| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
From the The Burnham Institute,*
La Jolla, California;
the University Department of Paediatric
Gastroenterology,
Royal Free and University
College School of Medicine, London, United Kingdom; the Institute of
Child Health,
University College, London,
United Kingdom; and St. Mark’s Hospital,§
Harrow, United Kingdom
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1,3-glucosyltransferase needed for the addition of the first
glucose to the dolichol-linked oligosaccharide. The maternal
mutation, C998T, causing an A333V substitution,
has been shown to cause CDG-Ic, whereas the two paternal
mutations, T391C (Y131H) and C924A (S308R) have not previously
been reported. The mutations were tested for their ability to rescue
faulty N-linked glycosylation of carboxypeptidase Y in
an ALG6-deficient Saccharomyces
cerevisiae strain. Normal human ALG6 rescues
glycosylation and A333V partially rescues, whereas the combined
paternal mutations (Y131H and S308R) are ineffective.
Underglycosylation resulting from each of these mutations is much more
severe in rapidly dividing yeast. Similarly, incomplete protein
glycosylation in the patient is most severe in rapidly dividing
enterocytes during gastroenteritis-induced stress. Incomplete
N-linked glycosylation of an HS core protein and/or
other biosynthetic enzymes may explain the selective localized loss of
HS and PLE.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
1,3-glucosyltransferase, cause
CDG-Ic.6-8
This enzyme is required for the addition of
the first of three glucose residues to LLO, and without the first
glucose, further glucosylation is prevented. The nonglucosylated
precursor oligosaccharide is a poor substrate for the
oligosaccharyltransferase complex and is inefficiently transferred to
proteins.9,10
Furthermore, the absence of glucose on the
oligosaccharides will affect the quality control and folding of the
proteins.11
PMM-deficient patients have severe
psychomotor and mental retardation, peripheral neuropathy, cerebellar
hypoplasia, and pericardial effusions among other
symptoms.1,12
They often have feeding problems and failure
to thrive.13
On the other hand, PMI-deficient patients
have normal intelligence, achieve developmental milestones, and do not
show neuropathy, cerebellar hypoplasia, or pericardial effusions.
Instead, they present with hypoglycemia, liver fibrosis, and
protein-losing enteropathy (PLE).14-16
The
1,3-glucosyltransferase-deficient patients (CDG-Ic) have a milder
CDG-Ia-like appearance.6,8
The deficiency has previously
been shown to result from a homozygous point mutation in the
ALG6 gene leading to an A333V change in the
protein.6,7
Here we describe a patient with this and two
other ALG6 mutations that severely compromise the
transferase function. The patient was initially seen for PLE, and an
intestinal biopsy showed a complete absence of the normal heparan
sulfate (HS) on the basolateral surface of
enterocytes.17,18
However, it improved after PLE subsided.
The ALG6 mutations seem to compromise N-linked
glycosylation most severely in rapidly dividing cells.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
S. cerevisiae cells were grown in standard yeast peptone dextrose and synthetic complete media.19 E. coli was grown in Luria-Bertani medium.20 Restriction enzymes and T4 DNA ligase were from Promega (Madison, WI). Oligonucleotides were from Genbase (San Diego, CA). Sequencing was performed using BigDye sequencing kit on an ABI 377 DNA sequencer, both from Applied Biosystems (Foster City, CA).
Enzyme Assays
The lysate from fibroblasts were centrifuged at 10,000 x g x 5 minutes and the supernatant used to assay activities of PMM (EC 5.4.28) and PMI (EC 5.3.1.8) as described21 and modified.22,23 The Dol-P-Glc synthase (EC 2.4.1.177) assay was previously described.24 GDP-mannose pyrophosphorylase (GDPMP, EC 2.7.7.13) assay has been described.25
Analysis of Serum Transferrin
Isoelectric focusing (IEF) analysis of transferrin was done using sera from the patient, defined cases of CDG-1a, CDG-Ib, and normal controls were as described previously.26
Analysis of the Genes
Total RNA was extracted using the RNAeasy kit (Qiagen, Valencia, CA). First-strand cDNA of the human ALG6 was synthesized at 45°C for 40 minutes using Superscript II reverse transcriptase (Life Technologies, Inc., Rockville, MD) and primer oVW118: 5' AAT GGT AAT TTC ATT TAT ACA TAG C 3'. The cDNA was used as template in a polymerase chain reaction with primers oVW117: 5' AAG AAG TGA TTG ACC ACG TTT 3' and oVW118 using the following cycle: 94°C for 2 minutes; 20x (94°C for 20 seconds, 48°C for 30 seconds, 70°C for 1 minute; 20x (94°C 20 for seconds, 48°C for 30 seconds, 70°C for 2 minutes); 70°C for 10 minutes. The 1700-bp fragment was used as template for the sequencing reaction and for further amplification with primers oVW132: 5' TTG CGC ACA GAA TTC CCC TCC CTA AAT 3' and oVW133: 5' GAC TGG AAC CTC GAG GAA ACA ATT TGT TTA GG 3' that introduced the restriction sites EcoRI and XhoI at each end of the ALG6 gene. Both the amplified cDNA and the recloned ALG6 fragments were sequenced essentially as described previously26 using primers oVW117 or 132, oVW 118 or 133, oVW126: 5' GCT GAT CTG CTG ATT 3', oVW105: 5' CCT AGG GTC ACT GGC, oVW106: 5' GTG TCA CTA CCA GTC 3', oVW108: 5' GAC TGG TAG TGA CAC, oVW109: 5' GCC AGT GAC CCT AGG 3', and oVW143: 5' GGG CAT TAG GAG TTC 3'.
Strains and Plasmids
The ALG6-deficient S. cerevisiae strain
YG227 (Mat ade2-101
ura3-52 his3
200 lys2-801
alg6::HIS3)27
was used
to express the human 1,3-glucosyltransferase gene ALG6.
Subcloning and transformation of E. coli and yeast were
performed using standard procedures.20
The expression
vector pWE85 was constructed by subcloning a
NotI-EcoRI fragment containing the strong
constitutive NOP1 promoter (a gift from S. Emr, University
of California, San Diego, CA) into the same sites of
pRS426.28
pWE137 and pWE126 encoding the maternal A333V
and the paternal Y131H, S308R 1,3-glucosyltransferase, respectively,
were constructed by digesting the polymerase chain reaction products
obtained from above with EcoRI and XhoI and
cloning the fragment into the same sites of pWE85. To separate the two
paternal mutations, pWE126 was digested with AvrII and
XhoI, and the fragment containing the DNA encoding the S308R
mutation was recloned into the same sites of pWE187 carrying the
wild-type human ALG6 gene. This resulted in pWE355. The
Y131H-encoding plasmid was constructed by digesting pWE126 with
EcoRI and AvrII and cloning of the fragment into
the same sites of pWE187, resulting in plasmid pWE215. All plasmids
were resequenced before transformation into the
ALG6-deficient S. cerevisiae strain
YG227.27
In Vivo Glycosylation of Carboxypeptidase Y (CPY)
S. cerevisiae strain YG227 containing one of the
following plasmids: pWE85 (without the ALG6 gene), pWE187
(containing wild-type human ALG6 cDNA), pWE134
(encoding A333V 1,3-glucosyltransferase), pWE126 (encoding Y131H,
S308R 1,3-glucosyltransferase), pWE215 (encoding Y131H
1,3-glucosyltransferase), and pWE355 (encoding S308R
1,3-glucosyltransferase), was used to investigate the effects
of the mutations in ALG6 on the in vivo
glycosylation of CPY. The transformed yeast cells were grown in SC
medium without uracil19
overnight at 30°C. The next day
the cells were diluted to an optical density (OD) at 600 nm of 0.05 and
followed spectrophotometrically. Cells were harvested at either
exponential phase (OD600 = 0.3 to 0.5) or early
stationary phase (OD600 
2.0) by
centrifugation. One OD600 unit of cells was
suspended in 10 µl of phosphate-buffered saline (PBS) with protease
inhibitors (Complete, Mini Ethylenediaminetetraacetic acid-free;
Boehringer Mannheim, Mannheim, Germany) and 10 µl of 2x sodium
dodecyl sulfate-loading buffer.20
After heating for 10
minutes at 100°C, the samples were loaded onto 10% sodium dodecyl
sulfate-polyacrylamide gels followed by Western blot using rabbit
antiserum against CPY (a generous gift from Jakob Winther, Carlsberg
Laboratory, Copenhagen, Denmark).
Immunohistochemical and Histochemical Analysis
Endoscopic biopsies of the small intestine were processed in formalin or snap-frozen in liquid nitrogen. Mucosal inflammation was assessed on frozen sections using monoclonal antibodies (mAbs) (DAKO, Cambridgeshire, UK) against CD3, CD4, CD8, CD19, CD25, and HLA-DR. Epithelial and basement membrane composition analysis used mAbs against laminin and collagen IV, (DAKO) and tenascin (Sigma, Gillingham, UK) using dilutions previously determined on sections of tonsil, spleen, small intestine, and colon. Bound antibody was localized by peroxidase or alkaline-phosphatase immunohistochemistry. Epithelial glycosaminoglycans (GAG) distribution was done on formalin-fixed tissue using antibody mAb 10E4 (Seikagaku, Abingdon, UK) that recognizes HS chains, and visualized with avidin-biotin (Vectastain Elite, Vector, Peterborough, UK). Histochemical detection of sulfated GAGs using cationic colloidal gold in PBS at pH 1.2 (polylysine gold, 1/100; Biocell Int., Cardiff, UK) together with a silver enhancer, as previously reported.29-31 Specific enzymatic digestions confirm that this technique detects HS in epithelial and endothelial cells, and chondroitin and dermatan sulfates within the mucosal tissues.29,31 Immunostaining of syndecan-1, the major HS core protein, was done as previously described,17 using mAb MCA681H (Serotec, Kidlington, UK) specific for human syndecan-1.32,33
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The patient was seen at age 3 months for profound PLE after gastroenteritis. The usual features seen in PLE (such as lymphangiectasia) were not found, and endoscopic small bowel biopsy showed only minimal abnormality with slight crypt elongation. The colonoscopy was normal. Proteinuria was absent. The postenteritis PLE normalized within a month, but during the next year, he had several episodes of life-threatening PLE after acute gastroenteritis, requiring large amounts of intravenous-infused albumin. He was cortisol-deficient (<50 nmol/L) and the episodes settled with cortisol replacement. At 14 months he had the first of several tonic-clonic fits, with a normal electroencephalogram.
At 9 months he showed prolonged blood coagulation time, Factor XI deficiency, and marked delay in motor development including peripheral hypotonia, but maintained normal reflexes and muscle bulk. Bayley scales showed global developmental delay to 5 months. He had little visual attention, no nystagmus, and visual evoked potentials were normal bilaterally, whereas magnetic resonance imaging scan showed delayed myelination.
The combination of periodic PLE and Factor XI deficiency prompted
transferrin IEF analysis (Figure 1)
, which showed an abnormal pattern
typical of CDG-I.34
Attempted therapy with oral mannose
gave no clinical improvement, and no significant increase in Factor XI
activity. He continues to show marked developmental delay, hypotonia,
and remains hypocortisolic, with severe Factor XI, antithrombin
III, heparin cofactor II, Protein C, and Protein S deficiencies.
Occasional convulsions occur, but repeated electroencephalogram
has been normal.
|
Small bowel biopsies were performed on two occasions, the first
during PLE and the second while the patient was well. Histological
assessment was mostly normal showing only slight villous blunting with
no crypt hyperplasia or inflammatory enteropathy. Lamina propria and
intraepithelial lymphocytes, CD3, CD4, CD9, CD19, and CD25 populations
were within normal limits on both samples (data not shown). The
epithelium during PLE was HLA-DR-negative (Figure 2A)
. Basement membrane laminin, collagen
IV, and tenascin distributions were normal (data not shown), but the
distribution of HS was abnormal, especially during acute PLE.
|
. A small amount of punctate, intraepithelial
staining was found, mostly above the nucleus (arrow in Figure 2B
). This
abnormal pattern and punctate staining (arrows in Figure 2, D and F
) is
seen using a mAb (10E4) specific for HS chains (Figure 2, D–G)
and
another that recognizes the major HS core protein in enterocytes,
syndecan-I17,18
(Figure 2, I and J)
. Sulfated GAGs and
immunoreactive HS were normal within the lamina propria and on vascular
endothelium in both biopsies. The subepithelial basement membrane
seemed intact. The second biopsy, taken while the patient was well
(Figure 2H)
, showed similar findings in only 11 of 16 identifiable
villi, whereas 5 of 16 villi showed a more normal, but slightly fainter
basolateral-staining pattern. This extended throughout the villus
in 2 of 5 and was patchy in 3 of 5. Villi with more normal
basolateral staining had considerably fewer HS-positive
intracellular aggregates. In contrast, both biopsies showed normal
basolateral HS staining in the patient’s colon (Figure 2K)
, stomach,
or esophagus (data not shown), suggesting that the pathology was
restricted to the small intestine. Analysis of Fibroblasts
Fibroblasts from the patient were analyzed to identify the
specific CDG defect. PMM, PMI, and GDP-Man pyrophosphorylase activities
were normal (Table 1)
. LLOs were isolated
from the patient’s and the control’s fibroblasts labeled with
[2-3H]mannose, and the size of the sugar chains
was analyzed by amine adsorption high pressure liquid chromatography
(Figure 3)
. The oligosaccharide from
control LLOs gave the expected pattern with the major peak co-eluting
with
Glc3Man9GlcNAc2.
Sugar chains from the patient were smaller and co-eluted with a
nonglucosylated sugar chain derived from LLO glycan made by a S.
cerevisiae strain with a nonfunctional ALG6
gene.27
This gene codes for an 1,3-glucosyltransferase
which is required for the addition of the first glucose residue
during LLO biosynthesis. The patient’s oligosaccharide was completely
sensitive to -mannosidase digestion, whereas digestion of the
corresponding oligosaccharide from control cells produced free
mannose and another product,
Glc3Man4GlcNAc2.
The smaller size of the patient’s oligosaccharide and its sensitivity
to -mannosidase digestion suggested that the first glucose residue
was not added during LLO biosynthesis. This could result from
deficiencies in either Dol-P-glucose synthesis or
1,3-glucosyltransferase, the defective enzyme in
CDG-Ic.6-8
Direct assay of Dol-P-glucose synthase
activity in the patient’s fibroblasts showed normal levels compared to
controls and to other CDG patients with known defects in other
genes26
(data not shown), suggesting a defective
1,3-glucosyltransferase.
|
|
Using reverse transcriptase-polymerase chain reaction we analyzed
the sequence of the ALG6 cDNA from the patient, his
asymptomatic parents, and several controls. The controls had the
expected published sequence for human ALG6, but the patient
had three base substitutions in the coding region of the gene. A C 
T at position 998 on the cDNA level (C998T) changes A V at position
333 of the Alg6 protein (A333V), the T391C mutation results in Y131H,
and the C924A mutation results in S308R (Figure 4)
. The first mutation is present in the
maternal allele (Figure 4
, top) and the second two in the paternal
allele (Figure 4
, middle). Previously, a homozygous C998T substitution
was reported in CDG-Ic patients6,7
indicating that this
homozygous substitution alone can be responsible for the phenotypes
seen in CDG-Ic.
|
ALG6-deficient S. cerevisiae strains
inefficiently glycosylate their glycoproteins.27
For
example, the vacuolar carboxypeptidase CPY has four N-linked
chains, but in alg6 strains, it lacks one or
two of these chains. Full glycosylation can be restored by
transformation with a plasmid carrying either the yeast or human
ALG6.6,7
We examined the functional
consequences of each mutation by testing its ability to restore normal
glycosylation to CPY. Expressing the various alleles from a multicopy
plasmid under a strong constitutive promoter showed that normal human
ALG6 completely restores CPY glycosylation in the
alg6 strain in the exponential growth phase
(generation time = 2.6 ± 0.3 hours) (Figure 5
, + ALG6). In contrast, the
maternal allele, encoding the A333V Alg6 protein, only partially
restores normal glycosylation because CPY glycoforms lacking both 1 and
2 oligosaccharide chains are seen. The paternal ALG6 allele
containing two substitutions resulting in Y131H and S308R cannot
correct CPY glycosylation, and the glycosylation pattern is similar to
yeast without Alg6p function (Figure 5
, -ALG6) where the
majority of CPY lacks oligosaccharide chains. To determine which of
these paternal mutations was more critical for activity, each was
tested separately for CPY glycosylation rescue. Expression of
ALG6 with only the Y131H substitution seems to have a
similar severity to A333V (Figure 5
, top). Some fully glycosylated CPY
is seen, but the majority of CPY still lacks 1 and 2 oligosaccharides.
S308R Alg6p alone weakly restores CPY glycosylation (Figure 5
, top).
Endoglycosidase H digestion to release all of the sugar chains,
produces only one band indicating that the size differences in
immunoprecipitated CPY are based on the number of oligosaccharide
chains (data not shown). These results suggest that this patient has a
more substantial loss of ALG6 activity than previously
described patients homozygous for A333V6
and for the
recently described S478P,7
which could explain the severe
clinical presentation.
|
, top
and bottom panels, respectively). We interpret this to mean that the
impaired glycosylation capacity is overwhelmed in fast-growing cells,
but tolerated in slow growing yeast cells, where the
glycosylation-demands are less. |
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Patients with several forms of CDG that result in protein underglycosylation often have gastrointestinal pathologies. For instance, PMI-deficient CDG-Ib patients have severe diarrhea, vomiting, and PLE.14-16 Small intestine biopsies show partial villus atrophy and the enterocyte endoplasmic reticulum is distended and engorged with insoluble precipitated proteins associated with the abundant chaperone BiP.14 Oral mannose therapy effectively relieves nearly all of the symptoms of these patients.14,40 PMM-deficient CDG-Ia patients also show a failure to thrive and many have gastrointestinal abnormalities. Duodenal biopsies show shortened villi, increased inflammatory cells in the stroma, and dilated smooth endoplasmic reticulum.13 Patients with CDG-Ie, caused by a defect in the synthesis of dolichol-P-mannose, have persistent reflux and poor weight gain despite caloric supplementation.26 It is likely that the high synthetic and growth demands placed on intestinal epithelial cells make them especially vulnerable to glycosylation-based pathologies.
We are uncertain about the basis of localized HS loss in this patient, but it is seen only in the small intestine, not in the more slowly turning over epithelial cells of the stomach, colon, and esophagus.41,42 Altered glycosylation may directly affect the synthesis and stability of the core protein, syndecan-1, or it may be indirect because many of the HS biosynthetic enzymes in the Golgi are most likely N-linked glycosylated. Intracellular accumulation of the small amount of residual HS-immunoreactive material in the nuclear region, perhaps the Golgi, suggests that the sulfated chains do not reach the surface. The syndecan-1 protein has two potential N-linked glycosylation sites (NFS at position 43 and NQS at position 231), and both are conserved between rat and human. The first potential N-linked glycosylation site is strongly conserved between mouse, hamster, rat, and human. All of the syndecan-1 proteins examined so far contain at least one potential N-linked glycosylation site, indicating that N-linked glycosylation is important. Underglycosylation may cause misfolding leading to degradation of the majority of core protein in the endoplasmic reticulum, whereas the remainder is inefficiently targeted. Thus, underglycosylation could directly or indirectly affect the interaction of HS with extracellular ligands involved in receptor-signaling complexes and other cell surface interactions.43
Severe PLE in children usually results from structural lesions where epithelial HS is preserved (eg, lymphangiectasia)44 or inflammatory enteropathy (eg, celiac disease, Crohn’s disease). The molecular basis is unknown, although albumin loss from inflammatory-based degradation of epithelial HS has been proposed as a pathogenetic mechanism in Crohn’s disease.30 The involvement of HS in PLE was also seen in a study of three infants with normal small bowel biopsies, who had massive PLE from birth and loss of HS from the basolateral enterocyte membrane.31 HS loss from the glomerular basement membrane leads to albumin leakage across the vascular endothelium in kidney45-48 and thus enterocyte HS deficiency may be analogous to congenital nephrotic syndrome.47 The findings here suggest that CDG-1c causes a similar loss of HS from the enterocyte basolateral membrane. Loss of syndecan-1 is also seen in reparative cells from patients with inflammatory bowel disease, and this is thought to compromise their ability to bind basic fibroblast growth factor.49 The beneficial effects of exogenous heparin in promoting healing may result from its ability to substitute for the missing HS chains and restore high-affinity growth factor binding and proliferation. In addition, heparin has also been reported to reduce PLE related to Fontan cardiac surgery in children.50 Therapy aimed at increasing epithelial HS expression may be clinically useful to prevent chronic intestinal protein loss.
The first biopsy of the patient taken during PLE showed a complete
absence of basolateral HS staining, whereas the subsequent biopsy taken
while well showed substantial basolateral HS in 
1 of 3 villi.
Intracellular HS only occurred in those villi where basolateral
staining was absent. It is striking that this child’s several episodes
of severe and prolonged life-threatening PLE all followed acute
gastroenteritis, where crypt cell proliferation increases to repair
damaged epithelium. We speculate that the already inefficient
glycosylation is overwhelmed by the increased epithelial turnover in
gastroenteritis, effectively leading to complete loss of HS in the
small intestinal epithelium. Failure to detect any HS abnormalities in
the patient’s colon, stomach, and esophagus because of slower cellular
turnover is supported by a twofold to threefold slower turnover in
these organs in rats.41,42
It is also consistent with our
finding that the ALG6 mutations impact CPY glycosylation
most severely in rapidly dividing yeast. This pattern of episodic
severe PLE, often associated with recurrent infections, rather than a
consistent gradual protein-loss, has been seen in other children with
CDG-I14
(and our unpublished observations). Unexplained
PLE is a life-threatening complication of many multisystemic
disorders. It is possible that enteric protein loss seen in some
individuals results from environmental stresses that exceed their
glycosylation capacity.
|
Acknowledgements |
|---|
|
Footnotes |
|---|
Supported by March of Dimes Grant FY99-205 (to H. H. F.), and by fellowship 990059/20 from the Carlsberg Foundation (to V. W.).
Present address of S. K. is Korea Basic Science Institute, Taejon 305-333, South Korea.
Accepted for publication August 29, 2000.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  A. M. Ostrow, H. Freeze, and J. Rychik Protein-Losing Enteropathy After Fontan Operation: Investigations Into Possible Pathophysiologic Mechanisms Ann. Thorac. Surg., August 1, 2006; 82(2): 695 - 700. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Bode, S. Murch, and H. H. Freeze Heparan Sulfate Plays a Central Role in a Dynamic in Vitro Model of Protein-losing Enteropathy J. Biol. Chem., March 24, 2006; 281(12): 7809 - 7815. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Bode, E. A. Eklund, S. Murch, and H. H. Freeze Heparan sulfate depletion amplifies TNF-{alpha}-induced protein leakage in an in vitro model of protein-losing enteropathy Am J Physiol Gastrointest Liver Physiol, May 1, 2005; 288(5): G1015 - G1023. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 V. Westphal, S. Kjaergaard, E. Schollen, K. Martens, S. Grunewald, M. Schwartz, G. Matthijs, and H. H. Freeze A frequent mild mutation in ALG6 may exacerbate the clinical severity of patients with congenital disorder of glycosylation Ia (CDG-Ia) caused by phosphomannomutase deficiency Hum. Mol. Genet., March 1, 2002; 11(5): 599 - 604. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. H. Freeze Update and perspectives on congenital disorders of glycosylation Glycobiology, December 1, 2001; 11(12): 129R - 143R. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
