod Genet. 21:89, 2004. PMID 15202737
291. Merritt RJ, Jenks BH: Safety of soy-based infant formulas containing isoflavones The clinical
evidence. J Nutr. 134:1220S, 2004. PMID 15113975
292. Mirzajani F, et al: The first study of galactose-1-phosphate uridyl transferase mutations in Iranian
galactosemia patients. Clin Biochem. 39:697, 2006. PMID 16765930
293. Misumi H, et al: Detection of UDP-galactose-4-epimerase deficiency in a galactosemia screening

program. Clin Chim Acta. 116:101, 1981. PMID 7318169
294. Mitchell B, et al: Reversal of UDP-galactose 4-epimerase deficiency of human leukocytes in

culture. Proc Natl Acad Sci USA. 72:5026, 1975. PMID 1748
295. Mizoguchi N, et al: Galactose metabolites in blood from neonates with and without
hypergalactosaemia detected by mass screening. Eur J Pediatr. 159:851, 2000. PMID 11079200
296. Mohandas T, et al: Studies with Chinese hamsterŒhuman somatic cell hybrids. Proc Natl Acad
Sci USA. 74:5628, 1977. PMID 271990
297. Mohandas T, et al: Assignment of GALT to chromosome 9 and regional localization of GALT,
AK1, AK3, and ACONS on chromosome 9. Cytogenet Cell Genet. 22:456, 1978. PMID 222547
298. Mohandas T, et al: Studies of somatic cell hybrids containing human translocations. Am J Hum
Genet. 31:586, 1979. PMID 292306
299. Monk AM, Holton JB: Galactose-1-phosphate uridyltransferase in cultured cells. Clin Chim Acta.
73:537, 1976. PMID 11916
300. Monteleone J, et al: Studies of a kindred. Am J Med. 50:403, 1971. PMID 5553957
301. Morland S, et al: Mutation of galactose-1-phosphate uridyl transferase and its association with
ovarian cancer and endometriosis. Int J Cancer. 77:825, 1998. PMID 9714048
302. Morrow R, et al: Ovarian failure in a young woman with galactosaemia. Ulster Med J. 54:218,
1985. PMID 4095816
302a. Levy H, Driscoll S, Porensky R, Wender D: Ovarian failure in galactosemia N Engl J Med.
310:50, 1984 PMID 6689742
303. Mumma JO: , et al: Distinct roles of galactose-1-P in galactose-mediated growth arrest of yeast
deficient in galactose-1-P uridylyltransferase (GALT) and UDP-galactose 4′-epimerase (GALE). Mol
Genet Metab (in press). PMID 6689742
304. Muralidharan K, et al: A naturally occurring šknock-outŠ mutation of human

galactose-1-phosphate uridyl transferase gene reveals an alternate pathway for galactose metabolism.
Am J Hum Genet. 59:A204, 1996.
305. Murphy M, et al: Genetic basis of transferase-deficient galactosaemia in Ireland and the
population history of the Irish Travellers. Eur J Hum Genet. 7:549, 1999. PMID 10439960
306. Murphy M, et al: Frequency distribution of the Q188R mutation in the Irish galactosaemic
population. J Inherit Metab Dis. 19:217, 1996. PMID 8739969
307. Nadler H, Chacko C, Rachmeler M: Interallelic complementation in hybrid cells derived from
human diploid strains deficient in galactose-1-phosphate uridyl transferase activity. Proc Natl Acad Sci
USA. 67:976, 1970. PMID 5289034
308. Nelsestuem G, Kirkwood S: The mechanism of action of the enzyme uridine diphosphoglucose
4-epimerase Proof of an oxidation-reduction mechanism with direct transfer of hydrogen between
substrate and the B-position of the enzyme-bound pyridine nucleotide. J Biol Chem. 246:7533, 1971.
PMID 4332554
309. Nelson C, et al: Verbal dyspraxia in treated galactosemia. Pediatrics. 88:346, 1991.
310. Nelson MJ, et al: Evaluation with MR imaging. Radiology. 184:255, 1992.
311. Ng W, et al: Galactokinase activity in human erythrocytes of individuals at different ages. J Lab
Clin Med. 66:115, 1965.
312. Ng W, et al: Biochemical studies of a human low-activity galactose-1-phosphate uridyl

transferase variant. J Inherit Metab Dis. 1:145, 1978. PMID 117249
313. Ng WG, Begren WR, Donnell GN: Galactose-1-phosphte uridyltransferase activity in

galactosaemia. Nature. 203:845, 1964.
314. Ng WG, W.R. Bergren WR, Donnell GN: An improved procedure for the assay of hemolysate
galactose-1-phosphate uridyl transferase activity by the use of 14C-labeled galactose-1-phosphate.
Clin Chim Acta. 15:489, 1967. PMID 6040456
315. Ng WG, et al: Prenatal diagnosis of galactosemia. Clin Chim Acta. 74:227, 1977. PMID 188570
316. Ng WG, et al: Deficit of uridine diphosphate galactose in galactosaemia. J Inherit Metab Dis.
12:257, 1989. PMID 2515367
317. Ng WG, et al: Biochemical and molecular studies of 132 patients with galactosemia. Hum Genet.
94:359, 1994. PMID 7927329
318. Ng WG, et al: Neonatal Screening in the Nineties. : Proceedings of the 8th International Neonatal
Screening Symposium and Inaugural Meeting of the International Society for Neonatal Screening.
1991, p. 181. PMID 7927329
319. Ning C, et al: Apparent galactose appearance rate in human galactosemia based on plasma
[ 13 C]galactose isotopic enrichment. Mol Genet Metab. 70:261, 2000.
320. Ning C, et al: Sucklings and 7-week-old animals fed a high-galactose diet. Mol Genet Metab.
72:306, 2001. PMID 11286504
321. Ning C, et al: Galactose metabolism by the mouse with galactose-1-phosphate uridyltransferase
deficiency. Pediatr Res. 48:211, 2000. PMID 10926297
322. Ning C, Segal S: Plasma galactose and galactitol concentration in patients with
galactose-1-phosphate uridyltransferase deficiency galactosemia Determination by gas
chromatography/mass spectrometry. Metabolism. 49:1460, 2000. PMID 11092512
323. Nishimura Y, et al: Clinical significance of portal vein imaging. J Inherit Metab Dis. 27:11, 2004.
PMID 14970742
324. Noelmans L, et al: Pregnancy and galactosaemia. J Obstet Gynaecol 26B. : 812, 2006.
325. Oberman A, et al: Galactokinase-deficiency cataracts in identical twins. Am J Ophthalmol.

74:887, 1972. PMID 464473
326. O’Brien M, Schofield P: Polyol-pathway enzymes of human brain Partial purification and
properties of aldose reductase and hexonate dehydrogenase. Biochem J. 187:21, 1980. PMID
6773519
327. Ohlsson A, Nasiell J, Von DÃ¶beln U: Pregnancy and lactation in a woman with classical
galactosaemia heterozygous for p.Q188R and p.R333W. J Inherit Metab Dis. 30:105, 2007. PMID
17143577
328. Okano Y, et al: Identification and characterization of a novel galactokinase variant, šOsaka,Š in
Asians. Am J Hum Genet. 68:1036, 2001. PMID 11231902
329. Openo K, et al: Epimerase-deficiency galactosemia is not a binary condition. Am J Hum Genet.
78:89, 2006. PMID 16385452
330. Ornstein KS, et al: Abnormal galactosylation of complex carbohydrates in cultured fibroblasts fom
patients with galactose-1-phosphate uridyltransferase deficiency. Pediatr Res. 31:508, 1992. PMID
1603629
331. Oyanagi K, et al: Uridine diphosphate galactose 4-epimerase deficiency. Eur J Pediatr. 135:303,
1981. PMID 7227386
332. Palmieri M, et al: Urine and plasma galactitol in patients with galactose-1-phosphate
uridyltransferase deficiency galactosemia. Metabolism. 48:1294, 1999. PMID 10535394
333. Panis B, et al: Untreated classical galactosemia patient with mild phenotype. Mol Genet Metab.
89:277, 2006. PMID 16621642
334. Panis B, et al: Body composition in children with galactosaemia. J Inherit Metab Dis. 28:931,
2005. PMID 16435185
335. Panis B, et al: Bone metabolism in galactosemia. Bone. 35:982, 2004. PMID 15454106
336. Panis B, Gerver W, Rubio-Gozalbo M: Growth in treated classical galactosemia patients. Eur J
Pediatr. 166:443, 2007. PMID 17024348
337. Panis B, van Kroonenburgh M, Rubio-Gozalbo M: Proposal for the prevention of osteoporosis in
paediatric patients with classical galactosaemia. J Inherit Metab Dis (in press)
338. Panis B, et al: Effect of calcium, vitamins K1 and D3 on bone in galactosemia. Bone. 39:1123,
2006. PMID 16782422
339. Park H, et al: The molecular basis of UDP-galactose-4-epimerase (GALE) deficiency

galactosemia in Korean patients. Genet Med. 7:646, 2005. PMID 16301867
340. Park H-D, et al: Molecular and biochemical characterization of the GALK1 gene in Korean
patients with galactokinase deficiency. Mol Genet Metab. 91:234, 2007. PMID 17517531
341. Parthasarathy R, Parthasarathy L, Vadnal R: Brain inositol monophosphatase identified as a

galactose-1-phosphatase. Brain Res. 778:99, 1997. PMID 9462881
342. Pesch L, Segal S, Topper R: Progesterone effects on galactose metabolism in prepubertal

patients with congenital galactosemia and in rats maintained on high galactose diets. J Clin Invest.
39:178, 1960. PMID 14432120
343. Petry K, et al: deficiency of glycolipids containing galactose or N-acetylgalactosamine and

accumulation of precursors in brain and lymphocytes. Biochem Med Metab Biol. 46:93, 1991. PMID
1931160
344. Pickering W, Howell R: Galactokinase deficiency Clinical and biochemical findings in a new
kindred. J Pediatr. 81:50, 1972. PMID 5034870
345. Piller F, Hanlon MH, Hill RL: Co-purification and characterization of UDP-glucose 4-epimerase
and UDP-N-acetylglucosamine 4-epimerase from porcine submaxillary glands. J Biol Chem.
258:10774, 1983. PMID 6885800
346. Platt A, et al: The insertion of two amino acids into a transcriptional inducer converts it into a
galactokinase. Proc Natl Acad Sci USA. 97:3154, 2000. PMID 10737789
347. Podskarbi T, et al: A genotype-phenotype correlation in classical galactosemia, in Demirkol M,

Shin Y (eds). : Diagnosis and Treatment of Inborn Errors of Metabolism. Turkish Society for PKU,
1996, p. 56. PMID 10737789
348. Podskarbi T, et al: Biochemical and clinical phenotypes. Eur J Clin Chem. 35:A27, 1997.
349. Podskarbi T, et al: Molecular characterization of Duarte-1 and Duarte-2 variants of

galactose-1-phosphate uridyltransferase. J Inherit Metab Dis. 19:638, 1996. PMID 8892021
350. Podskarbi T, Shin Y: Neonatal screening and differential diagnosis of galactosemia, in Levy H,
Grady G (eds). : 3rd Meeting of the International Society for Neonatal Screening. Boston, Third
International Society for Neonatal Screening,1998, p. 317. PMID 8892021
351. Prchal J, Conrad M, Skalka H: Association of presenile cataracts with heterozygosity for
galactosaemic states and with riboflavin deficiency. Lancet. 1:12, 1978. PMID 74495
352. Prestoz L, et al: Altered follicle-stimulating hormone isoforms in female galactosaemia patients.
Eur J Pediatr. 156:116, 1997. PMID 9039515
353. Quan-Ma R, et al: Galactitol in the tissues of a galactosemic child. Am J Dis Child. 112:477,
1966. PMID 5917984
354. Quimby BB, et al: Characterization of two mutations associated with epimerase-deficiency
galactosemia using a yeast expression system for human UDP-galactose-4-epimerase. Am J Hum
Genet. 61:590, 1997. PMID 9326324
355. Quimby BB, et al: Functional requirements of the active site position 185 in the human enzyme
galactose-1-phosphate uridylyltransferase. J Biol Chem. 271:26835, 1996. PMID 8900165
356. Rancour N, Hawkins E, Wells W: Galactose oxidation in liver. Arch Biochem Biophys. 193:232,
1979. PMID 453851
357. Reich S, et al: An unexpectedly high frequency of hypergalactosemia in an immigrant Bosnian

population revealed by newborn screening. Pediatr Res. 51:598, 2002. PMID 11978883
358. Reichardt J, Packman S, Woo S: 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, 1991. PMID 1897530
359. Reichardt JK, et al: Different molecular mechanisms for galactosemia. Genomics. 12:596, 1992.
360. Reichardt JKV, Berg P: Cloning and characterization of a cDNA encoding human
galactose-1-phosphate uridyl transferase. Mol Biol Med. 5:107, 1988. PMID 2840550
361. Reichardt JKV, et al: On the molecular nature of the Duarte and Los Angeles variants of
galactose-1-phosphate uridyl transferase. Am J Hum Genet. 53:A943, 1993.
362. Reichardt JKV, Levy HL, Woo SL: Molecular characterization of two galactosemia mutations and
one polymorphism Implications for structure-function analysis of human galactose-1-phosphate
uridyltransferase. Biochemistry. 31:5430, 1992. PMID 1610789
363. Reichardt JKV, Woo SLC: Molecular basis of galactosemia Mutations and polymorphisms in the
gene encoding human galactose-1-phosphate uridylyltransferase. Proc Natl Acad Sci USA. 88:2633,
1991. PMID 2100574
364. Ridel, K., N. Leslie, Gilbert D: An updated review of the long-term neurological effects of
galactosemia. Pediatr Neurol. 33:153, 2005. PMID 16087312
365. Riehman K, Crews C, Fridovich-Keil JL: Relationship between genotype, activity, and galactose
sensitivity in yeast expressing patient alleles of human galactose-1-phosphate uridylyltransferase. J
Biol Chem. 276:10634, 2001. PMID 11152465
366. Robertson A, Singh RH: Outcomes analysis of verbal dyspraxia in classic galactosemia. Genet
Med. 2:142, 2000. PMID 11397328
367. Roe T, et al: Childbearing by a galactosemic woman. J Pediatr. 78:1026, 1971. PMID 5577182
368. Rogers S, Segal S: Enhanced galactose metabolism in isolated perfused livers of folate-treated
suckling rats. Metabolism. 33:634, 1984. PMID 6738364
369. Ross KL, Davis CN, Fridovich-Keil JL: Differential roles of the Leloir pathway enzymes and
metabolites in defining galactose sensitivity in yeast. Mol Genet Metab. 83:103, 2004. PMID 15464425
370. Roth S, McGuire E, Roseman S: Evidence for cell-surface glycosyltransferases Their potential
role in cellular recognition. J Cell Biol. 51:536, 1971. PMID 5165268
371. Rubio-Gozalbo M, et al: Bone mineral density in patients with classic galactosaemia. Arch Dis
Child. 87:57, 2002. PMID 12089126
372. Rubio-Gozalbo ME, et al: The endocrine system in treated patients with classical galactosemia.
Mol Genet Metab. 89:316, 2006. PMID 16935538
373. Rutherford P, Davidson D, Matthai S: Dietary calcium in galactosaemia. J Hum Nutr Diet. 15:39,
2002. PMID 11903788
374. Salo WL, et al: The specificity of UDP-glucose 4-epimerase from yeast Saccharomyces fragilis.
Biochem Biophys Acta. 151:484, 1968. PMID 5636379
375. Sangiuolo F, et al: Biochemical characterization of two GALK1 mutations in patients with
galactokinase deficiency. Hum Mutat. 23:396, 2004. PMID 15024738
376. Sardharwalla I, et al: Pregnancy in classical galactosaemia, in Burman D, Holton J, Pennock C

(eds). : Inherited Disorders of Carbohydrate Metabolism. Lancaster, UK, MTP Press, 1980, p 125.
PMID 15024738
377. Sardharwalla I, Wraith J: Galactosaemia. Nutr Health. 5:175, 1987. PMID 3328119
378. Sardharwalla I, et al: A patient with severe type of epimerase deficiency galactosemia. J Inherit
Metab Dis. 11:249, 1988. PMID 3141714
379. Sauer M, et al: Pregnancy after oocyte donation to a woman with ovarian failure and classical
galactosemia. Fertil Steril. 55:1197, 1991. PMID 2037114
380. Sayle A, Cooper G, Savitz D: Menstrual and reproductive history of mothers of galactosemic
children. Fertil Steril. 65:534, 1996. PMID 8774282
381. Schadewaldt P, et al: Galactonate determination in urine by stable isotope dilution gas
chromatographyŒmass spectrometry. J Chromatogr B. 801:249, 2004.
382. Schadewaldt P, et al: Stable-isotope dilution analysis of galactose metabolites in human

erythrocytes. Rapid Commun Mass Spectrometry. 17:2833, 2003.
383. Schadewaldt P, et al: Age dependence of endogenous galactose formation in Q188R

homozygous galactosemic patients. Mol Genet Metab. 81:31, 2004. PMID 14728989
384. Schadewaldt P, et al: Unexpected outcome in untreated galactosaemia. Mol Genet Metab.
81:255, 2004.
385. Schulpis KH, et al: UDP galactose-4-epimerase deficiency in a 5.5-year-old girl with a unilateral
cataract. J Inherit Metab Dis. 16:903, 1993. PMID 8295413
386. Schulz J, et al: Mediators of galactose sensitivity in UDP-galactose 4′-epimerase-impaired

mammalian cells. J Biol Chem. 280:13493, 2005. PMID 15701638
387. Schulz J, et al: 32796, 2004. [PMID. : 15175331]
388. Schuster V, et al: Simultaneous occurrence of various mutations and polymorphisms in cis and in
trans of the galactose-1-phosphate uridyltransferase gene in a Turkish family with classical
galactosemia. J Mol Med. 76:715, 1998. PMID 9766850
389. Schwarz V: The value of galactose phosphate determinations in the treatment of galactosaemia.
Arch Dis Child. 35:428, 1960. PMID 13749191
390. Schwarz V, et al: Some disturbance of erythrocyte metabolism in galactosaemia. Biochem J.

62:34, 1956. PMID 13293149
391. Schweitzer S: Newborn mass screening for galactosemia. Eur J Pediatr. 154:S37, 1995. PMID
7671962
392. Schweitzer S, et al: Long-term outcome in 134 patients with galactosemia. Eur J Pediatr. 152:36,
1993. PMID 8444204
393. Schweitzer-Krantz S: Early diagnosis of inherited metabolic disorders toward improving outcome
The controversial issue of galactosaemia. Eur J Pediatr. 162:S50, 2003. PMID 14614623
394. Segal S: Komrower lecture The enigma and the challenge. J Inherit Metab Dis. 21:455, 1998.
PMID 9728327
395. Segal S, Berry G: Disorders of galactose metabolism, in Scriver C, et al: (eds). : The Metabolic
and Molecular Bases of Inherited Disease. New York, McGraw-Hill, 1995, p 967. PMID 9728327
396. Segal S, Blair A: Some observations on the metabolism of d-galactose in normal man. J Clin
Invest. 40:2016, 1961. PMID 16695871
397. Segal S, Blair A, Roth H: The metabolism of galactose by patients with congenital galactosemia.
Am J Med. 38:62, 1965. PMID 14247292
398. Segal S, Rutman J, Frimpter G: Galactokinase deficiency and mental retardation. J Pediatr.
95:750, 1979. PMID 490246
399. Sellick CA, Reece RJ: Contribution of amino acid side chains to sugar binding specificity in a
galactokinase, Gal1p, and a transcriptional inducer, Gal3p. J Biol Chem. 281:17150, 2006. PMID
16603548
400. Seyrantepe V, et al: Identification of mutations in the galactose-1-phosphate uridyltransferase

(GALT) gene in 16 Turkish patients with galactosemia, including a novel mutation of F294Y. Mutation
in brief no. 235. Online Hum Mutat. 13:339, 1999. PMID 10220154
401. Shaw C, Koen A: šGalactose dehydrogenase,Š šnothing dehydrogenase,Š and alcohol

dehydrogenase interrelation. Science. 156:1517, 1967.
402. Shield JP, et al: The relationship of genotype to cognitive outcome in galactosaemia. Arch Dis
Child. 83:248, 2000. PMID 10952646
403. Shih L, et al: Gene dosage studies supporting localization of the structural gene for
galactose-1-phosphate uridyl transferase (GALT) to band p13 of chromosome 9. Am J Med Genet.
19:539, 1984. PMID 6095663
404. Shin Y, et al: Uridine diphosphogalactose, galactitol and galactose-1-phosphate concentration in

patients with classical galactosemia. Proc Annu Symp SSIEM. 23:34, 1985.
405. Shin YS, et al: Three missense mutations in the galactose-1-phosphate uridyltransferase gene of
three families with mild galactosaemia. Eur J Pediatr. 155:393, 1996. PMID 8741038
406. Shin YS, et al: Two new mutations in the GALT gene which cause a GALT activity decrease by
40Œ50% of normal in red cells. J Inherit Metab Dis. 21:232, 1998. PMID 9686364
407. Shin YS, et al: Activity expression in patients with cataracts and mental retardation. J Inherit
Metab Dis. 23:383, 2000. PMID 10896300
408. Shin YS, et al: Developmental aspects of UDP-galactose pyrophorylase. Clin Chim Acta. 166:27,
1987. PMID 3038381
409. Shin YS, et al: Molecular and biochemical basis for variants and deficiency forms of
galactose-1-phosphate uridyltransferase. J Inherit Metab Dis. 22:327, 1999. PMID 10384398
410. Shin-Buehring Y, et al: The activity of galactose-1-phosphate uridyltransferase and galactokinase

in human fetal organs. Pediatr Res. 11:1003, 1977.
411. Shin-Buehring Y, et al: A simple assay for galactokinase using DEAE-cellulose column
chromatography. Clin Chim Acta. 74:1, 1977. PMID 12888
412. Siegel C, Sparks J, Battaglia F: Patterns of serum glucose and galactose concentrations in term
newborn infants after milk feeding. Biol Neonate. 54:301, 1988. PMID 3228558
413. Simonelli F, et al: Galactokinase activity in patients with idiopathic presenile and senile cataract.
Metab Pediatr Syst Ophthalmol. 15:53, 1992. PMID 1305705
414. Simoons F: A geographic approach to senile cataracts Possible links with milk consumption,
lactase activity, and galactose metabolism. Dig Dis Sci. 27:257, 1982. PMID 6804198
415. Skalka HW, Prchal JT: Presenile cataract formation and decreased activity of galactosemic
enzymes. Arch Ophthalmol. 98:269, 1980. PMID 7352874
416. Slepak T, et al: Intracellular galactose-1-phosphate accumulation leads to environmental stress

response in yeast model. Mol Genet Metab. 86:360, 2005. PMID 16169270
417. Slepak TI, et al: Involvement of endoplasmic reticulum stress in a novel classic galactosemia
model. Mol Genet Metab. 92:78, 2007. PMID 17643331
418. Sommer M, et al: Mutations in the galactose-1-phosphate uridyltransferase gene of two families
with mild galactosaemia variants. J Inherit Metab Dis. 18:567, 1995. PMID 8598637
419. Soni T, et al: Physicochemical and catalytic properties. Clin Chim Acta. 175:97, 1988. PMID
3168287
420. Sparkes R, et al: Assignment of GALT locus to 9cen leads to 9p22. Ann Hum Genet. 43:343,
1980. PMID 6249180
421. Srivastava S, Beutler E: Auxiliary pathways of galactose metabolism Identification of reaction

products of hexose 6-phosphate dehydrogenase and of šgalactose dehydrogenase. Š J Biol Chem.
244:6377, 1969. PMID 4390966
422. Stambolian D: Galactose and cataract. Surv Ophthalmol. 32:333, 1988. PMID 3043741
423. Stambolian D, et al: Cloning of the galactokinase cDNA and identification of mutations in two
families with cataracts. Nat Genet. 10:307, 1995. PMID 7670469
424. Stambolian D, et al: Cataracts in patients heterozygous for galactokinase deficiency. Invest
Ophthalmol Vis Sci. 27:429, 1986. PMID 3949470
425. Stambolian D, Scarpino-Myers V, Harris H: Purification of human galactokinase and evidence for
its existence as a monomer form. Biochim Biophys Acta. 831:306, 1985. PMID 4052432
426. Starling J, Keppler D: Metabolism of 2-deoxy-d-galactose in liver induces phosphate and uridylate
trapping. Eur J Biochem. 80:373, 1977. PMID 923584
427. Stefansson H, et al: Endometriosis is not associated with or linked to the GALT gene. Fertil Steril.
76:1019, 2001. PMID 11704127
428. Stenstam T, et al: Comparative study of their significance in different conditions. Acta Med
Scand. 177:1, 1946.
429. Stevens RE, et al: Idiopathic presenile cataract formation and galactosemia. Br J Ophthalmol.
73:48, 1989. PMID 2537652
430. Stibler H, et al: Carbohydrate-deficient transferrin in galactosaemia. Acta Paediatr. 86:1377,

1997. PMID 9475320
431. Sturiale L, et al: Hypoglycosylation with increased fucosylation and branching of serum transferrin
N-glycans in untreated galactosemia. Glycobiology. 15:1268, 2005.
432. Suzuki M, West C, Beutler E: Large-scale molecular screening for galactosemia alleles in a
pan-ethnic population. Hum Genet. 109:210, 2001. PMID 11511927
433. Tajima M, Nogi Y, Fukasara T: Primary structure of the Saccharomyces cerevisiae GALT gene.
Yeast. 1:67, 1985. PMID 2851900
434. Tedesco T, et al: The Philadelphia variant of galactokinase. Am J Hum Genet. 29:240, 1977.
PMID 194478
435. Tedesco T, et al: A population survey. Am J Hum Genet. 27:737, 1975. PMID 173184
436. Thoden J, Holden H: The molecular architecture of galactose mutarotase/UDP-galactose

4-epimerase from Saccharomyces cerevisiae. J Biol Chem. 280:21900, 2005. PMID 15795221
437. Thoden J, Holden H: The molecular architecture of human N-acetylgalactosamine kinase. J Biol
Chem. 280:32784, 2005. PMID 16006554
438. Thoden J, et al: Molecular structure of human galactose mutarotase. J Biol Chem. 279:23431,
2004. PMID 15026423
439. Thoden J, et al: Implications for type II galactosemia. J Biol Chem. 280:9662, 2005. PMID
15590630
440. Thoden JB, Gulick AM, Holden HM: Molecular structures of the S124A, S124T, and S124V
site-directed mutants of UDP-galactose 4 epimerase from Escherichia coli. Biochemistry. 36:10685,
1997.
441. Thoden JB, et al: Molecular structure of Saccharomyces cerevisiae Gal1p, a bifunctional
galactokinase and transcriptional inducer. J Biol Chem. 280:36905, 2005. PMID 16115868
442. Thoden JB, et al: Crystallographic evidence for tyr157 functioning as the active site base in
human UDP-galactose 4-epimerase. Biochemistry. 39:5691, 2000.
443. Thoden JB, et al: Accommodation of UDP-N-acetylglucosamine within the active site. J Biol
Chem. 276:15131, 2001. PMID 11279032
443a. Thoden J, Henderson J, Fridovich-Keil J, Holden H: Structural analysis of the Y299C mutant of
Escherichia coli UDP-galactose 4-epimerase. Teaching an old dog new tricks. J Biol Chem.
277:27528, 2002 PMID 12019271
444. Thoden JB, et al: X-ray structure of the human V94M-substituted UDP-galactose 4-epimerase. J
Biol Chem. 276:20617, 2001. PMID 11279193
445. Timson D: Functional analysis of disease-causing mutations in human UDP-galactose

4-epimerase. FEBS J. 272:6170, 2005. PMID 16302980
446. Timson D, Reece R: Functional analysis of disease-causing mutations in human galactokinase.

Eur J Biochem. 270:1767, 2003. PMID 12694189
447. Timson D, Reece R: Sugar recognition by human galactokinase. BMC Biochem. 4:16, 2003.
PMID 14596685
448. Timson DJ: The structural and molecular biology of type III galactosemia. IUBMB Life. 58:83,
2006. PMID 16611573
449. Townsend EJ, Mason H, Strong P: Galactosemia and its relation to Laennec’s cirrhosis Review of
the literature and presentation of six additional cases. Pediatrics. 7:760, 1951.
450. Trbusek M, Francova H, Kozak L: Galactosemia Deletion in the 5′ upstream region of the GALT
gene reduces promoter efficiency. Hum Genet. 109:117, 2001. PMID 11479743
451. Turnquist R, et al: Differential heat inactivation and further characterization of human liver
enzyme. Biochim Biophys Acta. 364:59, 1974. PMID 4433565
452. Tyfield L: Galactosaemia and allelic variation at the galactose-1-phosphate uridyltransferase

gene A complex relationship between genotype and phenotype. Eur J Pediatr. 159:S204, 2000. PMID
11216901
453. Tyfield L, et al: Mutation analysis and psychometric assessments in the galactosaemic population
of the British Isles. J Med Genet. 34:S18, 1997.
454. Tyfield L, et al: Preliminary evidence on 500 chromosomes for stratification by population and
region in Europe. Proceedings of the EurSci Found Conference on inherited disorders and their genes
in different European populations,[[AU. : Location?]] 1998.
455. Tyfield L, et al: Classical galactosemia and mutations at the galactose-1-phosphate uridyl
transferase (GALT) gene. Hum Mutat. 13:417, 1999. PMID 10408771
456. Tyfield L, Walter J: Galactosemia, in Scriver C, et al: (eds) The Metabolic and Molecular Bases of
Inherited Disease. New York. : McGraw-Hill, 2002.
457. Tygstrup N, Winkler K: Galactose blood clearance as a measure of hepatic blood flow. Clin Sci
(Lond). 17:9, 1958.
458. Vannas A, et al: Lens changes in a galactosemic fetus. Am J Ophthalmol. 80:726, 1975. PMID
1237236
459. Verma K, Jamwal D, Gupta S: A variant of galactosemia presenting as larger bilateral lenticular
opacities in an older child. Indian Pediatr. 16:375, 1979. PMID 511314
460. Viestenz A, et al: A case report]. Klin Monatsbl Augenheilkd. 218:121, 2001. PMID 11258124
461. von Reuss A: Zuckerausscheidung im Sauglingsalter Wein Med Wochenschr. 58:799, 1908.
462. Wada E: r-Galactonolactone in experimental galactosemic animals. Arch Biochem Biophys. 251:
215, 1986. PMID 3098177
463. Waggoner DD, Buist NRM, Donnell SN: Long-term prognosis in galactosemia Results of a survey
of 350 cases. J Inherit Metab Dis. 13:802, 1990. PMID 1706789
464. Waisbren S, et al: Speech and language deficits in early-treated children with galactosemia. J
Pediatr. 102:75, 1983. PMID 6848731
465. Walker D, Khan H: Some properties of galactokinase in developing rat liver. Biochem J. 108:169,
1968. PMID 5665881
466. Wallenfels K, Hucho F, Herrmann K: Enzymatically catalyzed mutarotation of aldoses Studies on

aldose-1-epimerase from E. coli. Biochem Z. 343:307, 1965. PMID 4290293
467. Walter JH, et al: Commentary. Arch Dis Child. 80:93, 1999. PMID 10325771
468. Walter J.H, et al: Generalised uridine diphosphate galactose-4-epimerase deficiency. Arch Dis
Child. 80:374, 1999. PMID 10086948
469. Wang BBT, et al: Molecular and biochemical basis of galactosemia. Mol Genet Metab. 63:263,
1998. PMID 9635294
470. Wang L, et al: H7 and identification of a new UDP-GlcNAc C4 epimerase gene. J Bacteriol.
184:2620, 2002. PMID 11976290
471. Wang Z, et al: Proton magnetic resonance spectroscopy of brain metabolites in galactosemia.
Ann Neurol. 50:266, 2001. PMID 11506413
472. Wasilenko J, Fridovich-Keil J: Relationship between UDP galactose 4′-epimerase activity and
galactose sensitivity in yeast. J Biol Chem. 281:8443, 2006. PMID 16452467
473. Wasilenko J, et al: Functional characterization of the K257R and G319E hGALE alleles found in
patients with ostensibly peripheral epimerase deficiency galactosemia. Mol Genet Metab. 84:32, 2005.
PMID 15639193
474. Webb A, et al: Verbal dyspraxia and galactosemia. Pediatr Res. 53:396, 2003. PMID 12595586
475. Webb P, et al: Milk consumption, galactose metabolism and ovarian cancer (Australia). Cancer
Causes Control. 9:637, 1998. PMID 10189050
476. Wedekind JE, Frey PA, Rayment I: Three-dimensional structure of falactose-1-phosphate

uridylyltransferase from Escherichia coli at 1.8 Ǻ resolution. Biochemistry. 34:11049, 1995.
477. Wedekind JE, Frey PA, Rayment I: The structure of nucleotidylated histidine-166 of
galactose-1-phosphate uridylyltransferase provides insight into phosphoryl group transfer.
Biochemistry. 35:11560, 1996.
478. Wehrli S, et al: Quantitation by nuclear magnetic resonance spectroscopy. Pediatr Res. 42:855,
1997. PMID 9396569
479. Wehrli S, et al: Metabolism of 13 C galactose by lymphoblasts from patients with galactosemia
determined by NMR spectroscopy. Mol Genet Metab. 77:296, 2002. PMID 12468275
480. Wells L, Fridovich-Keil JL: Biochemical characterization of the S135L allele of

galactose-1-phosphate uridylyltransferase associated with galactosemia. J Inherit Metab Dis. 20:633,
1997. PMID 9323558
481. Wells W, Pittman T, Egan T: The isolation and identification of galactitol from the urine of patients
with galactosemia. J Biol Chem. 239:3192, 1964. PMID 14245360
482. Wells W, et al: The isolation and identification of galactitol from the brains of galactosemic
patients. J Biol Chem. 240:1002, 1965. PMID 14284692
483. Wen Y, Bekhor I: Levels of expression of hexokinase, aldose reductase and sorbitol
dehydrogenase genes in lens of mouse and rat. Curr Eye Res. 12:323, 1993. PMID 8319491
483a. Lee A, Chung S: Demonstration that polyol accumulation is responsible for diabetic cataract by
the use of transgenic mice expressing the aldose reductase gene in the lens. Proc Natl Acad Sci USA.
92:2780, 1995 PMID 7708723
484. Wiesmann U, RosÃ©-Beutler B, SchlÃ¼chter R: Leguminosae in the diet The

raffinose-stachyose question. Eur J Pediatr. 154:S93, 1995. PMID 7671975
485. Williams V: Purification and some properties of galactose 1-phosphate uridyltransferase from
human red cells. Arch Biochem Biophys. 191:182, 1978. PMID 73656
486. Wilson DB, Hogness DS: The enzymes of the galactose operon in Escherichia coli. J Biol Chem.
239:2469, 1964. PMID 14235524
487. Winder A, et al: Partial galactose disorders in families with premature cataracts. Arch Dis Child.
58:362, 1983. PMID 6859916
488. Witting L, Haberland C, Brunngraber E: Ganglioside patterns in galactosemia. Clin Chim Acta.
37:387, 1972. PMID 5022102
489. Wohlers T, Fridovich-Keil JL: Studies of the V94M-substituted human

UDP-galactose-4-epimerase enzyme associated with generalized epimerase-deficiency galactosemia.
J Inherit Metab Dis. 23:713, 2000. PMID 11117433
490. Wohlers TM, et al: Identification and characterization of a mutation, in the human UDP
galactose-4-epimerase gene, associated with generalized epimerase-deficiency galactosemia. Am J
Hum Genet. 64:462, 1999. PMID 9973283
491. Wolfrom C, et al: Connection with cell growth and amino acid metabolism and possible bearing
on late-onset clinical symptoms. J Inherit Metab Dis. 16:78, 1993. PMID 8098071
492. Wong LJ, et al: Genotyping of 277 galactosemia patients. Am J Hum Genet. 59:A295, 1996.
493. Xu Y-K, et al: Decreased concentrations of uridine diphosphate galactose in erythrocytes and
cultured skin fibroblasts from classical galactosemia patients. Clin Chim Acta. 240:21, 1995. PMID
8582057
494. Xu Y-K, et al:. Radiochemical assay of minute quantities of galactose-1-phosphate

uridyltransferase activity in erythrocytes and leukocytes of galactosemia patients. Clin Chim Acta.
235:125, 1995. PMID 7554267
495. Yager C, et al: Galactitol and galactonate in red blood cells of galactosemic patients. Mol Genet
Metab. 80:283, 2003. PMID 14680973
496. Yager C, et al: Oxidation of galactose by galactose-1-phosphate uridyltransferase-deficient

lymphoblasts. J Inherit Metab Dis. 24:465, 2001.
497. Yager C, Wehrli S, Segal S: Urinary galactitol and galactonate quantified by isotope-dilution gas
chromatographyŒmass spectrometry. Clin Chim Acta 366 216, 2006. [PMID. : 14680973]
498. Yang J, et al: Studies on the substrate specificity of Escherichia coli galactokinase. Org Lett.
5:2223, 2005.
499. Yang J, et al: Structure-based engineering of E. coli galactokinase as a first step toward in vivo
glycorandomization. Chem Biol. 12:657, 2005. PMID 15975511
500. Yang Y-P, Corley N, Garcia-Heras J: Molecular analysis in newborns from Texas affected with
galactosemia. Hum Mutat. 19:82, 2002. PMID 11754113
501. Zekanowski C, Radomyska B, Bal J: Molecular characterization of Polish patients with classical
galactosaemia. J Inherit Metab Dis. 22:679, 1999. PMID 10399107
502. Zlatunich C, Packman S: Galactosaemia Early treatment with an elemental formula. J Inherit
Metab Dis. 28:163, 2005. PMID 15877205

Galactosemia - Scriver. MMBID 2006

Uploaded by

Copyright:

Available Formats

Galactosemia - Scriver. MMBID 2006

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Galactosemia - Scriver. MMBID 2006

Uploaded by

Copyright:

Available Formats

Chapter 72: Galactosemia

Chapter 72: Galactosemia

1. Galactosemia results from an impaired ability to metabolize galactose.

Dietary Sources of Galactose

Endogenous Production of Galactose

PATHWAYS OF GALACTOSE METABOLISM

The structures of galactose and glucose.

The Leloir Pathway

Conversion of β-D-galactose to α-D-galactose

Conversion of α-D-galactose to glucose-1-phosphate

abnormalities of development initiated in utero.

The Human GALK (GK1) Gene and cDNA

Total 1179 392

c This suggests multiple starting sites.

Submitted under GenBank accession number L76927 66

Structure and Catalytic Mechanism of Human GALK

Recently, a combination of biochemical and site-directed mutagenesis studies coupled with

Structure of human GALK defined by x-ray crystallography. (Courtsey of H. Holden.)

Galactose-1-phosphate uridylyltransferase (GALT)

The Human GALT Gene and cDNA

Structure and Catalytic Mechanism of Human GALT

the Gal-1-P, forming UDP-Gal, which is then released.

Uridine Diphosphate Galactose-4′-Epimerase (GALE)

The Human GALE Gene and cDNA

Nucleotide number Exon

Total 1488 c 348

c Including the 5-untranslated region of exon 1.

Submitted under GenBank accession number A 1022382 104

Structure and Catalytic Mechanism of Human GALE

Accessory Pathways of Galactose Metabolism

Reduction of Galactose to Galactitol

Oxidation of Galactose to Galactonate

An alternative oxidative pathway of galactose metabolism via galactonate.

The Pyrophosphorylase Pathway

The existence of the pyrophosphorylase, or UGP, pathway is supported by a preponderance of recent

Diagnosis: Acute Clinical Presentation, Enzyme Defect, and Metabolic Abnormality

Acute clinical presentation

Treatment and Outcome

Mutations at the GALK Locus

GALT DEFICIENCY (CLASSIC GALACTOSEMIA)

Diagnosis: Acute Clinical Presentation, Enzyme Defect, and Metabolic Abnormality

Acute clinical presentation

Variants of transferase-deficiency galactosemia

A number of studies have explored galactitol accumulation in galactosemia, addressing questions of

UDP-Gal and UDP-Glc

Role of Cell Type

Role of Residual GALT Activity

GALT Enzyme Assay

Treatment and Outcome

appropriate antibiotic should be administered intravenously. Unconjugated hyperbilirubinemia is usually

Soya Formula versus Elemental Formula

PARTial GALT Deficiency

Long-term management and outcome of patients on lifelong dietary restriction of galactose

linkages, is an important component of a number of plant compounds such as the raffinose

Additional or Alternative Treatments

Speech and Language

Ataxia and Other Late Neurologic Complications

2 Ethinyl estradiol at a dose of 5 mg/d.

Loestrin (an oral contraceptive preparation containing 20 mg ethinyl estradiol and a

• Physical development and growth throughout these years of pubertal induction.