Synthesis and Secretion of
Corticosteroid-Binding Globulin by Rat Liver: A SOURCE OF
HETEROGENEITY OF HEPATIC
CORTICOSTEROID-BINDERS
Jeffrey N. Weiser, … , Yung-Shun Do, David Feldman
J Clin Invest. 1979;
63(3)
:461-467.
https://doi.org/10.1172/JCI109323
.
Classical glucocorticoid receptors (type II) have a high affinity for synthetic and natural
glucocorticoids. We have previously demonstrated an additional binding site in kidney
cytosol (type III) which has a high affinity for corticosterone but a low affinity for
dexamethasone. In many ways, this binder resembles plasma corticosteroid-binding
globulin (CBG). The first goal of this study was to determine the organ distribution of the
type III binding sites. Cytosol was prepared from isolated cells to avoid plasma
contamination. Of the tissues examined, type III sites were found only in liver and kidney;
sites were absent from thymocytes, IM-9 lymphocytes, adipocytes, and bone cells. The
second goal of this study was to ascertain whether CBG is synthesized in liver and kidney.
Liver and kidney slices were incubated in vitro and the concentration of type III sites was
seen to rise in hepatic cytosol and incubating medium but not kidney. To verify the
impression that liver was synthesizing and secreting CBG, the following experiments were
performed: (a) To demonstrate that type III sites were CBG, steroid-binding profiles and
migration on polyacrylamide gel electrophoresis were shown to be identical for hepatic type
III sites and serum. (b) To indicate that the rise in type III sites was dependent on protein
synthesis, it was shown that cycloheximide blocked the appearance of new type III […]
Find the latest version:
Synthesis and
Secretion
of Corticosteroid-Binding
Globulin by
Rat Liver
A SOURCE OF HETEROGENEITY OF HEPATIC CORTICOSTEROID-BINDERS
JEFFREYN.WEISER,YUNG-SHUNDo,andDAVIDFELDMAN, Department of Medicine, Stanford University SchoolofMedicine, Stanford, California 94305
A B S
T R A C T
Classical glucocorticoid
receptors (type
II)
have a high affinity
for synthetic and natural
gluco-corticoids.
Wehave
previously
demonstrated
anaddi-tional binding
site
inkidney
cytosol
(type
III)
which
has a
high affinity for
corticosterone but
alow
affinity for
dexamethasone.
Inmany ways, this binder
resembles
plasma
corticosteroid-binding globulin (CBG).
The
first
goal of
this
study
wasto
determine the
organ
distribution of the
type
IIIbinding
sites.Cytosol
wasprepared from isolated cells
toavoid
plasma
contamina-tion.
Of the
tissuesexamined,
type
III sites werefound
only
inliver
and
kidney;
siteswereabsent
from
thymo-cytes,
IM-9lymphocytes, adipocytes, and
bone cells.
The second
goal of
this
study
wasto ascertainwhether
CBG
issynthesized
inliver and
kidney.
Liverand
kidney
slices
wereincubated
in vitroand the
con-centration
of
type
III sites was seento rise inhepatic
cytosol
and
incubating
medium but
notkidney.
Toverify
the impression that liver
wassynthesizing
and
secreting CBG, the
following
experiments
wereperformed: (a)
Todemonstrate that type
III sites wereCBG,
steroid-binding
profiles
and migration
onpoly-acrylamide gel
electrophoresis
wereshown
tobe
identical for hepatic
type III sites and serum.
(b)
To
indicate that the
rise intype
III sites wasdependent
on
protein
synthesis,
it wasshown that
cycloheximide
blocked the
appearance
of
newtype
III sites.(c)
Toestablish
that the type
III sites werebeing
secreted,
insitu
liver
perfusion
experiments
showed
time-de-Thisworkwaspresentedinpart atthe National Meeting of
theAmericanFederation for Clinical Research,San Francisco,
Calif., April 1978.
Dr.Weiser was supportedin part by the Howard Hughes
MedicalInstitute StudentResearch Program. This work was
done during Dr. Do's tenure as a postdoctoral fellow
sup-ported byaNational Institutesof Healthtraining grant5T32
AM07217. Dr. Feldman is an Investigator with the Howard
Hughes Medical Institute.
Receivedfor publication 21 July 1978 and in revisedform
12 October 1978.
pendent release of
new sites
intothe
perfusate.
Incon-clusion, liver synthesizes and
secretestype
IIIsites,
afinding previously suspected
but never
proved. The
presence of type III sites in kidney remains to
be
explained.
INTRODUCTION
Heterogeneity of
glucocorticoid-binding
proteins has
been described
ina
variety
of
target tissues (1-4). We
have previously demonstrated
three distinct corticoid
binders
inrat
kidney:
type
I, characterized by high
af-finity
for aldosterone and low affinity for
corticosterone,
was
felt
to
be the mineralocorticoid
receptor (5); type
II, characterized
by
high
affinity
for dexamethasone
and
corticosterone,
wasfelt
torepresent the
glueo-corticoid
receptor
(6); and
type
III, characterized by
high
affinity
for corticosterone but low
affinity
for
dexamethasone
(7), was
notpossible
toclassify
accord-ing to function. The type
IIIbinding site was
recog-nized to be
indistinguishable
inbinding
properties
from plasma
corticosteroid-binding globulin (CBG).1
In
addition
toclassical
glucocorticoid
receptors
(type
II)
several laboratories have described "CBG-like"
binders
in
a
variety
of other
tissues
including
liver
(2,
3), brain
and pituitary (8-9),
muscle
(10),
lung
(11),
and
uterus
(12). We
believe
the
CBG-like
binders in other
tissues and
the type III binder
inkidney may be the
same but will
use the designation type
IIIbinding
inthis paper to
specify
our
experimental
conditions.
Be-cause
the
binding
propertieg
of these cellular binders
are
virtually
identical
to CBG
(7), we
have been
con-cerned
aboutthe
difficulty
indiscriminating between
atrue
intracellular molecule and extracellular fluid
con-tamination
of
the tissues. At least for the kidney, we
IAbbreviations used in this paper: CBG,
corticosteroid-binding globulin; KRB, Krebs-Ringer Bicarbonate; MEM,
minimal essential medium.
were
able
todemonstrate
that the
typeIII
sites wereintracellular with
autoradiographic techniques (13).
The first
aimof this study
was to ascertainthe
organdistribution
of these
type III(CBG-like) binding
sites(i.e.,
sitesexhibiting
ahigh affinity for
corticosteroneand
low affinity for dexamethasone).
Webelieved that
knowing
whether the
binder
waspresentinall
tissuesor
restricted
toonly
selected
targetswould have
im-plications
in aconsideration
of
potential
functions of
this molecule.
Once
it wasdetermined
that,
of the
tis-sues
examined,
the binder
waslimited
toliver and
kidney,
we
turned
ourattention
tothe possibility that
these
organsmight
be
sitesof
biosynthesis
and
secre-tion
of plasma
CBG, thus
explaining
the intracellular
presence
of
type IIIbinders. The liver has
usually
been
considered the
siteof CBG
synthesis
but this fact has
not
been adequately
documented.
Aswill be detailed
in
this
paper,the
presentdata indicate that the liver
does indeed
synthesize
and
secrete type IIIbinders
which presumably
represent CBG. This
finding
provides
a basis
for
the intracellular localization of
type IIIsites inliver.
The
presenceof
type IIIbinding
sitesin
the
kidney
remains to
be
explained.
METHODS
[1,2-3H]Corticosterone (48 Ci/mmol), [1,3,4-3H]dexametha-sone (23 Ci/mmol), and [U-_4C]amino acid mixture (-260 mCi/mmol) were purchased from New England Nuclear (Boston,Mass.). Unlabeledsteroids (A grade) were obtained from Calbiochem-BehringCorp. (San Diego, Calif.). Conven-tionalreagents were ofanalytical gradeand werepurchased from SigmaChemical Co. (St. Louis, Mo.) unless otherwise specified. Sprague-Dawley, 160-200g, malerats (Simonsen Laboratories, Gilroy, Calif.) were adrenalectomized before useand maintainedon 0.9%saline drinkingwaterandPurina chowpellets (RalstonPurinaCo., St. Louis,Mo.)ad libitum. Rats weresacrificed by decapitation.
Preparation of isolated cells. All experiments employed adrenalectomizedrats. Liversandkidneyswereremoved after theviscera wereextensively perfusedviathevena cava with 50mlof iced Ca++-free Krebs-Ringer bicarbonate(KRB) buf-fer (pH 7.4) towashoutblood and then with20mlof Ca++-freeKRBcontaining 1mg/mlofcrudecollagenase (Worthing-tonBiochemicalCorp.,Freehold,N.J.: CLS II).Each kidney was decapsulated, halved, and the papilla excised and dis-carded. The tissue was rinsed, blotted, and sliced (275 ,um thickness) withaMcIlwaintissue
chopper
(5-7).Sliceswereplaced in a KRB solution with 1 mg/ml crude collagenase, 0.2%glucose, and 0.5% bovine serum albumin(GrandIsland Biological Co., Grand Island, N. Y.) and incubated for 60 min at 37°C in avigorously shaking waterbath in a
95:5%-O,:
CO2 atmosphere. Cells werefilteredthrougha250i.m
nylon mesh, centrifugedat 100g, andresuspendedthree times with cold KRB. This yielded isolated hepatocytes and renal tubules with excellentviabilityassessedbytrypan blueexclu-sion.
Adipocyteswereobtained fromepididymalfatpadswhich werediced into a Ca++-free KRB solution containing 3 mg/ml crudecollagenase,3 mM glucose,and 4%BSA in aNalgene flask(NalgeCo.,NalgeneLabware Div., Rochester, N. Y.) (14). The tissue was incubated as above, washed twice withiced
KRB containing 1% bovine serum albumin and once with KRB,and collected by flotation at 200 g.
The thymus was removed, cleaned of connective tissue, and placed in cold 0.9% saline. Pieces of thymus were rinsed throughasteel screen (100 ,um pore size) with iced incubating solution. Thymocytes were centrifuged at 250 g andwashed once in 0.84%NH4Cl to eliminate erythrocytes and twice with incubating solution.
The primary culture of bone cells from fetal rat calvariahas been described (15). Heat-inactivated calf serum was sub-stituted inthe procedure to avoid contamination with CBG. Monolayers were incubated with 1 mg/ml crude collagenase to removeextracellular collagen.Thebone cellswere centri-fugedat 200 gand broughtup twice inincubatingsolution.
Cultured IM-9 human lymphocytes (National Institutes of Health, Bethesda, Md.) were grown in a minimal essential medium (MEM) with Earle's salts (F-17, Grand Island Bio-logical Co.) that included 10% heat-inactivated calfserum. Cellswerecollected from thesuspension at 250 gand washed twice withincubating solution.
Preparation of tissue slices. Livers and(or)kidneys were perfused and slicedasdescribedfor isolated cellpreparation, except theperfusionmedium containedonly iced 0.9% saline. Sliceswerecarefully separated and rinsed severaltimes with acoldKRBsolution.Foreach sample,1gof sliceswasadded to4mlof sterileF-17 MEM(pH 7.4)with25 mMHepes in a Nalgene flask.Allflasks wereincubated withcontinuous 95: 5%-02:CO2at37°Cin ashakingwaterbath.Atthe end ofthe incubation, the flaskswereplaced on ice and the contents were homogenized with 20 strokes ofamotor-driven Teflon (Du Pont Co.,Wilmington, Del.) glass homogenizer. The homoge-nates were centrifugedat50,000 g for40min.The resulting high-speed supernate was stored at -20°C for subsequent assay.
Bindingassays. Type III binding was carried out by in-cubation in the presence of52 nM [3H]corticosterone plus 10-fold excess ofaldosterone and dexamethasone to prevent isotopebinding to type I (mineralocorticoid receptor) and type II (glucocorticoidreceptor) (5-7). Type II binding was car-ried outby incubation in the presence of 26 nM [3H]dexa-methasonewhich binds only minimally to type III sites (6, 7). Inboth systems,nonspecificbinding(<20% of the total) was definedasthatbindingresistant to100-fold unlabeled steroid andwas subtracted from total bindingto give specific bind-ing. Type IIbindingwas performed immediately on fresh tis-sue,whereastype IIIbinding could beperformed on frozen materialwithout loss of bindingactivity. Type IIincubations werecarried out at0°Cfor150 minandtype III at 37°Cfor 30min. Pilot studies verified that the conditionsemployed measuredpeak steady-state levels ofbinding in both assays. In bothsystems,unbound steroidwasremoved bycharcoal ad-sorption. Norite A
5%o
and dextran T70 0.5% were prepared in a bufferof 10 mMTris-HCland 1.5 mM EDTA (pH 7.4). A one-tenthvolume ofthe dextran-coatedcharcoal solution was added tocytosolorserum,vortexed,and allowed to sit for 20 min at 0°C. Charcoal wasremoved by centrifugation at 2,500 g for 10min andthe supernate radioassayed. The conditions of the assay gave similar results whensamples were measuredby SephadexG-50chromatography (PharmaciaFineChemicals, Div. of Pharmacia Inc., Piscataway, N. J.) (7). Protein was measured with Coomasie Brilliant Blue G-250 dye (Bio-Rad Laboratories,Richmond,Calif.)after the method of Bradford (16). Data are usually expressed per milligram of cytosol protein orwhere medium plus cytosol is used, per milligram of supernatant protein.When the starting material was isolated cells, the washed cellswere suspendedin hypotonicbuffer (10mMTris-HCl, 1.5mMEDTA, 10 mM monothioglycerol (pH 7.4), and lysed
by sonication. A high-speed supernate (50,000 g for 40 min) was prepared from the lysate and used forbinding studies. High-speed supernatant fractions from tissue slice experi-mentswerethawedand diluted withhypotonic buff'erto as-say plasma CBG and(or) type III binding by similar pro-cedures.
In whole cell binding studies,aliquots of cell suspensions were first counted electronically with a Coulter counter (Coulter Electronics Inc., Hialeah, Fla.). Cells were then incubated in F-17 MEM with appropriate concentrations of steroi(ds for 30 mnfim at 37°C in a 95:5%-02:CO2 atmosphere whileshakinig.1-mi ali(utiotswerethenremoved, place(din 1.5 ml micro testtubes, anidcentrifuigedat 800g for 1 mIi. After two washes with 1 ml of coldincubating solution, the cells werepouredonto Whatman GF/A filterpaper(Whatman, Inc., Clifton, N. J.) in a filtration manif'oldundervacuum. Filters were rinsed with 10ml ofincubating solution which elimi-nated unboundsteroids,and the dried filter paperwas radio-assayed. "Nonspecific" binding varied between 15 and 20% of total.
Cyclohexinmide
experiments. Theeffectof cycloheximide on the incorporation of amino acids into proteins and the synthesis of CBG binding sites was measured in parallel experiments. After preincubations with various concentra-tions ofcycloheximide for 90 min, aminoacid incorporation with a['4C]amino acidmixture wasdeterminedoverthe sub-se(Iuent120min.TCA (10%final concentration) was added to high-speed supernates of homogenates containing[14C]amino acids.TCAprecipitable protein wascollected at2,500 g for 10minand washed twice with TCA (5% final concentration). Theprecipitate wasdissolved by incubation in0.5 mlofNCS tissue solubilizer (Amersham Corp., Arlington Heights, Ill.) for24h at 50°C andradioassayed.Parallel flasks of slices were treatedsimilarlywithcycloheximideandthetype IIIbindilng
site contentof slices plus medium was measured.
Polyacrylamide gel electrophoresis experiments. Liver slices were incubated and aliquots of cytosol, incubation medium, and serum were labeled for type III sites as de-scribed above. In parallel experiments, unlabeled cortico-sterone was used andthemediumpulsedfor the final 60 min ofa 4-h incubation with 0.5 ,uCi ofa [14C]amino acid
mix-ture. Sanmples were then run on a slab gel electrophoresis apparatus(PharmaciaFineChemicals)at2°Cfor4h at6mA/ sample. Thegelwas7.5%polyacrylamideinabuffer of0.05 M Tris-HCI (pH 8.5). Thegelswere stainedforprotein or sliced into2.5-mm fractions and theradioactivity determined.
RESULTS
Tissue
distribution of
typeIIIbinding: cytoplasmic
assay.
The distribution
of
type III binding sites in
various tissues was examined. The selection of organs
was
based, in part, on our ability to prepare isolated
cells
and,
thus avoid extracellular fluid contamination.
Binding studies
were
performed
on thymocytes,
adipo-cytes,
IM-9lymphocytes, bone cells,
hepatocytes, and
renal tubules.
As
indicated
in
Fig. 1, with the
excep-tion
of
hepatocytes and renal tubules, none of these
tis-sues
exhibited
cytoplasmic type III binders. Type II
binding, was measured as a necessary condition for
type
IIIbinding; the presence ofthe labile type II
bind-ing site implied that the relatively
stable type III binder
had
not
been destroyed during cell preparation or in the
binding
procedures. The number of cytoplasmic type
C z
a
z
m E
z O O
0 ccnC
O-
E0 0
I iv
800
,600,
TYPE600
~~~~~TYPE
;400-200-T
NS NS
n
LIVER KIDNEY
11 E 11i1
BONE ADIPOSE
NS
FIGURE 1 Glucocorticoid binding sites in
cytosol prepatred
from isolatedcells.Isolatedcellswerepreparedasdescribed in NMethods. Cells were sonicated in hypotonic buffer and cytosol obtained. Binding studieswerethen performed with fresh cytosol; type II binding with 26 nM [3H]dexametha-sone and type III binding with 52 nM [3H]corticosterone in the presence of 10-fold unlabeled dexamethasone and aldosterone. Bound and free steroidwere separated by char-coal.Nonspecific binding w%as _23%of total
binding
and was subtractedineach case. Resultsareexpressed
permilligram
cytosolprotein.NSindicates that thedifferencebetween total binding aind nonspecific binding was not significant.Values shown are mean -SE.n = 6for liver and
kidney
andtn =4for the otherorgans.III
sites in renal tubules (212
fmol/mg
cytosol protein)
and
inhepatocytes (638 fmol/mg cytosol protein) was
two
to
three times that of the type II
value
in the same
tissue
prepared with the same experimental procedure.
In
our hands, the concentration of type II receptors is
always
lower after isolated cell preparation than when
cytosol
isdirectly prepared from intact tissues.
Tis.sue
distribution of
tlype
IIIbiinding:
whlole
cell
assay.
Ithas
been
suggested that the
subcellular
loca-tion
of
CBG-like
glucocorticoid binding
may
be
on
the
plasma membrane of pituitary cells (17), or liver cells
(18), or within the nucleus of hepatocytes (19). Because
our
experiments with high-speed supernates assayed
mainly cytoplasmic corticosterone binders, we also
used a
whole-cell binding
technique
with isolated cells
to
include all
subcellular
components as
potential sites
of
high affinity
binding.
Asshown in Fig. 2,
the results
of
whole
cell
binding procedures, like the cytoplasmic
binding
assays, showed type III binding only in kidney
of
the
cell
types examined. Because of
variable
and
methodologically
iniconsistent
restults,
we do not have
confidence
insimilar
experiments
done
with
hepato-cytes
that
gave
exceptionally high values and
adipo-cytes which, because they floated, were difficult to
quantitatively transfer to the filters (data not shown).
Because
the
type III binder was not present in all
tis-sues, we next turned
our
attention to the possibility that
intracellular
CBG-like material in liver and kidney
might be the result of synthesis of plasma CBG by
those organs.
j~TYPE150~~~~~~-
1I CLi
TYPE IIII
z
O
60-z
w
z
0 wU 40-(0
O
20-NS NS NS
-KIDNEY BONE IM-9 THYMUS
FIGURE2 Glucocorticoid binding sites in whole cells. Isolated cellswere
prepared
asdescribed
inMethods.
Bind-ingwasperformed by
the whole celltechnique;
type II bind-ingwith26 nM[3H]dexamethasone andtype IIIbinding
with 52 nM[3H]corticosteroneinthe presence of 10-fold unlabeled aldosterone and dexamethasone.Bound and free steroids
wereseparated by
washing cells on a filter.Nonspecific
binding
was
-20%
of total binding andwas subtractedineach
case.NSindicates thatthedifference between total and
nonspecific
binding was not significant. Values shown are mean+SE. n =4.
Measurement of type
IIIbinding
site
production by
liver and kidney slices.
Liver
and kidney slices were
incubated for
240
min and the total number of type III
binding
sites
was
determined
in
tissue
plus medium
at
hourly intervals.
As
shown
in Fig. 3, the number of
type
IIIbinding
sites per
milligram
of supernatant
0 600l
E /
500- LIVER
400-z I
x300 KIDNEY
> 200
0 60 120 180 240 INCUBATION TIME(minutes)
FIGURE 3 Theeffect of incubationtime on thetotalnumber oftypeIIIbindingsitesin liverandkidney slices. Slicesof liverandkidneywereincubatedat37°Cinaseriesofflasks. At the indicatedtime points,the tissue and medium in agiven flask werehomogenized and the high-speed supemate pre-pared andfrozen. The concentration of totalspecifictypeIII
bindingsitesincytosol plus mediumwasdeterminedbythe charcoalmethod andisexpressed per milligramofsupematant protein. Each value is the mean+SE. n=4 experimentsfor both liver andkidney.
protein was
unchanged
over time in
kidney slices but
linearly
increased
in liver slices from 309+32 fmol/mg
protein at 0 min to 579+86 fmol/mg protein after 240
min of incubation at 37°C. The almost
doubling of
the
total number of
type III sites per
flask appeared
to
indicate that liver was a site
of
type III
production.
Is
the type III binder CBG?
Anti-rat CBG antibody
was not
available
to us
for this study.
In
the absence of
specific antibody,
we
employed
the
steroidal-binding
properties as a means
of
characterizing
the
protein.
The
ability of
a
variety
of steroids
to
compete
for
type
IIIhepatic
[3H]corticosterone
binding sites was
deter-mined and
the
results
compared
to values for
plasma
CBG
binding
sites.
The hepatic samples
were
taken
from the bathing medium after
240-min liver slice
in-cubations which contained -66% newly
synthesized
binding
sites.
The
results
are
shown
inFig.
4.The
order
of affinities for
[3H]corticosterone
binding
sites
inboth
hepatic
and
plasma components was corticosterone
>
cortisol
=progesterone
>aldosterone
>dexa-methasone
>estradiol. Estradiol potentiated the
bind-ing
of
[3H]corticosterone
by
apparently
occupying low
affinity
sites. The
close
correspondence between
plasma
CBG
and liver
slice
high-speed
supernatant
binding
is
evidence that the newly produced binding
sites in
the liver
are
indistinguishable from plasma
CBG
binding
sites.
To
further
substantiate
that
the proteins
themselves
are
identical, polyacrylamide gel electrophoresis
experiments
wereperformed. The data indicate that
plasma
CBG
and
the hepatic protein from both
cytosol
and medium
migrate
inan
identical fashion
(Fig. 5).
The
nature
of the slower
migrating
[3H]corticosterone
binder
inthe
medium
isnot
clear. Taken
with
the
binding data, these results strongly
suggest type III
binding
sites
represent CBG
molecules.
Is
the type
IIIbinder newly synthesized protein?
Liver
slices were treated with increasing
concentra-tions
of
cycloheximide and [14C]amino acid
incorpora-tion into TCA
precipitable
material assessed. As
shown
inFig. 6B,
aconcentration
of
20ug/ml of
cyclo-heximide blocked
-90%
of
protein synthesis. As
shown
in
Fig. 6A,
this
concentration
of cycloheximide blocked
94%
of the
increase
intype
IIIbinding
ascompared
to
untreated control slices.
Lesser concentrations
of
cyclo-heximide
caused comparably
less inhibition of protein
synthesis
and
comparably
less blockage of type
IIIproduction.
The data suggest that the
production
of
type
IIIsites
isdependent
onprotein
synthesis.
Polyacrylamide electrophoresis
data confirm that one
of several
protein
peaks
from medium
and
cytosol
possessing
newly incorporated
[14C]amino
acids
co-migrates with plasma CBG binding sites
labeled
with
[3H]corticosterone
(Fig. 7).
Is
the type
IIIbinder
secreted
by
liver in situ?
In
isolated
insitu liver
perfusion
experiments, type
IIITEstradiol
B
Ato
removeextracellular fluid. It is difficult to reconcile
PLASMA
Estradiol our
findings
with
those
of Amaral
etal.
(19) and
Wertha-)examethasone
.mer and
colleagues (21) who
found, by immunologic
Aldosterone
methods,
aCBG-like protein
inlymphocytes
and
liver
/ D exam
h
etnuclei.
\
1
08Aidosterone
The
realization that the type
III
binder was limited
to certain organs
caused
ustoreconsider
ourprevious
rtisolI position
that the
binder
wasrelated
tohormone
actionrgesterone
\\Cortisol
(7,
13)
and
toexamineother
possibilities.
Because
the
Progesterone
liver and kidney both synthesize many productsand
because the liver
has
long
been
suspected
of
being
the
\
z\ \\
siteof CBG
synthesis,
weexamined
the
hypothesis
that
\tS
m
\ %e
type III intracellular binders might represent CBG
\\ O
\
O
synthesis.
This does
notappeartobe the
casefor
kidney
\\O
\
'@
(Fig. 3).
The following
data indicate that the liver synthesizes
Corticosterone!
and
secretes CBG
and that the type III binders
repre-ne
-
Corticosterone
\
sent
intracellular CBG. First, the type III binders
ap-10
102
0 1 101o2
pearindistinguishable
from CBG. A
comparison
of the
'ETITORCONCENTRATION (molar ratio)
type
IIIsteroid
binding
siteand
plasma
CBG
gives
similar
results (Fig. 4). Scatchard analysis has shown
FIGuRE 4 Acomparison of
the
steroidbinding profile
of type IIIsites inlivercytosol andplasma. (A)Livercytosol after240 min sliceincubation. (B) Plasma obtained from adrenalecto-mized rats. Cytosol or plasma was incubated with 2.6 nM [3H]corticosterone plus increasing concentrations of the in-dicated unlabeled steroid for 30 min at 37°C. Nonspecific bindingwas subtractedfrom allvalues. Totalbindinginthe absence ofcompetitor wastakenas 100%and was2,180 fmol/ mgplasma protein and 127fmol/mg cytosol protein.binding could be found
inincreasing
amounts inthe
perfusate (Table I).
Livers
from
intact rats wereper-fused
in situwith
a KRBbuffer solution
containing
38%
erythrocytes and
3%
bovine
serumalbumin
according
tothe
method of
Mortimore
(20). The
quantity of type
IIIbinding
sites in theperfusate
increased
in
a
roughly linear fashion between
30and
120 min. The
variability
inthe initial
rateof
secre-tion
(0-30
minperiod) reflects
the
fact
that the liver of
rat
2was
better washed
out
before the
perfusion
was
initiated. The
high
initial
levels
are,
therefore, felt
to
represent
washout
of residual plasma
CBG
plus the
secretion rate
of the first
30min.
DISCUSSION
This
work
began
as anattempttoelucidate the
natureof thetypeIIIorCBG-like
binding
siteinkidney.
Theorgandistribution data
(Figs.
1and2)
indicate thattype IIIbinding
is notuniversally
distributed along with
typeII orclassical
glucocorticoid
receptors but rather thatitspresenceisrestricted
tocertainorgans.Ofthose
organstested, only
the liver andkidney possessed
thetype III
binding
site.However,
in ourhands, all
tis-suesappeared
to possessthe
binder
until extensive efforts atwashing
and cellisolation
were undertakenI
0 a
z
4
a LUJ
U)
z
0
U
(0
Ul
LUJ
I-E
Cd) a
z
(0
FRACTION NUMBER
FIGURE 5 Polyacrylamidegel electrophoresisof
[3H]cortico-steronebinders. Liver sliceswereincubated for4hat370C. Attheconclusion of the incubation, aliquotsoflivercytosol and medium as well as serum were incubated with 26 nM
[3H]corticosterone±250-fold unlabeled corticosterone for30 minat370C. Thesesamples werethen subjectedtoslabgel electrophoresison7.5%polyacrylamideinabuffercontaining 0.05 M Tris-HCl(pH 8.5) for4hat2°Cat6mA/sample. The standards and molecular weights are: P, phosphorylase b
(94,000); B, bovine serum albumin (67,000); 0, ovalbumin
(43,000); and C, carbonicanhydrase (30,000).
A
0 L 0 1
80
M.
E E
z
0
z
co
z
60
20
BASAL
zz
2000
O <0
oC z
< E
Z
z K Sw
z
0 120
INCUBATIONTIME(min)
FIGURE6 The effect of cycloheximide on typeIII binding
siteproduction andaminoacid incorporation inliver slices. Slices were preincubated in cycloheximide for 90 min. (A)
Incrementin typeIIIbinding site production. The binding
sitesrepresenttissueplus medium. The basal levelwas214
fmol/mg ofsupernatantprotein. (B) Inhibition ofaminoacid
incorporation into TCA precipitableprotein. Aftera 90-min
preincubation in medium plus or minus cycloheximide, ["4C]amino acids wereadded for thenext 120 min.
that both proteins possess the same affinity for
[3H]-corticosterone
(data
notshown).
Furthermore,
electro-phoresis
onpolyacrylamide gel
showsidentical
mobility of plasma, cytosol,
and medium[3H]cortico-sterone
binders (Fig. 5).
Previous studiesshowed
multiple other
properties in common(7).
Second, the liver
appears tobe synthesizing the
binder. Support for this
argument isfound
instudies
showing
arise intypeIII sitesovertime(Fig.
3).[14C]-Amino
acids
areincorporated
into a proteinthat
co-migrates with serum
labeled
with[3H]corticosterone
(Fig. 7). Inhibition of the
rise in type IIIbinding by
concentrations of
cycloheximide
that inhibit -90% of amino acid incorporationisfurther evidence that the rise inbinding
site content isdependent
on proteinsynthesis.
Third, the liver
secretestypeIIIbinders
invitroand in situ. In vitro,the liver
slicesexhibited
atime-dependent release
oftype III sites into the mediumbathing
theslices.
This processwas alsoinhibited by
cycloheximide.
Our initial studies usinganinsituliverperfusion
system, indicatedaroughly
linear secretion oftype IIIbinding
sites into theperfusing
medium(Table I).
Although the
data indicate that the liverissynthesiz-ing and secreting
CBG,
we cannot rule out thepos-sibility
that intracellular CBGplays
an additionalphysiological
roleinliver. Because thekidney
doesnotFRACTION NUMBER
FIGURE 7 Polyacrylamide gel electrophoresis after
["4C]-amino acid incorporation into hepatic cytosol. Liver slices
were incubated for 4 h at37°C in a medium that included [14C]aminoacid mixture. At the conclusion of theincubation, aliquots ofliver cytosol and medium as well as serum
in-cubated with ['H]corticosterone were subjected to slab gel electrophoresis asdescribed inthelegendin Fig. 5.
TABLE I
Time-Course ofCBG Secretion byInSituPerfused Liver
l3HICorticosteronebindingsites
RatI Rat 2
CuLmulative Incremental Ctimutlative Incremental Time secretion secretion secretion secretion
mill pmol
0-30 112 112 54 54
0-60 153 41 89 35
0-90 184 31 135 46
0-120 236 52 179 44
Intact rats underwent in situ liver perfusion with serum-free perfutsate. Aliquots of perfusate were sampled at the indicated timepoints for [3H]corticosterone binding sites by the charcoal method. The first 30-min period represents some liver washout of plasma CBG in addition to the new secretion. The data is expressed as total picomoles secreted into the perfusion medium per unittime.
466 J. N. Weiser, Y-S. Do, and D. Feldman
BINDING SITE CONTROL PRODUCTION
Cycloheximide
- / /~~~~~5ug/ml
g 20jug/ml _
AMINO ACID INCORPORATION
Cycloheximide
ASpg/ml
,.20 pg/ml
BASALI
0.2
*u
tn
z
z
a: 2 w
Ue
z
0
I
C)
-.
E
C0
0
tn
0 a
C) z
0
z
-1<
0
0.
0
U
z
0
DC
0
z
0
appear
tosynthesize
CBG,
arole for the type III binder
in
that
tissue remains
tobe
determined.
The
studies
documenting
CBG-like binders
inother
organs
(8-12) have,
to avariable
degree, attempted
toexclude extracellular fluid
as a sourceof
thebinding
protein. However, as has become apparent
withbind-ing systems
for other steroid
hormones,
the exclusion
of
plasma
contamination is adifficult
problem.
Alpha-fetoprotein,
animportant estrogen
plasma
binding
pro-tein
inperinatal
rats,
mustbe
carefully
distinguished
from the
trueintracellular
estrogen receptor(22).
Addi-tionally,
it
isnow
believed that the
widespread
oc-currence
of
25-hydroxycholecalciferol binding
protein
in tissues is
the
result of plasma
contamination (23). In
this
later case,
the issue is more
complex,
as it appears
that a tissue and a
plasma
component
interact to
yield
the binding
protein.
ACKNOWLEDGMENTS
We thank Dr. Carl Mondon for performing the perfusion experiments.
This workwas supportedby National Institutesof Health grantAM18156.
REFERENCES
1. Feldman,D., J. W.Funder, andI. S.Edelman. 1972. Sub-cellular mechanisms in the action ofadrenal steroids. Am.J. Med.53:545-560.
2. Koblinsky, M., M. Beato, M. Kalimi, and P. Feigelson. 1972.Glucocorticoid-binding proteins ofratliver
cytosol.
II.Physicalcharacterization andpropertiesof th, binding proteins. J. Biol. Chem. 247: 7897-7904.
3. Litwack,G., R.Filler,S. A.Rosenfield,N.Lichtash,C. A. Wishman,
and
S.
Singer.1973. Liver cytosolcorticosteroid binderII, ahormonereceptor.
J. Biol. Chem. 248: 7481-7486.4. Agrawal, M.K., editor. 1977.Multiple molecularformsof steroidhormone receptors. Elsevier-North Holland,Inc., NewYork.
5. Funder,J. W., D.Feldman,and I. S. Edelman. 1973. The roles of plasma binding and receptor specificity in the mineralocorticoid action ofaldosterone.Endocrinology. 92: 994-1004.
6. Funder, J. W., D. Feldman, and I. S. Edelman. 1973. Glucocorticoid receptorsin the rat kidney: the binding of tritiated-dexamethasone.Endocrinology. 92: 1004-1013. 7. Feldman, D., J. W. Funder, and
I.
S. Edelman. 1973. Evidence fora newclassofcorticosteronereceptorsin rat kidney. Endocrinology. 92: 1429-1441.8. Koch, B., B. Lutz, B. Briaud, and C. Mialhe. 1976. Hetero-geneityofpituitaryglucocorticoid binding: evidence fora transcortin-likecompound.Biochim.Biophys.Acta.444: 497-507.
9. DeKloet,E. R.,andB. S. McEwen. 1976. A putative gluco-corticoidreceptorandatranscortin-like macromolecule in pituitary cytosol. Biochim. Biophys.Acta. 421: 115-123. 10. Mayer,N., N. Kaiser, N. Milholland,and F. Rosen. 1975. Cortisolbindingin ratskeletalmuscle.J. Biol. Chem. 250: 1207-1211.
11. Giannopoulos,G.1976. Acomparativestudy ofreceptors for natural and synthetic glucocorticoids in fetal rabbit
lung.J.
Steroid. Biochem. 7: 553-559.12. Milgrom, E., and E. E. Baulieu. 1970. Progesterone in uterusandplasmaI.Bindinginratuterus 105,000 g super-natant.Endocrinology. 87: 276-287.
13. Strum, J. M., D.Feldman,B.Taggart,D.Marver, and I. S. Edelman. 1975.Autoradiographic localization of cortico-steronereceptors (Type III) tothe collecting tubule ofthe ratkidney. Endocrinology. 97: 505-516.
14. Feldman,D.,andD.Loose. 1977.Glucocorticoid recep-tors inadipose tissue.Endocrinology. 100: 398-404. 15. Chen,T. L., L. Aronow, and D. Feldman. 1977.
Gluco-corticoidreceptorsandinhibition of bone cell growthin primaryculture. Endocrinology. 100: 619-628.
16. Bradford, M. M. 1976. Arapidand sensitivemethodfor the quantitation of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72: 248-254. 17. Koch, B.,B.Lutz-Bucher,B.Briaud,and C. Mialhe. 1977.
Glucocorticoid binding to plasma membranes of the
adenohypophysis.
J. Endocrinol. 73: 399-400.18. Terayama, H., N. Okamura, and T. Suyemitsu. 1976. Epinephrine and corticoid receptors in plasma mem-branes of liver and hepatomas. In Control Mechanismsin Cancer. W. E. Criss, editor. Raven Press, New York. 83-97.
19. Amaral,L., K. Lin, A. J.
Samuels,
andS.Werthamer. 1974. Human nuclear transcortin: itspostulated role ingluco-corticoid
regulation
of genetic activity. Biochim.Biophys.
Acta. 362:332-345.
20. Mortimore, G. E. 1963. Effects of insulin on release of glucose andureaby isolatedratliver.Am.
J.
Physiol.204: 699-704.21. Werthamer,S., A. J.Samuels, andL.Amaral.1973. Identi-fication and partial puriIdenti-fication of "transcortin-like" pro-tein within human lymphocytes. J. Biol. Chem. 248: 6398-6407.
22. Chen,T. L.,andD.Feldman. 1978. Distinctionbetween alpha-fetoprotein and intracellular estrogen receptors: evidence against the presence of estradiol receptorsin rat bone. Endocrinology. 102: 236-244.