| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ARTICLES |
Hoechst AG, Hoechst Marion Roussel (G.M., S.W., A.C.), Research Site Frankfurt, D-65926 Frankfurt am Main, Germany; and Laboratory of Molecular Cardiology (A.K., J.E.), Diabetes Research Institute, Auf’m Hennekamp 65, D-40225 Düsseldorf, Germany
Address all correspondence and requests for reprints to: Dr. Günter Müller, Hoechst AG, Hoechst-Marion-Roussel, Research Site Frankfurt, DG Metabolic Diseases, Building H825, D-65926 Frankfurt a.m., Germany.
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Analytical and metabolic labeling studies indicated the presence of inositol, glucosamine, mannose, galactose, and phosphate in these soluble mediators. Chemical and enzymatic modifications have suggested that they consist of a core structure of phosphoinositol, glycosidically linked to nonacetylated glucosamine, which is itself coupled to additional monosaccharides of varied composition (for a recent review see 7 . Two different mediators were separated by Cheng and co-workers by sizing columns or by ion exchange resins (8). Upon chemical analysis, both mediators were shown to be phosphoinositolglycan (PIG) compounds, as previously reported, but they contained two different inositols and amino sugars.
With the first elucidation of the complete structure of a GPI lipid, the membrane anchor of the variant surface glycoprotein, VSG, from Trypanosoma brucei, structural similarity was recognized between the polar headgroup of a GPI membrane protein anchor consisting of a highly conserved glycan core with myo-inositol, nonacetylated glucosamine, and three mannose residues in characteristic glycosidic linkage (for a review, see 9 , and the PIG substances with insulin-mimetic activity identified so far. This concept gained additional attractiveness by a putative straightforward mode of generation of these molecules by a single lipolytic and/or double lipolytic/proteolytic processing from free GPI lipids and/or GPI membrane protein anchors of the plasma membrane, respectively, which would serve as inactive precursors (for reviews see Refs. 2, 7, 10, 11). In fact, it was known for many years that treatment of insulin-sensitive cells with certain proteases or phospholipases induces some metabolic effects that resemble insulin action to a limited degree. Compatible with these findings was the demonstration that insulin induces release of a set of GPI-anchored plasma membrane proteins from the surface of 3T3 adipocytes (12, 13), isolated rat adipocytes (14), BC3H1-myocytes (15, 16) and hepatoma cells (17) and from membranes of skeletal muscle (18) via activation of a GPI-specific phospholipase and/or protease. Further evidence has accumulated for an involvement of regulated GPI metabolism in metabolic insulin signaling: 1) An antibody raised against a degalactosylated glycan fragment derived from the GPI membrane anchor of VSG from Trypanosoma brucei selectively inhibited some actions of insulin in intact BC3H1 myocytes (19). 2) Mutant K562 erythroleukemia cells that exhibit insulin-stimulated receptor and IRS-1 tyrosine phosphorylation but are totally deficient in GPI synthesis were completely unresponsive to insulin regarding stimulation of glycogen synthesis (20). 3) PIG structures exhibit acute hypoglycemic activity in normal and streptozotocin diabetic rats (21, 22) and can be isolated from normal subjects and in reduced amounts from type II diabetic patients (23).
However, from the beginning the GPI mediator concept was faced with two major problems: 1) The exact structure of the putative GPI cleavage products remained unknown. 2) Systematic investigations on the range of metabolic actions covered by a given chemically defined molecule and on their molecular mechanism of action were lacking. Most studies were performed with impure and chemically heterogenous preparations preventing exact determinations of the concentrations required and were restricted to single or a few selected assays for metabolic insulin actions, only.
These problems may be circumvented by preparation of PIG molecules from a defined source available in sufficient amounts, i.e. by proteolytic/lipolytic digestion in vitro of a single purified GPI-anchored protein. This approach for obtaining chemically defined PIG-peptides (PIG-P), linked via a phosphoethanolamine bridge to one or a few amino acids derived from the carboxyl terminus of the protein moiety, was used in two previous studies. 1) A pronase fragment was prepared from the trypanosomal soluble (PI-specific phospholipase C-degraded) VSG, consisting of the PIG coupled to the carboxyl terminal amino acid of VSG, aspartate (24). This PIG-P specifically and concentration-dependently inhibited isoproterenol-induced lipolysis in isolated rat adipocytes, liver microsomal glucose-6-phosphatase and cytosolic fructose-1,6-bisphosphatase in a cell-free system as well as gluconeogenesis in isolated hepatocytes. However, high concentrations of the fragment (100–200 µM for maximal effects, EC50-values of 62–130 µM) were required and their potency relative to that of insulin was not stated (24). 2) Digestion with PI-specific PLC of the proteinase K-fragment of GPI-anchored human erythrocyte acetylcholinesterase yielded a PIG linked to glycine (25). This PIG-P concentration-dependently antagonized the glucagon-stimulated activation of glycogen phosphorylase in intact rat hepatocytes. However, the cells used displayed limited insulin sensitivity, only, with insulin as well as the PIG-P (10 µM) reducing the glucagon response by 20–30% at maximum (25). In addition, both studies were restricted to the analysis of inhibitory effects of insulin and PIG-P on glucose and lipid metabolism.
Thus, it seemed useful to reevaluate the potential insulin-mimetic activity of PIG-P of known structure with respect to stimulatory effects of insulin on glucose and lipid metabolism. This was also justified by the fact that despite great efforts by numerous laboratories regulation of glucose transport, glycogen and lipid synthesis as well as of lipolysis by insulin via affecting the phosphorylation/dephosphorylation state of key metabolic enzymes remains largely unclear so far, leaving open a putative role of soluble PIG-P in mediating these processes (for a review, see 26 .
We decided to prepare chemically defined PIG-P from a GPI-anchored protein from the yeast Saccharomyces cerevisiae, the glycolipid-modified cAMP-binding ectoprotein, Gce1p, for the following reasons: 1) Albeit Gce1p is a minor plasma membrane protein of yeast (27, 28, 29), it can be isolated and purified in quantities sufficient for subsequent proteolytic and lipolytic cleavages in vitro. 2) The structure of the resulting PIG-P has been elucidated (M. Knauf, S. Petry, W. Frick, K. Sauber, manuscript in preparation). 3) The GPI anchor of Gce1p is susceptible to lipolytic/proteolytic processing in vivo during glucose-induced biosynthesis of the cell wall form of Gce1p (30, 31). PIG-P of similar or identical structure compared with the prepared ones should be generated in vivo, which may have signaling function. 4) A GPI-anchored protein homologous to Gce1p in yeast has been identified recently in cultured (13) and isolated adipocytes (32), which is lipolytically cleaved in response to insulin (13, 14). 4) GPI membrane protein anchors of yeast and higher eucaryotes possess identical glycan cores and phosphoethanolamine linkages to the carboxyl terminal amino acid of the protein moiety (33, 34). The only structural differences came from carbohydrate side chains branching off the common glycan cores. The presented work demonstrates that the chemically defined PIG-P obtained from yeast Gce1p exerts a broad spectrum of potent stimulatory effects on glucose and lipid metabolism in isolated rat adipocytes, cardiomyocytes, and diaphragms and offers a method for preparation of sufficient amounts of PIG molecules, which should facilitate studies on their molecular mechanism of action as well as on their potential use as antidiabetic drug.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
-mannosidases (Aspergillus
phoenicis, Jack bean) and PI-specific PLC (Bacillus
cereus) were provided by Boehringer Mannheim (Mannheim, Germany);
BioGel-P4 and ion exchange resins were from Bio-Rad Laboratories
(Munich, Germany); cAMP Sepharose, octyl/phenyl Sepharose CL-4B,
Sephacryl S-200, Mono Q and Sephadex LH-20 were bought from Pharmacia
Biotech (Freiburg, Germany); Wistar rats were delivered from the animal
breeding station of Hoechst AG (Kastengrund, Germany).
Purification of Gce1p
Gce1p with uncleaved GPI anchor was purified from lactate-grown
yeast cells (strain W303 1A, MATa,
canR1–100, ade2–1, his3–11, 15,
leu2–3, 112, trp1–1, ura3–1) (27). Ten liters of yeast culture
(3 x 108 cells/ml) were combined with 50 ml
suspension of yeast cells (5 x 107 cells/ml) that had
been metabolically labeled with
myo-[3H]inositol (29) to allow monitoring of
the purification of Gce1 protein and of the PIG-P derived from it.
After conversion of the yeast cells into spheroplasts by digestion with
Zymolyase 20,000, plasma membranes were prepared, purified by Ficoll
gradient centrifugation (30), and solubilized in TEB [25
mM Tris/HCl, pH 7.4, 0.5 mM EDTA, 1
mM dithiothreitol (DTT), 20 mM KCl, 0.1
mM phenylmethylsulfonylfluoride (PMSF), soy bean trypsin
inhibitor (10 µg/ml), 10 µM leupeptin, 1 mM
iodoacetamide, 5% glycerol, 0.35% ß-amidotaurocholate (35)] at
2–3 mg protein/ml. A previous study has demonstrated that extraction
of membranes with the detergent ß-amidotaurocholate results in a
considerable enrichment of GPI-anchored proteins toward integral
membrane proteins compared with other detergents (e.g. octyl
glucoside) that are commonly used for solubilization of GPI-anchored
proteins (35). After the initial solubilization of Gce1p (recovery
80–90% of total amount), the following purification steps were
performed with TX-100 (see below). The solubilized plasma membranes
were diluted with 10 volumes of ice-cold TX-114 (2%), 25
mM Tris/HCl (pH 7.4), 150 mM NaCl, the samples
were subjected to phase separation. Gce1p contained in the
detergent-enriched phase (two times reextracted) was purified by gel
filtration chromatography on Sephadex S-300, affinity chromatography on
N6-(2-aminoethyl)-cAMP Sepharose (elution with TEB
containing 0.1 mM cAMP) and phenyl Sepharose chromatography
(elution with the above buffer containing 0.75% TX-100 instead of
ß-amidotaurocholate). Elution from the columns of
[3H]inositol-labeled plasma membrane proteins and of a
purified [14C]inositol-labeled Gce1p marker protein was
followed in parallel by monitoring on line the 3H- and
14C-radioactivity. The partially purified Gce1p was
precipitated with polyethylene glycol 4000 (final concentration 12%).
Analysis of a small sample by SDS-PAGE and fluorography revealed the
presence of a single radiolabeled band at 54 kDa, only, which
comigrated with authentic photoaffinity-labeled Gce1p (for a see
comparison 36 . This demonstrated that Gce1p has been prepared to
radiochemical homogeneity and was the only protein in the preparation
modified with a GPI anchor. According to binding of
[3H]cAMP, Gce1p was enriched 200- to 500-fold (5–10 nmol
cAMP bound per mg protein) compared with total yeast membranes.
Preparation and purification of the PIG-P
Digestion with V8 protease.
The precipitated and washed (two
times with 0.8% polyethylene glycol 4000) Gce1p (400–500 nmol
corresponding to 250,000 to 450,000 dpm) was dialyzed against water,
reduced in volume (Speedvac concentrator) and solubilized in 3 ml of
0.1 M Hepes/KOH (pH 8.0), 20 mM
CaCl2, 0.5 mM DTT, 0.5% TX-100, 30 mg/ml V8
protease (Staphylococcus aureus). After incubation for
10 h at 50 C, a second aliquot of V8 protease and SDS (final
concentration, 1%) was added and the digestion continued for 10
h. The digestion mixture was boiled (15 min), centrifuged (10,000
x g, 10 min) and applied to a Sephacryl S-200 column
(1.5 x 90 cm) equilibrated in 20 mM sodium phosphate
(pH 7.0), 250 mM KCl, 0.1% TX-100, 0.5 mM DTT.
The elution of material during this and the following chromatographic
runs was followed by measurement of 3H-radioactivity
(Ramona Raytest on-line radioactivity monitor), UV absorption
(A220) and insulin-mimetic activity according to
stimulation of lipogenesis for each fraction. The
3H-labeled anchor-containing GPI-peptide had an elution
volume of 0.41 relative to the solvent marker
K2Cr2O7.
Gel filtration.
The peak fractions were pooled, concentrated
to 300 µl, mixed with 1.2 ml of ice-cold formic acid, and applied to
a water-cooled (4 C) Sephadex LH-60 column (1.5 x 65 cm)
equilibrated in ethanol, 88% formic acid (3:1, by vol.).
One-milliliter fractions were collected, immediately neutralized with
NaOH, and measured for 3H-radioactivity. (A control
experiment confirmed that the use of formic acid does not cause
cleavage of inositol-1,2-cyclic phosphate, presumably due to the short
running time [less than 10 min] and low temperature of the
chromatographic procedure. Decyclization of inositol-1,2-cyclic
phosphate as induced by treatment with HCl at high temperature [see
below] could be followed as a shift in the elution profile during SAX
anion exchange chromatography). The major radiolabeled digestion
product had an apparent molecular mass of about 2 kDa relative to
peptide standards. Fractions 83–100 were pooled, supplemented with
NaCl (final concentration, 150 mM) and TX-114 (final
concentration, 2%), incubated on ice (15 min) and centrifuged
(10,000 x g, 10 min, 4 C). The supernatant was removed
and warmed up to 37 C for 10 min to induce phase separation. After
centrifugation (10,000 x g, 5 min, 30 C), the aqueous
phase was removed. The remaining TX-114 phase containing about 85% of
the 3H-radioactivity was washed two times by addition of
the same volume of 100 mM Tris/HCl (pH 7.8), 1
mM EDTA, 0.5 mM DTT and a mix of protease
inhibitors (see above).
Octyl Sepharose chromatography.
The sample was extracted
with an equal volume of water-saturated 1-butanol. The butanol phase
was acidified and adjusted to 5% 1-propanol, 100 mM
ammonium acetate (pH 5.5). After heating (5 min, 100 C) and
centrifugation (10,000 x g, 10 min), the supernatant
was applied (2 ml/h) to a column (2.5 x 60 cm) of octyl Sepharose
CL-4B equilibrated in 5% 1-propanol, 100 mM ammonium
acetate (pH 5.5). The column was washed with 300 ml of equilibration
buffer and eluted with a linear gradient of 5–100% propanol in water
(10 ml/h) over 800 ml. Eight-milliliter fractions were collected and
measured for radioactivity. 3H-radioactivity eluted between
34–38% 1-propanol (as determined by the refractive index). The peak
fractions 71–79 were pooled, concentrated by evaporation,
freeze-dried, and redissolved in 8 ml of 5% 1-propanol. The
purification on octyl-Sepharose was repeated once and the eluted
radiolabeled material was lyophilized.
Digestion with PI-specific PLC.
The sample was dissolved in
200 µl of 20 mM Tris/acetate (pH 7.4), 0.1%
ß-amidotaurocholate, 1.5 U PI-specific PLC (Bacillus
cereus) and incubated for 8 h at 25 C and following addition
of further 5 U of PI-specific PLC for 16 h at 40 C under toluene
atmosphere (head-over rotation). After addition of 0.3 ml water, the
digest was extracted three times with 0.5 ml TX-100 (20%) to remove
toluene. After phase separation, residual toluene was removed from the
aqueous phase under a stream of N2. The aqueous phase was
subjected to TX-114 partitioning as described above. After phase
separation the aqueous phase was passed through a 2.5-ml column of
AG50X12(H+), and then eluted with 1 ml H2O. The
eluate was dried, dissolved in 100 mM ammonium formate (pH
2.8), and loaded onto a 5-ml Dowex 50W-X8 (100 mesh) cation exchange
resin column, equilibrated in 100 mM ammonium formate. The
[3H]inositol-containing PIG-P was eluted with the same
solvent, lyophilized, dissolved in water and passed through a C18
Sep-Pak column.
BioGel-P4 gel filtration.
The sample was chromatographed on
BioGel-P4 (400 mesh) in H2O containing 2 mM
DTT. The column (1.5 cm x 1 m) was held at 55 C with a water
jacket and eluted with H2O containing 2 mM DTT
(12 ml/h) in 1-ml fractions. Fractions 70–75 were pooled, dried, and
dissolved in 25 mM Tris/HCl (pH 8.0). The column was
calibrated in glucose units (Glu) using a glucose oligomer standard mix
(prepared as described for Dionex chromatography).
SAX anion exchange chromatography.
The sample was loaded
onto a SAX HPLC anion exchange column (0.5 x 5 cm) and eluted
with 10 ml of 40 mM triethylamine-formate (pH 4.8),
followed by a linear gradient of 60 ml of 40–400 mM
triethylamine-formate in 0.5-ml fractions. Fractions 93–97 were
pooled, dried, and dissolved in 50% methanol.
HPTLC analysis.
The sample was applied on Si-60 HPTLC plates
that were developed in chloroform, methanol, 30% ammonium hydroxide, 1
M ammonium acetate, H2O (180:140:9:8:23, by
volume) and monitored for radioactivity using a Berthold radioactivity
scanner. The major portion of the radiolabeled material remained at the
origin. It was eluted from the scraped silica by sonication in 50%
methanol, dried, dissolved in water, and applied onto another Si-60
HPTLC plate that was developed with 1-butanol, ethanol, ammonium
hydroxide, H2O (2:2:1:2, by volume). The plates were
scanned for radioactivity. Fractions 21–26 (corresponding a distance
of 21–26 mm from the origin) were scraped from the plate, eluted with
methanol, dried, dissolved in H2O containing 2
mM DTT and used for a second BioGel-P4 chromatography (see
above). Fractions 70–75 were used for analysis by Dionex
chromatography (after chemical and enzymic modification) and of
insulin-mimetic activity.
Determination of the concentration
PIG-P were hydrolyzed (6 M HCl, 16 h, 110 C, in
heat-sealed capillary tube). The amount of inorganic phosphate (2
mol/molecule) was determined. Tyrosine content (1 mol/molecule) was
measured by amino acid analysis after derivatization with
phenylisothiocyanate. Dried PIG-P was suspended in H2O
containing 2 mM DTT at a final concentration of 100
µM.
Chemical and enzymic modification of PIG-P
Nitrous acid deamination and sodium borohydride reduction.
The dried PIG-P (200,000 dpm) was dissolved in 50 µl of 0.1
M sodium acetate (pH 4.0), supplemented with 50 µl 0.5
M NaNO2, and incubated for 3 h at 25 C.
After addition of 20 µl 0.8 M boric acid, the mixture was
adjusted to pH 10.5 by supplementation of 25–40 µl 2 M
NaOH. The ("hot") reduction was initiated by addition of 40 µl
NaB3H4 (36 mM in 0.1 M
NaOH) and after 2 h completed (cold reduction) by addition of 100
µl NaBH4 and further incubation for 3 h. The samples
were supplemented dropwise with 85 µl of acetic acid (50%) and then
desalted by passage through 0.4 ml AG50X12(H+) (elution
with water), dried, further purified to remove boric acid (rotary
evaporation twice with 0.25 ml of acetic acid:methanol [5:95, by
volume], twice with 0.25 ml methanol, twice with 50 µl toluene) and
finally dissolved in water.
Dephosphorylation with hydrogen fluoride (HF).
The dried
sample (deaminated and reduced; 200,000 dpm) was supplemented with 50
µl aqueous HF (48%), incubated for 48 h on ice and then
transferred to 250–300 µl of frozen saturated LiOH, whereby the
final pH has to be adjusted to 3–5. After vortexing, centrifugation
and two washing cycles of the pellet fraction with 50 µl
H2O each, the combined supernatants were desalted (passage
through 0.2 ml AG50X12(H+) over AG3X4(OH-)
over 0.1 ml QAE-Sephadex A25(OH-), elution with 2.5 ml
water), filtered (0.2 µm membrane) and dried (rotary
evaporation).
Partial acetolysis.
The dried sample (100,000 dpm) was
treated with 100 µl pyridine, acetic anhydride (1:1) for 30 min at
100 C, and again dried. After addition of 60 µl acetic anhydride,
acetic acid, sulfuric acid (10:10:1) and incubation for 12 h at 37
C, the reaction was terminated by supplementation of 25 µl pyridine
and 1 ml H2O.
Decyclization.
Dried PIG-P was suspended in 100
mM HCl, incubated at 100 C for 15 min and subsequently
immediately placed on ice. After neutralization with NaOH, the samples
were dried and used for assay of insulin-mimetic activity.
Decyclization was controlled by a shift in the elution profile during
SAX anion exchange chromatography.
Digestion with -mannosidases from Aspergillus phoenicis and
Jack bean.
A quantity of 0.05 mU or 1 U of enzyme, respectively,
was incubated with substrate (50,000 dpm) in 25 µl 0.1 M
sodium acetate (pH 5.0) for 20 h at 37 C. After heating to 95 C
for 10 min the cleavage products were desalted, filtered and dried.
Pronase digestion.
Dried PIG-P was suspended in 100
mM NH4HCO3, 1 mM
CaCl2 and incubated at 37 C two times for 20 h with
1.4 mg pronase/ml each. After addition of 90% formic acid (25 µl/ml)
and centrifugation (4000 x g, 5 min) the resulting
supernatant was filtered (0.2 µm membrane), passed over a Dowex
50W-X8 column equilibrated with 100 mM ammonium formate (pH
2.8) and eluted with the same buffer. After neutralization, the samples
were dried and used for assay of insulin-mimetic activity.
Dionex anion exchange chromatography
The sample was applied on a Dionex CarboPac PA-1 anion exchange
HPLC equilibrated with 0.15 M NaOH/0.15 M NaOH,
0.25 M sodium acetate (95:5) at pH 13. The column was
eluted with 0.15 M NaOH and 0.15 M NaOH, 0.25
M sodium acetate from 95:5 (0 min) to 80:20 (50 min) at 0.6
ml/min. The column was calibrated in Dionex units by inclusion of 5
µl of a glucose oligomer standard mix (20 mg/ml) prepared by partial
acid hydrolysis (0.1 M HCl, 100 C, 5 h) to each run.
The internal standards were detected using a pulsed amperiometric
detector. The 3H-labeled fragments were followed by the
Raytest Ramona on-line radioactivity monitor.
Mono Q anion exchange chromatography
The filtered (0.45 µm membrane) sample in 100 µl ionized
water was injected into an HR5/5 Mono-Q column. The column was eluted
at a flow rate of 1 ml/min with deionized water from 0–1 min, with a
linear gradient to 50% of 350 mM ammonium acetate (pH 5.2)
from 1–8 min and with a linear gradient to 100% of the same buffer
from 8 to 25 min. A220 was followed on-line.
Isolation of ventricular cardiomyocytes
Ca2+-tolerant myocytes were isolated from male
Wistar rats (260–320 g) by perfusion of the heart with collagenase as
previously described (37, 38). The final cell suspension was washed
three times with HEPES buffer (130 mM NaCl, 4.8
mM KCl, 1.2 mM KH2PO4,
25 mM HEPES, 5 mM glucose, 20 g/ml BSA, pH 7.4,
equilibrated with oxygen) and incubated in silicone-treated Erlenmeyer
flasks in a rotating waterbath shaker at 37 C. After 20 min,
CaCl2 and MgSO4 (final concentration 1
mM) were added and incubation was continued until further
use. Cell viability was checked by determination of the percentage of
rod-shaped cells and averaged 90–95% under all incubation
conditions.
Isolation of white adipocytes (according to Ref. 39 with the
following modifications)
Male Wistar rats (140–160 g) were killed by cervical
dislocation. Epididymal fat pads were placed in KRH (25 mM
HEPES free acid, 25 mM HEPES sodium salt, 80 mM
NaCl, 1 mM MgSO4, 2 mM
CaCl2, 6 mM KCl, 1 mM sodium
pyruvate, 0.5% BSA) and then washed with KRB (12 mM
KH2PO4, 1.2 mM MgSO4,
4.8 mM KCl, 25 mM NaHCO3, 120
mM NaCl, 1.4 mM CaCl2, 5
mM glucose and 2.5% (wt/vol) BSA, bubbled with 5%
CO2/95% O2). Each pad was then cut into two to
three pieces. Two pieces each were incubated with 1.5 ml of digestion
buffer (10 mg collagenase in 9 ml KRB) for 15–30 min at 37 C in a
shaking water bath. Released adipocytes were washed two times with KRH
by flotation and then diluted to a final volume equal to 20 ml of KRH/g
of fat (final titer about 2.5 x 105 cells/ml).
Desensitization of adipocytes
Isolated rat adipocytes (106 cells) were incubated
in 20 ml of HEPES-buffered salt solution (40) containing either 5
mM glucose (normal cells) or 20 mM glucose, 16
mM glutamine, 10 nM insulin (desensitized
cells) for 20 h at 37 C under gentle shaking. Thereafter, the
cells were washed three times with HEPES-buffered salt solution lacking
glucose, glutamine, and insulin, and then incubated in 2.5 ml of KRH
(final titer about 3–4 x 105 cells/ml) for 15 min at
37 C to allow deactivation of the adipocyte glucose transport system.
For trypsin treatment, adipocytes in KRB (1.5 x 105
cells/ml) were incubated with 40 µg/ml trypsin for 20 min at 37 C
under gentle shaking. After addition of 0.4 ml bovine trypsin inhibitor
(10 mg/ml) and incubation for 5 min at 20 C, the cells were washed two
times and finally suspended in KRH (final titer see above).
Assays for insulin-mimetic activity
The rates for the basal and insulin-stimulated activities
determined by the individual assays as outlined below are well within
the ranges of published data and are given in the corresponding figure
legends each.
Lipogenesis.
Lipogenesis was measured as incorporation of
[3H]glucose into total toluene-extractable lipids
according to Ref. 41 with the following modifications. Nine hundred
microliters of adipocytes (3.5 x 105 cells) in KRH
containing 140 µM glucose were incubated with insulin or
PIG-P in 10-ml scintillation vials for 20 min at 37 C. Lipogenesis was
started by addition of 100 µl
D-[3-3H]glucose (4 µCi/ml KRH). After
incubation for 60 min at 37 C under an atmosphere of 5%
CO2/95% O2 and gentle shaking, 10 ml
toluene-based scintillation cocktail was added. The samples were
vigorously mixed and counted for radioactivity. dpm values for
radiolabeled lipids (as measured in the toluene phase) were corrected
for a background value determined for an incubation mixture containing
the same amount of [3H]glucose but lacking
adipocytes.
Glycerol-3-phosphate acyltransferase.
GPAT was assayed as
incorporation of [3H]glycerol-3-phosphate into
butanol-extractable lipids according to Refs. 42, 43 with the following
modifications. Two milliliters of adipocytes (6 x 105
cells) were incubated with insulin or PIG-P for 20 min at 37 C in 10-ml
scintillation vials under an atmosphere of 5% CO2 and
shaking (160 cycles/min). Washed cells were homogenized in 500 µl
homogenization medium containing 25 mM Tris/HCl (pH 7.4),
250 mM sucrose, 2 mM EDTA, 2 mM
DTT, 180 mM NaF, 1 mM NaVO3, 1
mM benzamidine, 10 µg/ml aprotinin, 2 µg/ml leupeptin,
0.1 mM PMSF by 10 strokes in a Teflon-in-glass
homogenisator. After centrifugation (1,000 x g, 5 min,
4 C) the infranatant below the fat cake was removed and centrifuged
(14,000 x g, 20 min, 4 C). The supernatant was
recentrifuged (150,000 x g, 60 min, 4 C) to obtain a
crude microsomal fraction. The pellet was suspended in 25
mM Tris/HCl (pH 7.4), 250 mM sucrose, 0.5
mM EDTA, 1 mM DTT, 0.1 mM PMSF at 2
mg protein/ml and stored in liquid N2 until determination
of GPAT activity (43). The assay mixture contained 0.2 mM
[3H]glycerol-3-phosphate (0.5 µCi), 150
µM palmitoyl-CoA and 100 µg microsomal protein in a
total volume of 0.5 ml and was incubated for 3 min at 37 C.
Glucose transport in adipocytes.
Glucose transport was
assayed as uptake of 2-deoxy-[3H]glucose according to
Ref. 44 with the following modifications. One milliliter of adipocytes
in KRH (3 x 105 cells) were incubated with insulin or
PIG-P for 20 min at 37 C. Initial rates of glucose transport were
measured by adding 50 µl KRH containing 0.5 µCi of
2-deoxy-D-[3H]glucose (final substrate
concentration, 0.1 mM) and after 5 min separating 100
µl-aliquots of the cells by centrifugation through 50 µl
dinonylphtalate. Preincubation of the cells with 20 µM
cytochalasin B was used to correct the values for simple diffusion and
nonspecific trapping of the hexose.
Glucose transport in cardiomyocytes.
Transport experiments
were performed at 37 C in HEPES buffer containing 1 mM
MgCl2, 1 mM CaCl2. The reaction was
started by pipetting a 50 µl aliquot of the cell suspension to 50
µl of HEPES buffer containing
3-O-[14C]methyl-D-glucose (final
concentration 100 µM). Carrier-mediated glucose transport
was then determined using a 10-sec assay period and
L-[14C]glucose to correct for simple
diffusion as described earlier (37, 45, 46).
Glucose transporter isoform 4 (GLUT4) translocation.
Adipocytes (7 x 106) were incubated in 100 ml of KRH
containing 5.5 mM glucose with insulin or PIG-P for 20 min
at 37 C under gentle shaking. Washed cells in KRH containing 2
mM KCN were suspended in 55 ml of 20 mM
Tris/HCl (pH 7.4), 250 mM sucrose, 1 mM EDTA,
0.2 mM PMSF, 100 µM benzamidine, 20 µg/ml
leupeptin, pepstatin, aprotinin, and antipain each, and homogenized
10-times in a Teflon-in-glass homogenizer. Plasma membranes and low
density microsomes were prepared as described (47). GLUT4 was
quantitatively immunoprecipitated with rabbit anti-GLUT4 antibodies
raised against the carboxyl-terminal 16 amino acids of rat GLUT4 and
protein A-Sepharose. The immunoprecipitated GLUT4 was separated by
SDS-PAGE and its amount evaluated by immunoblotting with the same
antibodies and [125I]protein A followed by
autoradiography of the dried gel (Kodak X-Omat AR). Quantitation was
performed by phosphorimaging of the dried gel (Molecular Dynamics,
Storm 840).
Glycogenesis.
Glycogenesis was assayed as incorporation of
U-[14C]glucose into ethanol-precipitable glycogen. Intact
hemidiaphragms with rib cage attached were dissected from male Wistar
rats and incubated in DMEM (15 ml/hemidiaphragm) containing 10% FCS,
5.5 mM glucose, 1% BSA, 50 U penicillin/ml, 50 µg
streptomycin sulfate/ml under constant bubbling of
O2:CO2 (95:5) for 60 min at 37 C. Subsequently,
the diaphragms were incubated with insulin or PIG-P in DMEM (15
ml/hemidiaphragm) for 30 min at 37 C in the presence of insulin or
PIG-P. Washed hemidiaphragms were then incubated with 5 µCi
[U-14C]glucose in 15 ml KRH containing 2 mM
glucose for 20 min at 37 C in the presence of insulin or PIG-P. After
blotting and cutting off the rib cage, the hemidiaphragms (80–100 mg
wet weight) were washed, frozen in liquid N2, and
homogenized in 2 ml of 25 mM Tris-HCl (pH 7.4), 150
mM NaF, 5 mM EDTA, 0.1 mM PMSF in
porcelain mortar on ice. One hundred microliters of the supernatant
(200–500 µg protein) obtained by centrifugation (10,000 x
g, 15 min, 4 C) was used for glycogen determination by
precipitation of glycogen with ethanol (43).
Glycogen synthase (GS).
GS activity was assayed by measuring
the incorporation of D-[14C]glucose from
UDP-[14C]glucose into glycogen. Ten milliliters of
adipocytes in KRB (about 3 x 105 cells) were
incubated with insulin or PIG-P in a shaking water bath under an
atmosphere of 5% CO2 for 30 min at 37 C. Washed cells were
frozen in liquid N2 and then homogenized in 0.5 ml of 10
mM N-[Tris(hydroxymethyl)methyl]glycine buffer (pH 7.5)
containing 10 mM EDTA, 150 mM KF, 5
mM DTT by ten strokes in a Teflon-in-glass homogenizer on
ice. The homogenate was centrifuged (5500 x g, 2 min,
10 C). The infranatant below the lipid layer was cleared from residual
lipids by two additional centrifugations and served as source for GS.
The reaction was started by adding 30 µl of the homogenate to 60 µl
of a reaction mixture (prewarmed at 30 C) containing 33 mM
Tris/HCl (pH 7.8), 0.2 mM UDP-[U-14C]glucose
(4 µCi), 6.7 mg glycogen, 150 mM KF and 0.1
mM/10 mM glucose-6-P. After incubation for 20
min at 30 C, the reaction was terminated by addition of 2 ml of 66%
ethanol, 10 mM LiBr (-20 C), rapid mixing and filtration
over prewetted Whatman GF/C glass-fiber discs. The filters were washed
five times with 5 ml of 66% ethanol each at 25 C, dried, and measured
for radioactivity. Blank values determined by adding the homogenate to
tubes containing the complete reaction mixture plus ice-cold
ethanol were subtracted from the total values each. The fractional
velocity was calculated as described (43). Measurements of GS activity
in homogenates from adipocytes that had been incubated in the presence
of 0.1 mM glucose instead of 5 mM glucose (the
concentration routinely used) did not reveal significant differences
with respect to both the activity ratio and the effect of PIG-P or
insulin. Presumably, the dilution of the limited amount of
glucose-6-phosphate, which accumulates during incubation with 5
mM extracellular glucose, during the subsequent preparation
of the homogenate and the GS assay was high enough to prevent
allosteric activation of GS. According to our experience, incubation of
the isolated rat adipocytes in the presence of 5 mM glucose
has a positive impact on their viability, in general, and insulin
sensitivity (glucose transport, glycogen synthesis), in particular.
Lipolysis.
Lipolysis was assayed as release of glycerol. One
milliliter of adipocyte suspension in KRB (3 x 105
cells) was incubated with insulin or PIG-P for 10 min at 37 C under an
atmosphere of 5% CO2. After addition of 1 µM
isoproterenol and 1 unit of adenosine deaminase, the incubation was
continued for 60 min. Following addition of 200 µl of extraction
medium containing 120 mM Hepes/KOH (pH 7.4), 6
mM isobutylmethylxanthine, 60 mM EDTA, 180
mM NaF, the cells were homogenized by 10 strokes of a
Teflon-in-glass homogenizer. The extract was centrifuged (10,000
x g, 15 min, 4 C). The infranatant below the fat cake was
stored in liquid N2 and used for glycerol determination
(43).
Protein kinase A.
PKA was assayed by incorporation of
32P from [32P]ATP into histone H1. Twenty
microliters of the cleared adipocyte extract (see "lipolysis") was
added to 80 µl of a reaction mixture containing 20 mM
3-(N-morpholino)propanesulfonic acid (pH 7.2), 4
mM DTT, 12.5 mM MgCl2, 0.2
mM PMSF, 1 mM isobutylmethylxanthine, 50
µM [32P]ATP (2.5 µCi) with or without 1
µM cAMP. After incubation for 10 min at 30 C, the
reaction mixture was supplemented with 1 ml of ice-cold 20%
trichloroacetic acid, 10 mM ATP, 5 mM
NaPPi, 100 mM NaF. The precipitates were
centrifuged (12,000 x g, 5 min, 4 C) and then
dissolved in 100 µl of 1 N NaOH. After addition of 1 ml
of 20% trichloroacetic acid, 5 mM NaPPi, 10
mM ATP, the precipitates were decanted onto Whatman GF/C
glass fiber filters under vacuum prewetted with washing solution (20%
trichloroacetic acid, 5 mM NaPPi). After
washing (five times with 3 ml each of washing solution), the filters
were dried and counted for radioactivity. Blank values determined by
inclusion of 0.6 mg/ml PKA inhibitor from bovine heart into the
reaction mixture were subtracted from the total values each. The PKA
activity ratio indicating the portion of PKA active at the time-point
of homogenization compared with total enzyme content was expressed as
the ratio between phosphorylation of histone H1 with and without
cAMP.
Miscellaneous procedures
Published methods were used for protein determination (43),
generation of the trifluoroacetic acid (TFA)-fragment from the PIG-P
(48), immunoprecipitation and immunoblotting of GLUT4 (47) and
immunoprecipitation/immunodepletion of the PIG-P with anti-CRD
antibodies (28, 30).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
). It was then subjected
to hydrophobic interaction chromatography on octyl-Sepharose (Fig. 1B).
The radiolabeled GPI anchor peptide eluted as a sharp peak between
34–38% 1-propanol (fractions 71–79). It was subsequently treated
with exogenous PI-specific PLC from Bacillus cereus to
remove diacylglycerol from the anchor. After recovery from the aqueous
phase of TX-114 partitioning, the [3H]inositol-labeled,
polar PIG-P eluted during BioGel-P4 chromatography at a volume
corresponding to 8–8.5 glucose units as the major peak (fractions
70–75). It was further purified by anion-exchange chromatography on a
SAX HPLC column (Fig. 1D). PIG-P (fractions 93–97) constituted the
only peak material with respect to both radiolabel and A220
and was finally subjected to two successive HPTLC runs. The material
migrating as the single radioactive and UV-absorptive spot during the
second HPTLC plate (Fig. 1E) was recovered (fractions 21–26). It was
analyzed by BioGel-P4 chromatography (Fig. 1F) and shown to be
radiochemically and spectroscopically pure. The elution behavior
(fractions 70–75) from the column, calibrated with a glucose oligomer
standard mix, with a retention time of 8.5 glucose units corresponding
to about 1.5 to 1.6 kDa is compatible with the apparent molecular mass
of a polar fragment derived from Gce1p by digestion with pronase and
bacterial PI-specific PLC (29).
|
). This material presumably consists of proteolytic
degradation fragments (peptides, amino acids) from Gce1p itself and
from proteins which are not modified with a GPI anchor (see above) and
contaminate the Gce1p preparation as well as of lipids associated with
the GPI anchor of Gce1p. Previous studies demonstrated that Gce1p can
be isolated from yeast plasma membranes in TX-100 insoluble complexes
(49), which in mammalian cells harbor GPI lipids, glycosphingolipids
and cholesterol, in addition (for a review, see 50 .
|
). The relative increases in the
percentage area of peak material eluting at 2.92–2.98 min (as
indicated by an arrow) out of total material from the
corresponding step (C0 [5.8%]->D [31.4%] 5.4-fold;
D
E [63.2%], 1.7-fold; EF [97.9%], 1.6-fold) correlated well
with the relative enrichments of the insulin-mimetic activity
calculated for the total material from the corresponding purification
steps with respect to A220 (C0D, 8.8-fold;
DE, 1.8-fold; EF, 1.8-fold; see Table 1
). In addition, the peak
materials eluting at 2.92–2.98 min contained more than 95% of the
[3H]inositol radiolabel of the total materials from step
C0, D, E and F, each (data not shown). These findings
strongly suggest that the peak material (at 2.92–2.98 min) contains
the insulin-mimetic activity. Thus, since A220 is a very
sensitive parameter for detection of organic as well as of anorganic
substances, the material obtained after the final purification step (F)
is assumed to be chemically homogenous.
|
F (about 17%) can be explained by
a large amount of (phosphorylated) inositol moieties contaminating the
PIG-P at stage C0 that are generated by lipolytic digestion
from phosphoinositol lipids copurifying with detergent-solubilized
Gce1p. The enrichment of GPI-anchored proteins and phospholipids in
detergent-resistant complexes is well documented (49, 50). This
(phosphorylated) inositol-containing material obviously did not exhibit
significant insulin-mimetic activity (see 3.3-fold enrichment with
regard to the amount of [3H]inositol from
C0F). Furthermore, after immunodepletion with anti-CRD
antibodies, only a minor portion of insulin-mimetic activity was left
in the fractions C0 (12.8%) F (3.4%) each, whereas
the immunodepleted fraction C0 still contained about 80%
of total [3H]inositol (data not shown). This argues for
the absence of inositol-1,2-cyclic phosphate in the phosphorylated
inositol-containing material contaminating the PIG-P throughout the
purification steps C0F. The decline in the residual
insulin-mimetic activity left after immunodepletion during step
C0F reveals the removal of insulin-mimetic activity that
is not due to the PIG-P structure and thereby confirms the efficiency
of the purification procedure. At the end, more than 96% of the total
activity contained in fraction F could be immunoprecipitated with
anti-CRD antibodies, i.e. the vast majority of active
molecules harbors a terminal inositol-1,2-cyclic phosphate residue. Structural analysis of the material from step F revealed the presence of components, only, typically for the polar head group of yeast GPI anchors and of the carboxyl terminal tripeptide of Gce1p, Tyr-Cys-Asn (determined by sequence determination using manual Edman degradation) that are covalently linked (data not shown, manuscript in preparation). Determination of free amino groups in the intact PIG-P demonstrated the presence of two such groups per mol of myo-inositol, one belonging to the nonacetylated glucosamine, the other one to the tyrosine residue. Other amino acids (determined after total acid hydrolysis by the ninhydrin method) were not found in the PIG-P preparation. This provides strong evidence that the PIG-P is derived from a single GPI-anchored protein, Gce1p, and demonstrates its chemical homogeneity. Furthermore, the GPI anchor of Gce1p appears to be very closely related to the trypanosomal or vertebrate GPI anchors because, after cleavage by PI-specific PLC, it is recognized by antibodies against the cross-reacting determinant of soluble VSG from Trypanosoma brucei. This suggests a structural arrangement of the glycan core constituents of Gce1p in a fashion similar or identical to that of VSG.
The PIG-P obtained after the final purification step F were
analyzed by Dionex anion exchange chromatography. For this, they were
deaminated by nitrous acid treatment and then reduced with radiolabeled
sodium borohydride for replacement of (radiolabeled)
inositol-1,2-cyclic phosphate with radiolabeled 2,5-anhydromanitol
(AHM). The material was then subjected to aqueous HF dephosphorylation
for removal of the terminal phosphoethanolamine moiety. Subsequently,
three different chemical or enzymic treatments were performed to
specifically cleave off mannose residues before analysis of the size of
the degradation fragments radiolabeled at the terminal AHM by Dionex
chromatography (Fig. 3). The radiolabeled HF-fragment
was eluted at 3.2 Dionex units corresponding to the neutral glycan core
with a single carbohydrate side chain branching off,
Man1–2Man1–2Man1–6Man1–4AHM (Fig. 3A). It thus constitutes the
glycan portion of 80% of the whole spectrum of GPI anchors made by
yeast cells (34). The lack of terminal phosphoethanolamine allowed
access of mannose-specific exoglycosidases to this structure.
-mannosidase (Aspergillus phoenicis) specific for
-D-Man(1, 2)D-Man glycosidic bonds released
a radiolabeled fragment from the deaminated, reduced and
dephosphorylated PIG-P eluting at 2.2 Dionex units compatible with the
structure, Man1–6Man1–4AHM (Fig. 3B). -mannosidase (Jack bean),
specific for any terminal, nonsubstituted, -D-Man,
generated a radiolabeled fragment from the neutral glycan core eluting
at 1.0 Dionex units corresponding to a structure completely devoid of
the mannose residues, i.e. AHM (Fig. 3C). Finally, partial
acetolysis, which is relatively specific for Man()1–6Man glycosidic
bonds (but also cleaves Man()1–4AHM to a minor degree), produced a
radiolabeled structure from the neutral glycan core eluting at 1.1
Dionex units consistent with Man1–4AHM as well as a small amount of
free radiolabeled AHM (Fig. 3D). Taken together, the size analysis of
the exoglycosidase digestion and partial acetolysis products of the
deaminated, reduced and dephosphorylated PIG-P by Dionex anion exchange
chromatography is in full agreement with a conserved glycan core
structure of the headgroup of GPI anchors common for protozoa and
mammals, Man1–2Man1–6Man1–4GlcN1–6inositol-1-phosphate, which is
modified with a single Man1–2 branch at the distal mannose residue
typical for the 80%-component of total yeast GPI anchors (34). In
addition, this structure was confirmed by 1) NMR spectroscopy of the
complete as well as deaminated/dephosphorylated PIG-P (Knauf, M.,
unpublished results); 2) GC/MS analysis of the chemically or
enzymatically treated PIG-P (Petry, S., unpublished results); and 3) GC
analysis of the complete GPI anchor peptide generated by pronase
digestion of Gce1p (29). In the latter study, the purified GPI anchor
peptide was dephosphorylated with HF, the oligosaccharides hydrolyzed
with trifluoroacetic acid, the sugars reduced with NaBH4
and then acetylated, and the final products separated and identified by
GC and amino acid analysis. Using these methods, the stoichiometry of
the individual components was determined as follows: ethanolamine: Man:
GlcN: inositol = 1.2: 3.9: 1: 1.6.
|
). This amount was sufficient for performing a complete
concentrations-response curve for a single insulin-mimetic activity
with adipocytes (each point with five independent incubations). The
assays with diaphragms required larger fermentations of yeast on
industrial scale (up to 1000 liters). Gce1p is a low abundance protein
with respect to total membrane proteins compared with certain mammalian
or protozoal GPI-anchored proteins. However, the possibility of large
scale production of a GPI-anchored protein and PIG-P derived thereof by
yeast is of great advantage, which obviously represents a prerequisite
for a systematic study of the insulin-mimetic activity of PIG
structures by performing a multitude of different assays based on
various cellular systems over a broad range of concentrations. Previous
reports using PIG-P from human erythrocyte acetylcholinesterase and
trypanosomal VSG were clearly limited in this respect (Refs. 24, 25;
see also first section of this article).
|
). In normal adipocytes,
lipogenesis (performed at low glucose) was stimulated in a
concentration-dependent manner by both insulin (up to 22-fold;
EC50 = 0.5 nM) and PIG-P (up to 15.5-fold;
EC50 = 1 µM). In adipocytes preincubated in
primary culture in the presence of high concentrations of glucose,
glutamine and insulin to induce desensitization of the cells toward
insulin (40, 47), the insulin stimulation of lipogenesis was reduced to
about 40% of the normal response (10-fold at maximum). In contrast,
the maximal responsiveness of the desensitized adipocytes toward PIG-P
was significantly less diminished (to about 80% of the normal
response) with no apparent change of the EC50 values. In
fact, in desensitized adipocytes, the maximal activation of lipogenesis
was higher for PIG-P compared with that for insulin. In adipocytes
pretreated with a limited concentration of trypsin to cleave off the
insulin binding -subunit of the insulin receptor, insulin completely
failed to increase lipogenesis. There was no concomitant impairment of
cell viability as confirmed by stimulation of lipogenesis by 1
mM NaVO3 to about 90% of the maximal insulin
effect (data not shown). PIG-P activated lipogenesis to about 80%
compared with that of untreated cells with EC50-values
remaining at 1–2 µM. These data provided first hints
that 1) PIG-P exerts certain insulin-mimetic effects on isolated rat
adipocytes; 2) PIG-P may interfere with the insulin signaling cascade
at a site downstream of the putative defect, which is responsible for
the desensitization; and 3) PIG-P does not bind to the insulin
receptor.
|
) was activated by insulin in a concentration-dependent
manner up to 2.5-fold (EC50 = 0.5 nM) when
assayed with a crude microsomal fraction that had been prepared from
incubated normal adipocytes. PIG-P increased GPAT activity in a
concentration-dependent manner up to 1.8-fold (EC50 = 3–5
µM) which accounted for 75–85% of the maximal insulin
effect. Additional experiments demonstrated that the GPAT activity
associated with a crude mitochondrial fraction (14,000 x
g-pellet, see Materials and Methods), which in
contrast to its microsomal counterpart cannot be inactivated by
N-ethylmaleimide, was not affected at all by incubation of
the adipocytes with either PIG-P or insulin (data not shown). This is
in agreement with a previous report on BC3H-1 myocytes (42)
that the microsomal N-ethylmaleimide-sensitive form of GPAT,
but not the mitochondrial N-ethylmaleimide-insensitive one,
is under insulin control. Glucose transport (Fig. 6) was
increased up to 15-fold after incubation of normal adipocytes with both
insulin (EC50 = 0.3 nM) and PIG-P
(EC50 = 1–1.5 µM). In desensitized
adipocytes, glucose transport activation by insulin was drastically
impaired with respect to both maximal responsiveness (5-fold) and
sensitivity (EC50 = 3–4 nM). In contrast,
PIG-P stimulated glucose transport to about the same maximal response
with slightly elevated EC50 values (2–2.5
µM). In consequence, in desensitized adipocytes the
maximal glucose transport activity elicited by PIG-P (at 20
µM) exceeded that by insulin about 3-fold. The
differential desensitization of the glucose transport system for
stimulation by insulin and PIG-P is well correlated to that of
lipogenesis (see above). This reflects the fact that at low glucose
concentrations (<0.1 mM) in the incubation medium glucose
transport is rate-limiting for lipid synthesis in isolated
adipocytes.
|
|
). The EC50-values for
activation of glucose transport and GLUT4 translocation by PIG-P were
comparable (2–3 µM), whereas the maximal stimulation
factor for translocation (about 5-fold) is considerably lower than that
for transport (Fig. 7B). Presumably, this reflects the limited
responsiveness of the GLUT4 translocation assay due to contamination of
the isolated plasma membranes with low density microsomes as suggested
by a similar discrepancy for insulin (about 6-fold stimulation of
translocation, see Fig. 7). However, an increase of the intrinsic
glucose transport activity of GLUT4 by PIG-P in addition to stimulation
of GLUT4 translocation cannot be excluded so far.
|
). Control
experiments with equivalent volumes of digestion mixtures lacking PIG-P
added to adipocytes before incubation with insulin demonstrated that
the dilution of the incubation mixtures was sufficient and the test
period short enough to avoid significant interference with insulin
signaling (data not shown). Thus, a putative reduction of PIG-P
activity in stimulating lipogenesis and glucose transport would most
likely be the consequence of the loss of certain structural
determinants of the PIG-P. Treatment with nitrous acid at pH 3.5 and
subsequent reduction leads to deamination of the nonacetylated
glucosamine separating phosphoinositol from the glycan-peptide portion
with its terminal AHM. Treatment with aqueous HF causes
dephosphorylation, separating the inositolglycan portion from the
phosphate and ethanolamine-peptide moieties. Treatment with mild acid
cleaves the myo-inositol-1,2-cyclic phosphodiester,
resulting in PIG-P with terminal inositol-1(2)-monophosphate. Treatment
with pronase removes the two terminal amino acids, Tyr, Cys, from the
peptide portion resulting in phosphoinositol-glycan-Asn. Treatment with
TFA cleaves between the nonacetylated glucosamine and the proximal
mannose residue separating phosphoinositol-glucosamine from the
glycan-peptide portion with the mannose residues attached. In
comparison to the untreated PIG-P, each treatment caused a considerable
impairment of the maximal activation (by 10 µM PIG-P) of
lipogenesis and glucose transport (Fig. 8). Most efficient were nitrous
acid deamination, aqueous HF dephosphorylation, and TFA cleavage with
reductions to almost basal values, compatible with cleavage of the
PIG-P within its glycan core structure. Opening of the terminal cyclic
phosphate moiety resulted in a somewhat lower, but still significant,
deactivation to 2- to 3-fold above basal values. Removal of the two
amino terminal amino acids from the tripeptide decreased the activity
to 50–60% of that of the untreated PIG-P (see Discussion).
Immunodepletion of the PIG-P preparation with anti-CRD antibodies
revealed that about 70% of the total insulin-mimetic activity (with
respect to stimulation of both glucose transport and lipogenesis) could
be removed from the particular lot of PIG-P (used for the experiments
of Fig. 8). In that lot, the portion of PIG-P molecules harboring the
intact inositol-1,2-cyclic phosphate moiety was about 30% of total
material, only (due to prolonged storage at pH 6.5) as determined by
quantitative and calibrated immunoblotting with anti-CRD antibodies.
This means that the specific insulin-mimetic activity of PIG-P with
inositol-1,2-cyclic phosphate is about 3.7-fold higher than that of
PIG-P with decyclized inositol-monophosphate attached (the remaining
portion contained in that lot of PIG-P). This value is in reasonable
agreement with the 5.4-fold higher specific activity of the
inositol-1,2-cyclic phosphate containing PIG-P as calculated for the
PIG-P preparation of the purification scheme (Table 1; 11% of total
PIG-P molecules did not cross-react with anti-CRD antiserum and
contained 3.4% of total insulin-mimetic activity). We conclude that
the tripeptide and inositol-cyclic-phosphate moieties at the termini of
the PIG-P are not absolutely required for but significantly enhance the
insulin-mimetic activity of PIG-P.
|
, insulin produced a 3.3-fold
stimulation of 3-O-methylglucose transport using maximally
effective concentrations of the hormone. Parallel incubations with
PIG-P (2 µM) showed that they produced a 2.3-fold
increase in glucose transport corresponding to 70% of insulin
action.
|
). The higher
EC50-value of PIG-P (8–10 µM) required for
the isolated diaphragms compared with adipocytes may be muscle-specific
or based on a diffusion barrier for the PIG-P to reach (via
the endothelium) the individual muscle fibers of the diaphragm. The
even more pronounced discrepany of the EC50-values for
insulin between diaphragm and adipocytes as well as the similarity of
the EC50-values for PIG-P between cardiomyocytes (see
above) and adipocytes favor the latter possibility. The key metabolic
enzyme for glycogen synthesis is GS, which is activated by insulin both
in an allosteric (via increase of glucose-6-phosphate) and
covalent (via dephosphorylation) manner. At low
glucose-6-phosphate concentrations (<0.1 mM), the
dephosphoform of GS will be active. GS activity was increased in a
concentration-dependent manner upon incubation of intact diaphragms
with PIG-P or insulin and subsequent assay of a homogenate fraction in
the presence of 0.1 or 10 mM glucose-6-phosphate (Fig. 11). The increments of the maximal fractional velocity,
i.e. the portion of active dephosphoform of GS at the time
point of homogenization, were comparable between PIG-P (2.8-fold
vs. basal) and insulin (2.6-fold vs. basal). The
corresponding EC50-values for PIG-P (8 µM)
and insulin (10–20 nM) were in good agreement with those
for glycogenesis (see above).
|
|
). Upon incubation of adipocytes, insulin and PIG-P
reduced in a concentration-dependent manner the rate of
isoproterenol-induced lipolysis to almost basal levels and to about
30% above basal, respectively. Compared with the other insulin-mimetic
effects of PIG-P on isolated rat adipocytes (see above), the
concentration-response curve was shifted to the left (EC50
= 0.5–0.8 µM). The increase of sensitivity of fat cells
for the antilipolytic activity of PIG-P corresponds well to the high
sensitivity toward insulin (EC50 = 0.05–0.07
nM), which may be taken as evidence for (in part) similar
molecular mechanisms used by PIG-P and insulin for inhibition of
hormone-sensitive lipase.
|
). Incubation of rat adipocytes with PIG-P before
treatment with isoproterenol led to a concentration-dependent decline
of the PKA activity ratio from the maximal isoproterenol-induced value
(set at 1) to 0.25–0.30, which corresponded to about 70% of the
maximal insulin effect. The EC50 value of PIG-P for
inhibition of PKA (0.3–0.5 µM) was in good correlation
to that for inhibition of lipolysis (see Fig. 12). This suggests that
the major mechanism for inhibition of agonist-stimulated lipolysis by
PIG-P relies on the reduction of the active portion of PKA, resulting
in a lower phosphorylation and activity state of hormone-sensitive
lipase rather than in its direct inhibition. Additional experiments are
required to discriminate between a decline of cytosolic cAMP levels
and/or an inhibition of the intrinsic activity of PKA independent of
the cAMP levels induced by PIG-P to lower the PKA activity ratio.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The PIG-P used in the present study, which has been generated from a yeast GPI-anchored protein by successive proteolytic and lipolytic cleavage, is chemically homogenous and of defined structure. It potently mimics a broad spectrum of metabolic insulin actions (i.e. stimulation of glucose transport, GLUT4 translocation, lipogenesis, GPAT, glycogenesis, GS, and inhibition of isoproterenol-induced lipolysis, PKA) in isolated fat and muscle (cardiomyocytes, diaphragm) cells. Perhaps the most challenging insulin-mimetic activity of the PIG-P represents the stimulation of glucose transport in isolated rat adipocytes, diaphragms and cardiomyocytes, because for PIG structures derived from free glycolipids controversial results have been published. PIG purified from rat liver does not modulate glucose transport in isolated rat adipocytes (52), but PIG-like molecules (although structurally ill-defined) from other sources, such as calf plasma (53) or rat adipocytes (54), activate glucose transport in this system. These discrepancies may rely on the biological source of PIG and thus ultimately on structural differences. Recent preliminary studies on the structure-activity relationship of the PIG(-P) used in the present report hint into this direction (Frick, W., A. Bauer, M. Bauer, S. Wied, G. Müller, manuscript in preparation). A PIG-P with a single asparagine residue left as peptide portion, which had been prepared from yeast Gce1p as described here but using pronase instead of V8 protease, or chemically synthesized PIG structures harboring the complete glycan portion of the isolated PIG-P but lacking the terminal phosphoethanolamine-peptide portion stimulate/inhibit the key enzymes of glucose metabolism (i.e. GS, GPAT, PKA) almost as efficiently as the authentic PIG-P but, in contrast to those, show very moderate or even no effect on glucose transport in isolated rat adipocytes.
The potency of the PIG-P with respect to both maximal activity and
EC50 values and its availability in sufficient quantity
should enable detailed studies on its target at the molecular level in
normal and desensitized adipocytes as well as in muscle cells. There
are several possibilities for the molecular mechanism how PIG-P can
exert pleiotropic effects on glucose transport and metabolism in
insulin-sensitive cells: 1) binding of the PIG-P (perhaps
via different structural determinants) to distinct cell
surface receptors (which certainly do not include the insulin receptor
[see Fig. 4] but may be identical with receptor proteins for
lipolytically cleaved GPI-anchored proteins identified in plasma
membranes of isolated rat adipocytes (32)); 2) after transport across
the plasma membrane [which has been described for PIG molecules from
rat liver after incubation with rat hepatocytes (55)], binding of
PIG-P to distinct intracellular target proteins, which are coupled to
the glucose transport system (GLUT4 translocation machinery) and the
individual key metabolic enzymes (GS, GPAT, hormone-sensitive lipase);
3) induction of a signaling cascade that branches to GLUT4 and the
metabolic enzymes. In case of the latter mechanism and in analogy to
insulin signal transduction, tyrosine phosphorylation of insulin
receptor substrate-1/2 (IRS-1/2) may be the primary target of PIG-P
because this seems to be prerequisite for most if not all insulin
actions (for a review, see 56 . In contrast,
phosphatidylinositol-3'kinase (PI-3'kinase), which is located
immediately downstream of IRS-1/2 in the insulin signaling cascade (56)
and has to be stimulated by insulin for glucose transport activation to
occur (57), is apparently involved in some but not all metabolic
insulin actions (58). In consequence, a putative direct activation of
PI-3'kinase by PIG-P (which does not involve tyrosine phosphorylation
of IRS-1/2) may not be sufficient to induce all the observed
insulin-mimetic effects. Currently, we are studying the effect of PIG-P
on IRS-1 and PI-3'kinase and the participation of PI-3'kinase in the
insulin-mimetic activity of PIG-P (Eckel, J., and G. Müller,
manuscript in preparation). Thus, knowledge of the molecular mode of
PIG-P action may also provide insights into the regulation of lipid and
glycogen synthesis by insulin because PIG-P and insulin may share
part(s) of their signal transduction mechanisms.
The PIG-P we are using is derived from a GPI-anchored protein and harbors a phosphoethanol-peptide moiety in addition to the phosphoinositolglycan part. This molecule seems to be almost as potent as insulin in stimulating glucose transport and metabolism and significantly more effective than PIG structures derived from free glycolipids (for which insulin-dependent production in insulin-sensitive cells has been reported; for a review see 7 . However, its unusual structure raises some questions on the physiological relevance because the (insulin-induced) biosynthesis would require double processing of a GPI-anchored protein. In fact, this has been speculated from previous findings that insulin causes lipolytic cleavage of some GPI-anchored proteins in cultured fat and muscle cells (see first section of this article). We confirmed this for Gce1 (which is a GPI-anchored protein homologous to yeast Gce1p) in 3T3 adipocytes (13). In addition, we demonstrated a rapid loss of myo-inositol-phosphate from the lipolytically cleaved Gce1, indicating a secondary processing event within the residual glycan portion of the anchor or the carboxyl terminus of Gce1. However, so far there is no direct evidence for the (insulin-dependent) release of PIG-P from Gce1 (or other GPI-anchored proteins) in insulin-sensitive cells.
|
Acknowledgments |
|---|
Received January 30, 1997.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  P. McLean, S. Kunjara, A. L. Greenbaum, K. Gumaa, J. Lopez-Prados, M. Martin-Lomas, and T. W. Rademacher Reciprocal Control of Pyruvate Dehydrogenase Kinase and Phosphatase by Inositol Phosphoglycans: DYNAMIC STATE SET BY "PUSH-PULL" SYSTEM J. Biol. Chem., November 28, 2008; 283(48): 33428 - 33436. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Chalasani, R. Vuppalanchi, N. S. Raikwar, and M. A. Deeg Glycosylphosphatidylinositol-Specific Phospholipase D in Nonalcoholic Fatty Liver Disease: A Preliminary Study J. Clin. Endocrinol. Metab., June 1, 2006; 91(6): 2279 - 2285. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. R. F. Nascimento, L. M. A. Lessa, M. R. Kerntopf, C. M. Sousa, R. S. Alves, M. G. R. Queiroz, J. Price, D. B. Heimark, J. Larner, X. Du, et al. Inositols prevent and reverse endothelial dysfunction in diabetic rat and rabbit vasculature metabolically and by scavenging superoxide PNAS, January 3, 2006; 103(1): 218 - 223. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Anderwald, G. Muller, G. Koca, C. Furnsinn, W. Waldhausl, and M. Roden Short-Term Leptin-Dependent Inhibition of Hepatic Gluconeogenesis Is Mediated by Insulin Receptor Substrate-2 Mol. Endocrinol., July 1, 2002; 16(7): 1612 - 1628. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. K. Edens, L. A. Reaves, M. S. Bergana, I. L. Reyzer, P. O'Mara, J. H. Baxter, and M. K. Snowden Yeast Extract Stimulates Glucose Metabolism and Inhibits Lipolysis in Rat Adipocytes in Vitro J. Nutr., June 1, 2002; 132(6): 1141 - 1148. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. Müller, S. Wied, and W. Frick Cross Talk of pp125FAK and pp59Lyn Non-Receptor Tyrosine Kinases to Insulin-Mimetic Signaling in Adipocytes Mol. Cell. Biol., July 1, 2000; 20(13): 4708 - 4723. [Abstract] [Full Text]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
