Abstract
purpose. Macular corneal dystrophy (MCD) is a rare corneal dystrophy
that is characterized by abnormal deposits in the corneal stroma,
keratocytes, Descemet’s membrane, and endothelium, accompanied by
progressive clouding. It has been classified into three
immunophenotypes—MCD types I, IA, and II—according to the serum level
of sulfated keratan sulfate (KS) and immunoreactivity of the corneal
tissue. Recently, mutations in a new carbohydrate sulfotransferase gene
(CHST6) encoding corneal glucosamine N-acetyl-6-sulfotransferase (C-GlcNac-6-ST) have been
identified as the cause of MCD. Mutation screening of the CHST6 gene has been undertaken to identify the
underlying mutations in five unrelated British families with MCD.
methods. DNA was extracted from venous blood obtained from all participants, and
the coding region of CHST6 was amplified by polymerase
chain reaction (PCR). The PCR products were analyzed by direct
sequencing and restriction enzyme digestion. Enzyme-linked
immunosorbent assay (ELISA) was performed to assess the presence of KS
in serum from the probands of MCD-affected families participating in
the study.
results. Six novel missense mutations—four homozygous and two compound
heterozygous—were identified in the CHST6 gene. The
ELISA showed that the disease in all patients participating in the
study was of MCD type I, including the subtype IA.
conclusions. These novel mutations are thought to result in loss of corneal
sulfotransferase function, which would account for the MCD
phenotype.
Macular corneal dystrophy (MCD, On-line Mendelian
Inheritance in Man (OMIM) 217800; provided in the public domain by the
National Center for Biotechnology Information, Bethesda, MD, and
available at http://www3.ncbi.nlm.nih.gov/omim) is an autosomal
recessive disorder characterized by bilateral progressive stromal
clouding and central corneal thinning.
1 2 MCD accounts for
10% to 75% of the corneal dystrophies that require penetrating
keratoplasty, depending on the population.
3 4 The disease
becomes evident in the first decade of life, defined by fine, diffuse,
superficial clouding in the central stroma that extends to the
periphery and usually involves the entire thickness of the cornea by
the second decade of life. These opacities are more superficial and
prominent in the central cornea and are deeper and more discrete in the
periphery.
MCD has been classified into three immunophenotypes—I, IA, and
II—based on measurement of the serum level of sulfated keratan sulfate
(KS) by enzyme-linked immunosorbent assay (ELISA)
5 and an
immunohistochemical evaluation of the corneal tissue.
6 In
MCD type I, neither the cornea nor the serum contains appreciable
levels of sulfated KS, whereas in MCD type II, there is detectable KS
in the cornea and serum.
7 The third subtype, IA, in which
sulfated KS is absent in the cornea and the serum but can be detected
in the keratocytes, has been identified in subjects from Saudi
Arabia.
8 All three immunohistochemical subtypes have the
same clinical phenotype.
Histologically, MCD is characterized by accumulation of
glycosaminoglycans between the stromal lamellae, underneath the
epithelium, and within the keratocytes and endothelial
cells.
9 10 An abnormality in the metabolism of KS has been
implicated in the pathogenesis of MCD
11 and has been
attributed to an error in a specific sulfotransferase involved in the
sulfation of KS.
12 13 14
MCD types I and II had previously been linked to chromosome 16
(16q22).
15 16 17 Recently, mutations in a new carbohydrate
sulfotransferase gene (
CHST6) encoding corneal glucosamine
N-acetyl-6-sulfotransferase (C-GlcNac-6-ST) have been
identified as the cause of MCD.
18 A number of missense
mutations were described in patients with MCD type I
18 19 20 or type IA.
20 In MCD type II, deletions and/or
rearrangements in the upstream region of
CHST6 18 and recently a missense
mutation
20 were reported.
In our study, mutation screening of CHST6 in five unrelated
British families with MCD and with no clinically apparent systemic
manifestations, revealed four homozygous and two compound heterozygous
mutations in CHST6, all of the missense type. All the
mutations involve the substitution of amino acids highly conserved
across related carbohydrate sulfotransferases and/or represent the
nonconservative substitution of nonpolar for polar residues and are
thought to result in loss of CHST6 gene function.
The study had the approval of Moorfields Eye Hospital Ethics
Committee and conformed to the tenets of the Declaration of Helsinki.
Informed consent was obtained from all participants for clinical and
molecular genetic study. Autosomal recessive MCD in our patients of
white origin was diagnosed on the basis of pedigree structure and the
following clinical features: bilateral symmetrical superficial stromal
cloudiness studded by small, irregular, rounded, gray-white anterior
stromal patches. The diagnosis was confirmed by histopathologic
examination of corneal buttons from all patients after keratoplasty.
Genomic DNA was extracted from 10-mL blood samples obtained,
according to standard protocols, from all subjects involved in the
study. The coding region of CHST6 was amplified by
polymerase chain reaction (PCR). Each PCR was performed in a 50-μL
reaction mixture containing genomic DNA (200 ng), primers (0.4 μM
each), MgCl2 (1.5–2 mM), deoxynucleoside
triphosphates (dNTPs; 0.2 mM), 1× PCR buffer (containing 10 mM
Tris-HCl [pH 8.3], 50 mM KCl, and 0.1% gelatin) and Taq polymerase (0.5 U; Bioline, UK). Amplification reactions were
performed under the following conditions: 3 minutes of denaturation at
94°C followed by 35 cycles of denaturation at 94°C for 1 minute,
annealing at 60 to 68°C for 1 minute, extension at 72°C for 1
minute, and a further extension step at 72°C for 5 minutes.
For amplification of the
CHST6 coding region the following
primer pairs were used: for the 5′-coding region, CK71h-intrn
(5′-GCCCCTAACCGCTGCGCTCTC-3′) and CK71h-R1180
(5′-GGCTTGCACACGGCCTCGCT-3′) designed by Akama et al.
18 For the middle coding region, two pairs of primers were designed:
CK71M-F1 (5′-GACATGGACGTGTTTGATGC-3′) and CK71M-R1
(5′-GCACGATGCCGTTGTCAC-3′); CK71M-F2 (5′-GCTCAACCTACGCATCGTG-3′) and
CK71M-R2 (5′-ATCCGTGGGTGATGTTATGG-3′); and for the 3′-coding region the
following primer pair was designed: CK71L-F (5′-GAGCCGCTGGCAGAAATC-3′)
and CK71L-R (5′-TGCACCATGCACTCTCCTC-3′). A mismatch primer pair was
also designed, sulfohin-F (5′-CTGTGCGACATGGACGAGT-3′) and sulfohin-R
(5′-CACCACGTGGCTGTAGGAG-3′) to confirm one of the homozygous mutations
(in family A, F107S) which does not alter any known restriction enzyme
site. The mismatch primer created an
HinfI site in
conjunction with the mutation.
For direct sequencing, PCR products were purified (Qiaquick PCR
purification kit; (Qiagen, Crawley, UK) and sequenced using an
automatic fluorescence DNA sequencer (ABI Prism 373A; Perkin Elmer,
Foster City, CA), according to the manufacturers’ instructions.
Nucleotide sequences were compared with the published cDNA sequence of CHST6 (GenBank accession number AF219990; GenBank is
provided in the public domain by the National Center for Biotechnology
Information, Bethesda, MD, and is available at
http://www.ncbi.nlm.nih.gov/genbank), and mutations identified were
excluded from 100 control chromosomes of white origin by restriction
enzyme digestion.
Five enzymes were used to confirm the mutations
identified
. HinfI, with recognition site 5′…
G
ANTC… 3′, was used to study the
homozygous T1012C and C906T in families A and E, respectively.
ApaI, with recognition site 5′…
GGGCC
C… 3′, and
BstNI, with
recognition site 5′… CC
(A/T)GG… 3′,
also were used to confirm the heterozygous changes C783T and T1291G in
families B and C, respectively. Similarly,
DrdI, with
recognition site 5′… GACNNNN
NNGTC…
3′, was used to confirm the homozygous C1309T mutation in family D. And
finally,
BanII, with recognition site 5′… G (A/G) GC
(T/C)
C… 3′, was used to confirm the
homozygous G905T change in family E.
The PCR products digested by HinfI, BstN I, and BanII restriction enzymes were analyzed by 6% nondenaturing
polyacrylamide gel electrophoresis (Protogel; National Diagnostics,
Atlanta, GA) and were visualized by staining with ethidium bromide.
Products digested by ApaI and Drd I were analyzed
on 3% agarose gels (Bio-Rad, Herts, UK).
Inhibition ELISA for KS detection using the monoclonal antibody
5-D-4 (ICN Biochemicals Ltd., Basingstoke, UK) was performed to assess
the presence of KS in serum obtained from the probands of MCD-affected
families participating in the study. The method was performed as
described previously,
21 22 except that an antibody
(Aggrecan; Sigma Chemical Co., Poole, UK) was used to coat the
plates and 2, 2′-azino-di-[3-ethylbenzthiazoline sulfonate 6]
(ABTS; Roche Molecular Biochemicals, Lews, UK) was used as the
substrate. The plates were read at 410 nm with a reference wavelength
of 490 nm on a flow plate reader (ICN Biochemicals Ltd.). The ELISA was
sensitive to KS concentrations of 10 to 10,000 ng/mL.
Six novel missense mutations were identified within the
coding region of the
CHST6 gene in five unrelated families
with MCD type I
(Table 1) . None of these changes was detected in the control population
of the same ethnic origin, as determined by restriction enzyme digest
on 100 chromosomes, confirming that they are likely to be pathogenic
mutations. In addition, they were shown to segregate with the disease
by sequencing and restriction enzyme analyses
(Fig. 1) .
Four of these mutations are homozygous and were detected in families A,
D, and E. In family A, a T1012C transition at the second nucleotide
position of codon 107 was identified, resulting in a
phenylalanine-to-serine substitution
(Figs. 2A 2B) . In family D, a C1309T transition at the second nucleotide position
of codon 206 was found, with an alanine-to-valine substitution
(Figs. 2G 2H) . In family E, two consecutive homozygous changes were
identified. The first is a G905T transversion, located at the third
nucleotide position of codon 71, which changes the amino acid from
glutamic acid to aspartic acid. The second change, a C906T transition
located at the first nucleotide position of codon 72 leads to a
proline-to-serine substitution in the same patient
(Figs. 2I 2J) .
Although the first mutation predicts a change in the amino acid within
the same group, the second alteration changes the amino acid residue to
a different anionic group.
In families B and C two heterozygous changes were identified, with a
different mutation on each allele (i.e., compound heterozygotic
mutations), which could account for such autosomal recessive
inheritance. The first heterozygous change, a C783T transition,
occurred at the first nucleotide position of codon 31, predicting a
change of amino acid proline to serine
(Figs. 2C 2D) . The other
change, a T1291G transversion, was identified at the second nucleotide
position of codon 200, changing the amino acid from leucine to arginine
(Figs. 2E 2F) .
To assess the significance of these mutations, an alignment of the
primary sequence of
CHST6 with other published carbohydrate
sulfotransferases,
23 24 25 26 27 a computer program for
simultaneous alignment of many nucleotide or amino acid sequences to
detect or demonstrate homology between new sequences and existing
families of sequences (Clustal W, ver. 1.7; SGI, Mountain View, CA) was
used
28 (Fig. 3) . With the exception of a P31S substitution (families B and C) that
represents a nonconservative change but occurs at a position where
residue type is not conserved across the carbohydrate
sulfotransferases, all the mutations occur at positions where the
residue type is highly conserved across all the carbohydrate
sulfotransferases. Furthermore, mutations F107S (family A), L200R
(families B and C), and P72S (family E) are all nonconservative changes
involving substitution of a nonpolar residue for a polar residue that
may affect
CHST6 enzyme activity. Two of the substitutions
are conservative changes, E71D (acidic→acidic, family E) and A206V
(nonpolar→nonpolar, family D). Of these, E71D occurs in tandem with
the P72S substitution (which is more likely to be disease causing) and
thus may not itself have a major deleterious effect. It is not
immediately apparent, however, why the A206V substitution should be
disease causing, although conservative substitutions elsewhere in
CHST6 have been reported to underlie MCD type I in Icelandic
families.
19
In the alignment, 5′- and 3′-phosphate-binding (PB) domains that
interact with adenosine 3′-phosphate 5′-phosphosulfate (PAPS), a
sulfate donor for sulfotransferases, are boxed
(Fig. 3) .
29 30 These domains form part of the vital active site
region of the enzyme. Two of the new mutations identified in this
study, L200R and the conservative substitution A206V, occur within the
3′-PB domain.
Inhibition ELISA performed on serum from the patients with MCD showed
little or no inhibition of the monoclonal antibody 5-D-4. Control
experiments using increasing KS concentrations effectively inhibited
5-D-4 binding. These data from our patients are consistent with MCD
type I or type IA.
In this study we collected five unrelated British families with
clinically and histopathologically diagnosed MCD and screened the CHST6 gene, which is reported to be involved in the
pathogenesis of this disorder by affecting the synthesis of KS.
KS proteoglycans (lumican, keratocan, and mimecan) are present in the
cornea as the major class of proteoglycans and are thought to play an
important role in corneal transparency.
31 32 33 The
sulfate group of KS appears to be crucial for its biological function,
because the degree of sulfation of KS increases during corneal
development
34 35 and the synthesis of unsulfated KS is
observed in the corneas of patients with MCD.
36 37 In view
of the autosomal recessive inheritance of the condition, MCD probably
results from a deficiency in sulfotransferase specific for proper
sulfation of KS.
10
We have identified six novel missense mutations in the probands from
all families participating in this study. The mutations were detected
by direct sequencing of CHST6 PCR products. Using
restriction enzyme analysis, we were able to confirm the mutations in
the probands and exclude them from unaffected family members and 50
normal individuals. To measure levels of sulfated KS in the serum of
our patients with MCD, ELISA was performed and revealed barely
detectable levels in all patients, confirming that disease in our
patients was of MCD type I or IA.
All mutations detected occur at positions in the protein where the
residue type is highly conserved across carbohydrate sulfotransferases
and/or involve nonconservative substitutions of nonpolar residues for
polar residue. One possible exception is E71D (family E) but this
occurs as part of a double, homozygous substitution with P72S. It seems
unlikely that E71D on its own would result in absence of protein
function, because it entitles a replacement of an amino acid with
another that is chemically similar. Two mutations occur in the 3′-PB
domain, an essential part of the active site responsible for PAPS
binding. Of these, L200R involves a nonconservative substitution of a
polar for a nonpolar residue at a site that is completely conserved
across carbohydrate sulfotransferases. The A206V substitution, although
representing a conservative change at a position that is not so highly
conserved across carbohydrate sulfotransferases, is nonetheless
predicted to result in a structural change at a highly sensitive region
of the protein and, when modeled on the crystal structure of mouse
estrogen sulfotransferase, result in nonfunctionality due to structural
constraints (data not shown).
It has recently been shown that there is a decrease in the activity of
N-acetylglucosamine-6-sulfotransferase (GlcNAc6ST) in the
cornea of patients with MCD and that this results in the formation of
poorly or nonsulfated KS and causes corneal opacity.
38 More recently, it has been reported that mouse intestinal GlcNac6-O-ST
(mIGn6ST) has the same activity as human corneal GlcNac6-O-ST
(C-GlcNA6ST) suggesting that mIGn6ST is the orthologue of human
C-GlcNA6ST and functions as a sulfotransferase to produce KS in the
cornea. Moreover, the amino acid substitutions in human C-GlcNA6ST
resulting from missense mutations in
CHST6 found in patients
with MCD abolished the sulfotransferase activity by functional
inactivation rather than protein degradation or
mislocalization.
39
The homozygous mutations detected in our patients infers an essential
role of C-GlcNA6ST, the CHST6 protein product, in the
production of normally functioning KS, whereas its inactivation is
responsible for the MCD phenotype. Furthermore, the heterozygous
changes identified affected both alleles in two families, suggesting
that these compound heterozygous mutations could also result in the MCD
phenotype.
We conclude that mutated carbohydrate sulfotransferase could result in
loss of function of the sulfotransferase enzyme required for proper
sulfation of KS, an essential element for corneal transparency, leading
to the MCD phenotype.
The authors thank John Dart, David Gartry, and Frank Larkin for
help with the clinical data; the genetic nurses, Catherin Willis and
Catherine Plant, for samples collection; Nigel Fullwood for advice and
guidance; and all patients who participated in this study.