Mechanisms of glucocorticoid type II receptor regulation and activation in brain


Material Information

Mechanisms of glucocorticoid type II receptor regulation and activation in brain
Physical Description:
234 leaves : ill. ; 29 cm.
Densmore, Charles L., 1954-
Publication Date:


Subjects / Keywords:
Glucocorticoids -- physiology   ( mesh )
Receptors, Glucocorticoid -- isolation & purification   ( mesh )
Brain -- drug effects   ( mesh )
Neuroscience thesis Ph.D   ( mesh )
Dissertations, Academic -- Neuroscience -- UF   ( mesh )
bibliography   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph.D.)--University of Florida, 1987.
Bibliography: leaves 202-232.
Statement of Responsibility:
by Charles L. Densmore.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000904993
oclc - 17945246
notis - AEL4006
sobekcm - AA00006107_00001
System ID:

Full Text








I greatfully acknowledge the interest, support and patience of my

supervisory committee including Drs. Kimon Angelides, Steven Childers,

Robert Cohen, and Kathleen Shiverick. I am especially thankful for the

never-tiring, unselfish support of my advisor and friend, Dr. William G.


I am particularly indebted to Seyed M. Emadian and Yun-Chia (Jenny)

Chou who have helped me in more ways than they realize.

I would also like to thank Mary Rupp for always seeing to it that

my experiments ran smoothly and Kimmy Bloom for the painstaking artwork.

Finally, this work would never have been possible without the

constant love and support of my wife Anita.



ACKNOWLEDGEMENTS..................................................... ii

ABSTRACT .................... ................................... v


I GENERAL INTRODUCTION..................................... 1


Introduction........................... ........ .......... 13
Materials and Methods .................................... 15
Results.................................................. 19
Discussion.............................................. 43

BINDING CAPACITY.......................................... 52

Introduction............................................ 52
Materials and Methods................................... 62
Results................ ............. ... .................. 66
Discussion.............................................. 108

GLUCOCORTICOID RECEPTORS.................................. 127

Introduction............................................ 127
Materials and Methods................................... 143
Results.................................................. 148
Discussion.............................................. 161


GLUCOCORTICOID RECEPTOR.................................... 171

Introduction........................................... 171
Materials and Methods................................... 177
Results................................ ...... .. 180
Discussion ............................................... 193

VI CONCLUDING REMARKS......................................... 196

BIBLIOGRAPHY..................................................... 202

BIOGRAPHICAL SKETCH............................................ 233

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



Charles L. Densmore

August 1987

Chairman: William G. Luttge
Major Department: Neuroscience

Glucocorticoids have been shown to have profound effects on the

mammalian nervous system, most of which have been shown to result from a

receptor-mediated change in concentration or activity of key proteins.

This change is a consequence of an action of the glucocorticoid-

receptor complex on the molecular machinery that controls gene

expression. Upon binding of a steroid by the receptor, it undergoes a

temperature-dependent activation, correlated with changes in a number of

receptor properties, before it can effectively interact with its nuclear

acceptor sites.

The surface hydrophobicity of the glucocorticoid receptor in brain

cytosol was examined using hydrophobic interaction chromatography.

There was a clear increase in hydrophobicity upon activation, a finding

with important implications for the interactions between activated

receptors and numerous nuclear components.

The successful purification of the glucocorticoid receptor allowed

for an investigation of activation relatively free of cytosolic factors

that could interfere with, or contribute to, the process. Although

activation of unpurified receptors led to a 50-fold increase in

DNA-cellulose binding, there was no increase when purified receptors

from brain or liver were subjected to identical activation conditions.

However, other changes in purified receptors associated with activation,

including an increase in surface hydrophobicity, implied the possibility

of a multistep process.

Unoccupied glucocorticoid receptors can exist in a non-steroid-

binding or down-regulated state which can be up-regulated to a

steroid-binding state upon addition of sulfhydryl reducing reagents. An

investigation of the hydrodynamic and biochemical nature of the up- and

down-regulated receptors revealed that they exhibited the same

sedimentation properties, but differed in surface hydrophobicity. In

addition, the up-, but not the down-, regulated form proved susceptible

to irreversible inactivation by some sulfhydryl reactive reagents,

allowing for a determination of the respective populations of the two

forms in cytosol or homogenate.

Sedimentation of the unoccupied receptor was examined by post-

labeling procedures and was found to differ in the presence (9-10 S

only) or absence (9-10 & 4-5 S) of molybdate. This effect of molybdate

was mimicked somewhat by heat-stable cytosolic factors.




Glucocorticoid steroid hormones are known to have profound

metabolic, neuroendocrine and behavioral effects in the mammalian brain.

Glucocorticoids, in addition to regulating ACTH secretion (for review

see Keller-Wood and Dallman, 1984), produce alterations in mood, changes

in detection and recognition of sensory stimuli, changes in sleep and in

the extinction of previously acquired habits, and alterations of

electric activity of the brain (for reviews see Raynaud et al., 1980;

McEwen et al., 1982; Schraa and Dirks, 1982; Luttge, 1983; De Kloet,

1984; Reese and Gray, 1984; Meyer, 1985). Since most of these reviews

were written, many new reports concerning the role of glucocorticoids in

the nervous system have appeared. A number of studies have further

investigated the localization of glucocorticoid binding in different

regions of the central nervous system including the spinal cord (Clark

et al., 1981; Duncan and Stumpf, 1984; Orti et al., 1985), the locus

ceruleus (Krasuakaya, 1982), the hypothalamus (Kato, 1983;

Ribarac-Stepic et al., 1984), the caudate-putamen (Defiore and Turner,

1983), the hippocampus (De Kloet et al., 1984), or a variety of

different brain regions (Tornello et al., 1981; Dordevic-Markovic et

al., 1983; Alexis et al., 1983; Sarrieau et al., 1983, 1984; Birmingham

et al., 1984; Ribarac-Stepic et al., 1984).

Many recent studies have focused on the role of glucocorticoids in

the development of the nervous system. Receptor binding studies in the



developing brain imply an early sensitivity to glucocorticoids (Kitraki

et al., 1984), particularly in the cerebellum (Pavlik and Buresova,

1984). The general in vivo effect of glucocorticoids on brain growth

and development appears to be one of suppression (Slotkin et al., 1982;

Doerner, 1983; Meyer, 1983; DeKosky et al., 1984; Pavlik et al., 1984)

although adrenalectomy of young rats was shown to result in

hypomyelination of the central nervous system (Preston and McMorris,

1984) indicating that the effects of glucocorticoids on neural

development are complex. Using a brain cell culture system, Stephens

(1983) reported that glucocorticoids (cortisol) stimulate rather than

inhibit the processes of myelinogenesis and brain development.

Investigations of the development of synapses between cholinergic

neurons and retina muscles in a cell culture system led to the

hypothesis by Puro (1983) that glucocorticoids regulate the development

of mechanisms which couple neuronal depolarization with release of

neurotransmitter. Lending some potential support to this hypothesis are

the findings of Leeuwin et al. (1983) who reported that glucocorticoids

induced significant changes in the morphology of synaptic vesicles of

cholinergic phrenicc) nerve terminals. Smith and Fauquet (1984)

reported that glucocorticoids indirectly stimulate adrenergic

differentiation in cultures of migrating and premigratory neural crest

by selectively enhancing catecholaminergic properties in neural crest

cells that had already been exposed to an appropriate signal of another

kind. Finally, a variety of behavioral, biochemical and morphological

teratologies of the central nervous system are known to be induced by

glucocorticoids (Bohn, 1984; Pratt et al., 1984).

Glucocorticoids have profound and complex effects on brain

metabolism. Podvigina et al. (1983) reported that a variety of

glucocorticoids inhibited respiration and phosphorylation and increased

glycolysis in slices and homogenates of rat cerebral cortex, hippocampus

and hypothalamus. In agreement with glucocorticoid binding data, these

workers found hippocampus to be most sensitive and cortex to be least

sensitive. Interestingly, Kamenov (1981) has suggested that the degree

to which glucocorticoids affect respiration may itself be under the

indirect control of the nervous system and may vary seasonally in some

animals. These and many other metabolic functions in brain are

modulated by glucocorticoids via an impressive number of enzymes. Patel

et al. (1983) showed recently that corticosterone treatment resulted in

an increase in glutamine synthetase activity in rat brain. These

findings were confirmed in mouse brain cell culture by Stephens (1983)

who found that cortisol also increased the activity of 2',3'-cyclic

nucleotide phosphohydrolase while it caused a decrease in arylsulfatase

A activity (thought to be important in myelinogenesis). Earlier work

regarding the role of glucocorticoids in glia is reviewed in McEwen

(1984). The activity of glycerol 3-phosphate dehydrogenase, another

enzyme thought to be involved in myelination and present in high

concentrations in oligodendrocytes, was recently shown to be reduced in

rat brain after adrenalectomy. Rat brain citrate synthetase activity,

on the other hand, has been recently shown to undergo an actinomycin

D-reversible decrease upon administration of cortisol (Sharma and

Patnaik, 1984). These same workers reported earlier that cortisol

differentially regulates the activity of malate dehydrogenase isoenzymes

in rat brain (Sharma and Patnaik, 1982). Complex results have been

obtained for tyrosine aminotransferase activity in brain as well.

Mishunina and Babicheva (1981) reported that a single injection of

cortisol to rats caused a reduction of tyrosine aminotransferase

activity in the hypothalamus and medulla oblongata while causing an

increase in the activity in the hippocampus. Glucocorticoid injection

has also been shown to increase arginase activity in guinea pig brain

(Bjelakovic and Nikolic, 1983) and to "modulate" ornithine decarboxylase

activity in rat brain (De Kloet et al., 1983). Interestingly, tyrosine

hydroxylase activity in the superior cervical ganglion was shown to be

increased by injections of the synthetic glucocorticoid, dexamethasone,

but not by endogenous glucocorticoids (corticosterone or cortisol)

regardless of dose (Sze and Hedrick, 1983). Eranko et al. (1982), using

fluorescent histochemical techniques, reported that injection of newborn

rats with cortisol caused a great increase in the number of the small,

intensely fluorescent cells and the appearance of similar small cells

with intense immunohistochemical reactions for tyrosine hydroxylase

(TH), dopamine-B-hydroxylase (DBH) and phenylethanolamine (norepi-

nephrine) N-methyltransferase (PNMT) in the superior cervical ganglion.

At the same time, the epinephrine content and the PNMT activity of the

ganglion greatly increased, while no significant changes were observed

in the dopamine or norepinephrine content or TH or DBH activity. PNMT

expression has also been shown to be influenced by glucocorticoids in

cells derived from the neural crest (Bohn, 1983).

Single acute doses of dexamethasone have been recently shown to

affect the levels of monoamines and their metabolites in rat brain

(Rothschild et al., 1985). These researchers reported significant

increases in dopamine levels in the hypothalamus and nucleus accumbens

of dexamethasone-treated rats when compared with saline-treated rats,

whereas no significant effect of dexamethasone on dopamine levels in

frontal or striatal brain areas was found. These workers also found

that dexamethasone led to a significant increase in serotonin in

hypothalamus, whereas dexamethasone caused a significant decrease in

serotonin in the frontal cortex. These findings may help to explain the

observations that corticosteroids lower prolactin levels and can induce

psychiatric disturbances as well as the finding that effectively ill

patients with high post dexamethasone cortisol levels are frequently


Other recent findings of relevance to the effects of gluco-

corticoids on neurotransmitter systems include the report by Mishunina

and Kononenko (1983) that a single injection of cortisol increased the

uptake of gamma-aminobutyric acid (GABA) by synaptosomes of hippocampus,

hypothalamus and cerebral cortex. However, in vitro studies by these

same workers resulted in a glucocorticoid-dependent increase in GABA

uptake by only the hypothalamus synaptosomes, indicating differential

mechanisms of glucocorticoid action in these different brain regions.

Haidarliu et al. (1983) reported that perfusion of the rat neostriatum

with dexamethasone after stereotaxic injections of [3H]-labeled dopamine

into the caudate nucleus increased basal but inhibited potassium-

stimulated [3H]dopamine release and yield of metabolites. Slotkin et

al. (1982) reported that neonatal exposure to dexamethasone resulted in

alterations in norepinephrine synthesis and turnover in both central and

peripheral sympathetic neurons. Further demonstration of the

relationship between glucocorticoids and central neurotransmitter

systems is evident from studies reporting a decrease in the

corticosterone binding capacity of hippocampus after administration of

6-hydroxydopamine (Weidenfeld et al., 1983) or 5,7-dihydroxytryptamine

(Siegel, 1983). Similarly, the possible role of medial forebrain bundle

catecholaminergic fibers in the modulation of glucocorticoid negative

feedback effects has been further elucidated by investigating the

effects of 6-hydroxydopamine treatments in rat brains (Feldman et al.,

1983). In addition to effects on transmitter uptake and synthetic and

degradative enzyme activity, some of the effects of glucocorticoids on

catecholaminergic function are known to be via catecholaminergic

receptors (see review by Davies and Lefkowitz, 1984).

Many of the actions of glucocorticoids outside the brain, such as

their effects on lipolysis, gluconeogenesis, glycogenesis and

glycogenolysis, are regulated primarily by cAMP or hormones acting

through cAMP leading to the notion that glucocor,ticoids have

potentiating (permissive or synergistic) effects on the action of many

hormones that act through cAMP as the intracellular second messenger

(for review see Liu, 1984). Such wide-ranging effects of gluco-

corticoids on the effects of these other hormone systems, thought to be

via the regulation of the autophosphorylation of cAMP-dependent protein

kinase, will likely be shown to play an important role in the

glucocorticoid modulation of neural function as well.

A recent finding of potential importance to the field of epilepsy

research involves the discovery that natural, but not synthetic, gluco-

corticoids modulate the binding of [3H]muscimol, a GABA agonist, in

crude synaptosomal membranes and in brain sections of rats (Majewska et

al., 1985). In general, nanomolar concentrations concentrations of

corticosterone and pregnenolone-sulfate enhanced mucimol binding in

cerebral cortex, cerebellum, thalamus and hippocampus via an apparent

increase in the affinities of GABA receptors. More recently, two

metabolites of the steroid hormones progesterone and deoxycortico-

sterone, were shown to be potent barbiturate-like ligands of the GABA

receptor-chloride ion channel complex (Majewska et al., 1986). At low

concentrations both steroids inhibited binding of the convulsant

t-butylbicyclophosphorothionate to the GABA-receptor complex and

increased binding of benzodiazepine flunitrazepam. In addition,

chloride uptake into isolated brain vesicles was stimulated and the

inhibitory actions of GABA in cultured rat hippocampal and spinal cord

neurons were potentiated. These findings may help to explain the

ability of certain steroid hormones to rapidly alter neuronal excit-

ability (see below) and may provide a mechanism for the anesthetic and

hypnotic actions of naturally occurring and synthetic anesthetic


The effects of glucocorticoids on the electrical activity of

neurons has been studied at the level of single neurons by the method of

microiontophoresis. This technique has been used to study glucocort-

icoid-induced changes in activity of neurons in a variety of brain

regions including hypothalamus (Mandelbrod et al., 1981), medial septal

area (Papir-Kricheli and Feldman, 1981), reticular formation (Avanzino

et al., 1983) and hippocampus (Belyi et al., 1982). In most of these

studies glucocorticoids were found to inhibit some neurons, excite some

neurons and have no effect on still other neurons in the same brain

region, although the ratios of the three groups varied between regions.

Peripherally, the effects of glucocorticoids on neuromuscular function

have also been investigated. These effects are generally facilitatory

(see review by Hall, 1983), although there have been reports to the

contrary (Korneeva and Emelyanov, 1981).

The role of glucocorticoids in the regulation of physiological

reactions to stress has been widely studied (for review see Munck et

al., 1984; Sapolsky et al., 1986) and at least some of these stress-

related functions of glucocorticoids are known to involve the central

nervous system. One example that has received a good deal of attention

recently involves the effects of glucocorticoids on stress-induced

analgesia. Adrenalectomy, hypophysectomy and/or glucocorticoid

treatment have all been shown to have significant effects on the

stress-induced (cold water, footshock, etc.) inhibition of pain

responsiveness. Markel et al. (1984) reported that adrenalectomy

significantly reduced the threshold for footshock sensitivity in rats

and that steroid replacement restored, or even potentiated, this pain

sensitivity. Similarly, Mousa et al. (1983) reported that analgesia

induced in rats by cold-water swim stress and measured by the tail-flick

and hot-plate methods was significantly antagonized after IP pretreat-

ment for three days with dexamethasone. These investigators go on to

suggest that since naloxone also attenuated the analgesia developed by

the cold-water swim stressor, corticosteroids may have a role in

modulating stress-induced analgesia via the endogenous opiate system.

Evidence of this was presented recently by Terman et al. (1984) who used

inescapable foot shock as a stressor and the tail-flick test as a

measure of analgesia. These workers, using multiple doses of gluco-

corticoids, found that lower doses potentiated and higher doses

significantly reduced stress analgesia in both sham and hypophysect-

omized animals. Despite some evidence to the contrary (Baron and

Gintzler, 1984), these results, along with the finding that nonopiod

analgesia is unaffected by dexamethasone (Lewis et al., 1980) and that

corticosterone exerts the same biphasic dose-dependent effects on

morphine analgesia (Terman et al., 1985), suggest an opiod action for

glucocorticoids, perhaps at the receptor level. Of some potential

relevance to this topic is a report that adrenalectomy significantly

potentiated and corticosterone administration attenuated stress-induced

decreases in norepinephrine contents in the hypothalamus and thalamus,

and the increases in 3-methoxy-4-hydroxyphenylethyleneglycol sulfate (a

principle metabolite of norepinephrine) levels in the hypothalamus,

amygdala and thalamus (Nakagawa et al., 1983).

The idea that glucocorticoids, via the adrenocortical axis, may

have an important role in the process of aging has been examined in

depth recently (Sapolsky et al., 1986). It has been shown that the aged

male rat is impaired in terminating the secretion of glucocorticoids at

the end of stress. This hormonal excess is thought to be due to

degenerative changes in a region of the brain which normally inhibits

glucocorticoid release; the degeneration, in turn, appears to be caused

by cumulative exposure to glucocorticoids (Sapolsky and McEwen, 1985;

Sapolsky, 1986). Together, these effects form a feed-forward cascade in

the aging subject with potentially serious pathophysiological


In addition to the above findings regarding the role of

glucocorticoids in the nervous system, glucocorticoids are clinically

important in the field of neurosurgery, most notably for the treatment

of various kinds of brain edema related to intracranial tumors, cerebral

infarctions and head injuries (Bouzarth and Shenkin, 1974;

Sugiura et al., 1980; Yu et al., 1981).

Although there is limited evidence of glucocorticoid binding

directly to synaptic (Towle and Sze, 1983) and other (Duval et al.,

1983) membranes, most of the glucocorticoid effects amenable to

experimentation have been shown to result from a receptor-mediated

change in concentration or activity of key proteins. This change is a

consequence of an action of the glucocorticoid-receptor complex on the

molecular machinery that controls gene expression. Recent advances such

as recombinant-DNA and monoclonal-antibody techniques have offered new

ways of exploring this machinery. Recently, a monoclonal antibody

against the rat liver glucocorticoid receptor in combination with the

indirect immunoperoxidase technique has made it possible to demonstrate

glucocorticoid receptor-immunoreactive nerve and glial cell nuclei all

over the tel- and diencephalon of the male rat (Fuxe et al., 1985).

Since the glucocorticoid receptor qualifies as one of the few proteins

known to control gene expression in eukaryotes, it is not surprising

that glucocorticoids have become a fashionable tool for the molecular

biologist (for reviews see Ringold et al., 1983; Lan et al., 1984; Hager

et al., 1984; Rousseau, 1984a; Scheidereit et al., 1986). A major step

in the understanding of the structure and function of the glucocorticoid

receptor was made recently when the primary structure of a functional

human glucocorticoid receptor cDNA was identified and subsequently

cloned (Govindan et al., 1985) and expressed in vitro (Hollenberg et

al., 1985). The focus of the following review and the proposed

research, however, will be the receptor molecule itself and the changes

in this molecule brought about by the binding of a steroid ligand which

enable it to interact with the machinery that controls gene expression.

Although little has been said in this brief review regarding the

structure or function of the receptors in brain for other classes of

steroids (i.e. estrogens, progestins, androgens), it is particularly

relevant to briefly note recent findings regarding the so called

mineralocorticoid, or "type I", receptor in neural tissues. The

relevance of mineralocorticoid receptors to glucocorticoid actions in

neural tissues stems first from the fact that [3H]aldosterone has been

shown to bind to glucocorticoid, or "type II", receptors as well as type

I receptors in brain cytosol (Anderson and Fanestil, 1976) by virtue of

biphasic Scatchard plots for [3H]aldosterone, but not [3H]dexamethasone.

Later reports indicated that inclusion of the new synthetic gluco-

corticoid RU 26988 (unlabeled), which binds to type II but not type I

receptors or corticosteroid binding globulin (Mogiulewsky and Raynaud,

1980), linearizes the [3H]aldosterone Scatchard plot,(Beaumont and

Fanestil, 1983; Emadian et al., 1986), leaving only a high-affinity

component assumed to be the type I receptor. A second point of

relevance involves the fact that in addition to aldosterone,

corticosterone has a high affinity for the type I receptor (Krozowski

and Funder, 1983), which has led some to believe that corticosterone may

be a major, if not the primary, ligand for type I receptors in brain.

This topic is of great concern to those investigating glucocorticoid

actions in the nervous system and is currently the subject of

investigation by others in this lab. It should be noted, however, that

synthetic glucocorticoids, such as dexamethasone or triamcinolone

acetonide, have a much lower affinity for the type I receptor and

usually possess "purer" glucocorticoid properties than corticosterone as

determined by their physiological actions in peripheral tissues

(Meyer, 1985). In addition to their high specificity for type II

glucocorticoid receptors, other advantages of using these synthetic

glucocorticoids include the fact that their rate of dissociation from

the receptor is considerably slower than that of natural glucocorticoids

and have an insignificant affinity for corticosteroid binding globulin.

For these and other reasons, the research reported here was carried out

using dexamethasone and/or triamcinolone acetonide exclusively, and

therefore all further references to "glucocorticoid receptors" will

imply type II receptors for adrenal steroids unless otherwise indicated.

Major areas of glucocorticoid receptor research that are of

particular relevance to the glucocorticoid modulation of nervous system

function include the following: 1) The stabilization and potential

mechanisms of up- and down-regulation of unoccupied glucocorticoid

receptors. 2) Activation of the glucocorticoid-receptor complex to the

nuclear-binding form and subsequent interaction with the genome. 3)

Purification of the glucocorticoid receptor and subsequent detailed

physicochemical analysis.




Hydrophobic interactions basically result from the adherence of

non-polar solutes or groups in aqueous solution, resulting,

consequently, from their repulsion from the strong polar-polar

interactions of water molecules. Like its ionic properties, a

molecule's surface hydrophobicity depends upon its primary structure.

A change in the surface hydrophobicity of glucocorticoid-receptor

complexes associated with activation has been suggested by workers

reporting an increase in the partitioning of these complexes after

activation into the polyethylene glycol (PEG) layer in an aqueous

dextran-PEG two phase system (Andreasen, 1978 and 1982; Andreasen and

Mainwaring, 1980) and by Luttge et al. (1984d) who, in addition to an

increase in partitioning into PEG, found a large increase in the

fractional adsorption of [3H]triamcinolone acetonide-receptor complexes

to glass fiber (Whatman GF/C) filters following activation. This

putative increase in hydrophobicity does not appear to be unique to the

glucocorticoid receptor, since similar transformations following

activation have been reported recently for estrogen (Gschwendt and

Kittstein, 1980; Murayama and Fukai, 1983), androgen (Bruchovsky et al.,

1981) and progesterone (Lamb and Bullock, 1983) receptors. Although the

functional significance of this increase in hydrophobicity is unknown,


it may contribute to the increased binding affinity of the activated

receptor for DNA. Increasing the hydrophobicity of the receptor may

also increase its rate of permeation through the nuclear membrane.

Hydrophobic interactions have also been reported to account for the

high affinity binding of glucocorticoids to their receptors (Wolf et

al., 1978; Alfsen, 1983; Eliard and Rousseau, 1984). According to these

workers, binding requires both sides of the steroid to be enveloped by

the receptor.

The influence of hydrophobic interactions on the structure and

steroid-binding properties of glucocorticoid receptors has recently been

investigated by means of chromatography on Sephacryl S-300 or Lipidex

1000 or by incubation with charcoal or phospholipase C (Bell et al.,

1986). These workers suggest that receptor activation is preceded by

structural changes associated with the loss of a lipid factor from the

complex. They also reported that non-polar steroid antagonists, and

lipophilic compounds such as phenothiazines, bound to secondary,

hydrophilic sites on the receptor and exerted allosteric effects on the

primary steroid-binding site implying that hydrophobic interactions may

be important determinants of the structure and properties of

glucocorticoid receptors.

One particularly useful method for investigating the hydrophobic

characteristics of steroid receptors is hydrophobic interaction

chromatography. Typically this involves characterizing the chromato-

graphic behavior of the receptor in question on a series of hydrophobic

matrices of decreasing polarity. Bruchovsky et al. (1981) investigated

the hydrophobic properties of the rat prostatic nuclear androgen

receptor using gamma-amino-alkyl derivatives of agarose with varying

alkyl substituent lengths. These workers found the adsorption of

receptor to the modified agarose gels increased with the length of the

alkyl substituent. However, the presence of the terminal amino group in

these gels introduces the possibility of ionic effects which could

complicate the interpretation of the hydrophobic interactions (Shaltiel,

1974). More recently, Lamb and Bullock (1983) investigated the hydro-

phobic properties of the rabbit uterine progesterone receptor using a

series of alkyl agarose columns of increasing alkyl chain length (no

terminal amino group). No such studies have been conducted on

glucocortcoid receptors.

In addition to gaining further knowledge regarding the process of

receptor activation, more detailed hydrophobic interaction studies of

both activated and unactivated forms of the glucocorticoid-receptor

complex, as well as the non-steroid bound form of the receptor, are

likely to provide valuable information concerning the physicochemical

properties of each of these receptor forms. Such studies might also

provide potentially useful means by which partial purification of these

different receptor forms can be achieved.

Materials and Methods

Chemicals, Steroids and Isotopes

[6,7-3H]Triamcinolone acetonide, or 9a-fluoro-llb,16a,17a,21-tetra-

hydroxy-1,4-pregnadiene-3,20-dione-16,17-acetonide, ([3H]TA, specific

activity = 37 Ci/mmol) was purchased from New England Nuclear (Boston,

MA). Sephadex G-25 (fine) was obtained from Pharmacia Fine Chemicals

(Piscataway, NJ). Dithiothreitol (DTT) and 4-(2-hydroxyethyl)-l-piper-

azineethane-sulfonic acid were courtesy of Research Organics (Cleveland,

OH). Sodium molybdate (Na2Mo04), calf thymus DNA-cellulose, glycerol,

PPO (2,5-diphenyloxazole) and dimethyl POPOP (1,4-bis[2(4-methyl-5-

phenyloxazoyl)]benzene), phenyl agarose and alkyl agaroses were all

purchased from Sigma Chemical Co. (St. Louis, MO). Scinti Verse II was

purchased from Fisher, Inc. (Fair Lawn, NJ). All other chemicals and

solvents were reagent grade.


All studies used female CD-1 mice (Charles River Laboratories,

Wilmington, MA) that were subjected to combined ovariectomy and

adrenalectomy approximately 1 week prior to each experiment in order to

remove known sources of endogenous steroids. Both operations were

performed bilaterally via a lateral, subcostal approach under barbi-

turate anesthesia, and mice were given 0.9% NaCI (w:v) in place of

drinking water. On the day of the experiment mice were anesthetized

with ether and perfused slowly through the heart with ice-cold HEPES-

buffered saline (20-30 ml, isotonic, pH 7.6).

Cytosol Preparation and Steroid Binding

Brains were removed from the perfused animals and homogenized (2 x

10 strokes at 1000 rpm) in 4 volumes of ice cold Buffer A (20 mM HEPES,

2 mM DTT and 20 mM Na2MoO4, pH 7.6 at 0 C) in a glass homogenizer with

a Teflon pestle milled to a clearance between the pestle and homogeniz-

ation tube of .125 mm on the radius (to minimize rupture of the brain

cell nuclei (McEwen and Zigmond, 1972). The crude homogenate was

centrifuged at 100,000 g for 20 min and the supernatant recentrifuged at

100,000 g for an additional 60 min to yield cytosol. During these

centrifuge runs, and during all other procedures, unless otherwise

indicated, careful attention was paid to maintaining the cytosol at

0-2 C. Final protein concentrations were typically in the 6-8 mg/ml

range. For experiments involving occupied (activated or unactivated)

receptors, cytosol samples were incubated with 20 nM [3H]TA for 24 to 40

hours at 0 C with or without a 200-fold excess of unlabeled TA.

Experiments involving hydrophobic interaction chromatography of

unoccupied receptors required that the individual fractions collected

from the hydrophobic and control columns be postlabeled with 20 nM

[3H]TA +/- 4 um [1H]TA. Prior to hydrophobic interaction chromatography

of unlabeled receptors, cytosol was run on Sephadex G-25 columns equili-

brated and eluted with the appropriate buffer to be used for

subsequently eluting the hydrophobic columns.

Activation and Removal of Unbound Steroid

[3H]TA-labeled cytosol samples to be activated were first run on

Sephadex G-25 columns (0.6 x 14 cm) equilibrated with buffer containing

20 mM HEPES and 2 mM DTT only. This column run resulted in the removal

of both free steroid and molybdate, allowing for uninhibited activation

(Luttge and Densmore, 1984; Luttge et al., 1984a,b,d). The bound

fraction was then incubated at 22 C for variable periods of time to

determine the minimum period of time required to achieve total

activation. The degree of activation was determined by DNA-cellulose

binding assay. For the purpose of hydrophobic interaction chromato-

graphy of activated receptors, the maximally activated receptor

preparation was again run on a second Sephadex G-25 column equilibrated

with the appropriate buffer (the same buffer used to equilibrate and

elute the subsequent alkyl agarose columns). For unactivated receptors,

the cytosol was treated in an identical fashion except that the first

Sephadex G-25 column was equilibrated and eluted with HEPES buffer

containing 20mM molybdate and 2 mM DTT and the bound fraction was

subjected to a 0 C (rather than a 22 C) incubation prior to running on

the second Sephadex G-25 column (identical to the second column run for

the activated preparation).

In the case of postlabeling of unoccupied receptor fractions or

monitoring of dissociation occurring during the hydrophobic column runs

of prelabeled receptors, all or part of each hydrophobic column fraction

was again subjected to Sephadex G-25 chromatography to remove free


DNA-cellulose Binding Assay

Calf thymus DNA-cellulose, prepared originally by the method of

Alberts and Herrick (1971), was added to HEPES buffer containing 2 mM

DTT to yield a final concentration of 10 mg/ml (4.1 mg DNA/g

DNA-cellulose). In a typical assay (run as duplicates) a 100 ul aliquot

of [3H]TA- labeled cytosol was added to 300 ul of the DNA-cellulose

slurry. The mixture was then vortexed gently and incubated at 0 C for

60 min. Assay tubes were oscillated at 150 rpm throughout the

incubation and vortexed gently every 10 min. The DNA-cellulose was

collected by centrifugation (2000 x g for 5 min), the supernatant

discarded and the DNA-cellulose pellet washed 3 times with 1 ml of HEPES

buffer plus 2 mM DTT. The entire pellet was resuspended in deionized

water and transferred to scintillation vials for determination of bound

radioactivity. DNA- cellulose was prepared in buffer not containing

molybdate in order to reduce the concentration of molybdate during the

binding assay, since high concentrations of molybdate were found to

reduce the efficacy of receptor binding to DNA-cellulose (Luttge et al.,


Receptor Stability Determinations

Prelabeled receptor preparations (both activated and unactivated),

were run on Sephadex G-25 columns equilibrated and eluted with 50 mM

molybdate and varying concentrations of KC1 (0 2,000 mM) in order to

determine the effects of ionic strength on receptor stability. A

fraction of the bound fraction was immediately run on a second Sephadex

G-25 column prior to specific binding determination, while the remainder

was incubated for 4 hr at 0 C prior to being run on a second Sephadex

G-25 column and specific binding determined.


The initial experiment in the investigation of glucocorticoid

receptor hydrophobicity involved hydrophobic interaction chromatography

of activated and unactivated glucocorticoid receptor complexes on a

limited Shaltiel series using a series of elution buffers of decreasing

hydrophobic character. This experiment used a series of hydrophobic

gels created by covalently attaching alkyl groups of increasing chain

length to agarose including propyl (n = 3 carbons), hexyl (n = 6), octyl

(n = 8), decyl (n 10) and dodecyl (n = 12) agarose, along with

unmodified agarose as a control. Activated and unactivated receptor

preparations equilibrated in 600 mM KC1 and 50 mM molybdate were applied

to each of the columns (containing 1 ml of gel) and then eluted step

wise with the following buffers (each of which contained 50 mM molybdate

in HEPES buffer): 600 mM KC1, 300 mM KC1, 0 mM KC1, 10% glycerol and 30%

glycerol. Results (not shown) indicated that activated preparations

were retained longer than unactivated and that this difference was the

largest on the hexyl agarose columns. Both receptor forms appeared to

be eluted completely with higher salt concentrations on all gels except

octyl, decyl and dodecyl which exhibited variable retention of both

forms of the receptor in the presence of the more hydrophobic elution


Further refinement of the procedure to investigate receptor

hydrophobicity next involved hydrophobic interaction chromatography of

activated and unactivated glucocorticoid receptor complexes on a more

complete Shaltiel series of alkyl agarose columns using a series of

elution buffers of decreasing hydrophobic character. In addition to the

original gels used, butyl (n=4), pentyl (n=5) and phenyl (n=P) agarose

were added and the gel volume was increased to 1.5 ml. In addition,

smaller fractions (0.5 ml) were collected to increase resolution and a

different series of step wise buffer elutions were used including: 600

mM KC1, 0 mM KC1 and 10% ethylene glycol (all in HEPES buffer plus 50 mM

molybdate). A final elution of 6 M urea was used to remove those

receptors (and other proteins) that were strongly retained throughout

all of the elution steps. With the higher resolution resulting from

smaller fraction size, it was obvious that activated receptors were

retained longer than unactivated receptors run on butyl, pentyl and

hexyl agarose columns. On all three columns, however, both forms of the

receptor were eluted completely, though at different rates, by 600 mM

KC1. Phenyl agarose strongly retained the activated receptor, whereas

most of the unactivated form of the receptor was eluted rapidly under

the high ionic strength conditions. Octyl, decyl and dodecyl agarose

retained the receptors more than the other gels except that there were

no significant differences between the activated and unactivated forms

on these columns. Overall, the pentyl agarose exhibited the greatest

differences in retention between the activated and unactivated forms

while still allowing for both forms to be eluted completely under the

same buffer conditions.

Since the results of the previous experiments indicated that any

binding retained by any of the hydrophobic columns after adequately

washing with 600 mM KC1 could not be eluted with anything tested short

of denaturing solutions such as 6 M urea, hydrophobic interaction

chromatography of activated and unactivated glucocorticoid receptor

complexes on a Shaltiel series using only 600 mM KC1 was performed.

This experiment was similar to the previously described experiment

except that only one elution buffer was used instead of a series of

different buffers. The results basically replicated those of the

previous experiment in that the activated form of the receptor

consistently shows a greater degree of retardation on most of the

columns with pentyl, hexyl and phenyl agarose providing the greatest

discrimination between the two forms (Figure 2-1).

Since high ionic strength elution buffers (HEPES buffer plus 50 mM

molybdate and 600 mM KC1) were able to discriminate between the

activated and unactivated forms of the receptor on virtually all of the

columns tested while allowing all of the binding to be eluted, the

effects of using lower or higher ionic strength buffers to equilibrate

and elute the hydrophobic columns were investigated. Because of

uncertainties regarding the activation of unactivated glucocorticoid

receptors and the stability of glucocorticoid binding under high salt

conditions, hydrophobic interaction chromatography of activated and

unactivated glucocorticoid receptor complexes on a Shaltiel series

equilibrated and eluted with only HEPES buffer and 50 mM molybdate (no

KC1) was first attempted (Figure 2-2). Proply, butyl, pentyl and hexyl


W 0J0N40
4-Ul I d CA
0 CU )0 6 4 0 II
o (3i )-I 0 0 C
r4 Q W n 0 4 4 i
U i r W i 4 0 0 x II
il = 0 1 "4 m
4 H a col 4 a C 4co
0 CC Ca O CB Cda 0 C" W
uo 0 0 a0 0 m 0 C 1O
0 u a)39 u 0. w 0 ao 0

s: O 2* ocUc-4I oo

-o -4 r4 ) co ro 0 w
a 0g 0 a a- >% C H 0 w c
P4- S E- r-41 3 O

( -C 'O 0 mi 3 CO
a U 93 CD 0 v (L0 0 1
E-4 U r-4 o a m >,
C 4-'0 0 4J 0 ^ Q 4Ul U
roa 4) 00$a W HCU C f
44u 4 4 r. (U -3 s 0 0
Aji 4-i o 3 a ar-i -O
0 0 0 w3 0 a B 0 0 Ai 0 HI

CCO 4 4 i
U 0 a >-.4 O 0)L O
v U % oo B- c4 o
> n 0 T, s l^ a s H en

co ) 9 4 04
o .4 W I= 1w00
(1 0 e'I0 z co 0 w a u- r-4 co ca
co 0 I3 0 0) "4 0 C o0 U C -l

> "-I ca -W w 4

3 0 00 0o ua
C 0 4 -N4 O t i-T 11CO

Se0 4 e 0OMO -I H 0 O
0: 0 w on a 0 0 00

O,-4 en 00 cn > CUi 0 a
OH O H CH -o

o *n > 4) I a 0H
$4 4 P 0 4J U 3 0 0 4i o3


i000 o !?o n
O0 0 0to0Sa 0 >%

0 -L 0 -
$4 r-4 0 4 t> i4 0 0 C) en

rS4 Co o 44- 0 0

0 *W oN -4
Hi cW a> U 0c

a J40 0 0 41 01 jj
t-1 0 0 4 O H o
U 0 0 0 3 Is 0 >4 0

W-r 4 CO 4-aJ co -W m 0 p& 4 'V
q) 4-1 B I-t4 C

41 4 0 c A 0 0 r. I C
0 co $ 4I 4 3 4l Z0 *

jA oS o t > t O
0i >- 0 vu 0) Sa cu I
4M 0 a a) co 3 0) 4
0 1 44 O 4 4-W to3 e n
w 0 4 u > i 4o C a) .4 1
x0 H = il 0 r 1 W 0 i0 (l

.10 M 3 a ca u 0 4 > ) (0

-4 u 0a wv 0u u p a
1 1 "d3%O o 0 0
6 -W 04 6 4 ) (D 4 La
$4 0. a) UO 3 1 0 3 w 0u 3 ? 04
0c u a = 4 3 U ? 0 4 %4

0 0 0 0 0 0
D0O o u D cu

IDIJ, O 0o jUc9j ld








N- N

0 0 0


Q 44

om rCT6 C 1
w40 a
0 4 44 4
4.4 04

0 W
0 P. 1 0a s0

SE CO 3 .w .0

r-4 003 0

HOI 00
10 C s 4i a

" -40

0 l 4
0 o 00 -4

4C 4 1 0 0
> 0 0 0 0 w
-H a (U pa j-
41 r-4 4

E 0
c 0 4h
co a c 4 co
0 W -4 .0 0
CO r4 >1 4J.

> w0 0 w
4-14 0 $ -v- 44

SC0 4-4 a'
4 $4 W) 0)
0 r0 w '-

V a w 0

-W ... -4
o o co

0 0 0
60 4- r.
0 r--4C3 0- 0 431
SO- C (V 0

S3 m U 04

Z-4 -0 0
cr (a u a

0 4 0 0 CO
"4 "1 r-4

W Q104.W 0
o w h 4 i q l w

- W 0 C 4

0 04lC0
41 0 w w 0
Q r- c0 .0 0

(0 0 1-4 40.

0 0 4-4 0 -4
MC 0 l 3r%

s Lr 0

0~- 0 4'0 41
SSr4 ci
0 I4 r4-

S0.0 0 0CO

0 0 w- 0
w u 0 0

Z 0. CO 0 0)
bjr o J0<

0. -a
x r-
0 II
$4 ) 0l

V ,-4 W
0 S ,'00
-4 00

u co
e 4a

ca cs au
C 00

0 0 -4
a 00
OU Q II r1

e r 1

-o 0 0

0 ,4o -1
0U 0

S..4I 0
0 0
3 00

0 d 0
-4 06 II
0- 00

a c

0C 0 40

0 I
0n u

C. O
0 0r II

O 00
O' 0

-0 0 0

0d 0
0 0 4.1
0 *0 0

44 VO
4 0)
0 -4 0 co
0 c0. 0

0 0 00
0 0a
4 )D r-.4

0 0 0 >%
10 1- 10

0 0 -

0 0 SFl 0 -

o ofc a *a



I IIl I l


S0 4 0 0

v cll 4-- "" l i -

a10101 o u a2:Jad

agaroses all showed an increased affinity for the unactivated receptor,

resulting in varying degrees of increased retention when compared to

elution by higher salt. Interestingly, pentyl agarose showed the

greatest retention of the unoccupied receptor of any of the gels. In

contrast, the unactivated receptor exhibited a decrease in retention on

the phenyl agarose. There was no appreciable differences in the elution

profiles obtained from octyl, decyl or dodecyl agarose in high or low

salt. As was the case for unactivated receptors, propyl, butyl, pentyl

and hexyl agaroses all showed an increased affinity for the activated

receptor under low salt conditions, resulting in varying degrees of

increased retention when compared to elution by higher salt. Interest-

ingly, pentyl agarose exhibited the greatest retention of any of the

gels examined for both the activated and the unactivated forms of the

receptor. There was no appreciable differences in the elution profiles

obtained from octyl, decyl, dodecyl or phenyl in high or low salt.

The finding that octyl agarose exhibited greater retention of both

forms of the receptor than decyl agarose, which exhibited greater

retention than dodecyl agarose regardless of ionic strength (one would

have expected the reverse if these interactions were truly hydrophobic),

led to a more extensive investigation of the interactions between the

glucocorticoid receptor and these highly hydrophobic gels. Although

free steroid is removed from the cytosolic preparations immediately

prior to applying these preparations on the hydrophobic columns and

binding with [3H]TA has been shown to exhibit high stability and very

slow dissociation even under high salt conditions, the possibility that

the hydrophobic alkyls might themselves be destabilizing the receptors

via strong hydrophobic interactions had not been examined. This

involved replicating the previously described experiment whereby

activated and unactivated receptors where applied to the complete

Shaltiel series of columns equilibrated and eluted with HEPES buffer

plus 50 mM molybdate and 600 mM KC1 except that each fraction eluted

from the column was divided into two equal halves. One half was counted

directly as before while the other half was subjected to an additional

bound/free separation on Sephadex G-25 columns. When the profiles

determined by each of these methods for each of the gels were compared,

there were no significant differences observed except in the case of the

octyl, decyl and dodecyl agaroses. None of the counts eluting from

these three columns eluted in the bound fraction during the subsequent

Sephadex G-25 bound/free separation.

Since the results from the preceding experiment indicated how

receptor instability and/or dissociation could complicate the

interpretation of the results from hydrophobic interaction chromato-

graphy of these receptors, and since many of these experiments required

higher concentrations of molybdate and potassium chloride than are

generally encountered throughout most of the literature regarding

glucocorticoid receptors, effects of these parameters on receptor

binding properties obviously warranted further study. In this

experiment, cytosol was prepared in the usual fashion in the presence of

20 mM molybdate and 2 mM DTT and labeled with 40 nM [3H]TA plus or minus

unlabeled TA for 24 hours to achieve maximal binding. After this

initial incubation, aliquots were taken to determine total and non-

specific binding. Next, aliquots of prelabeled cytosol were applied to

Sephadex G-25 columns preequilibrated with every possible combination of

the following concentrations of sodium molybdate (0, 20 or 50 mM) and

KC1 (0, 50, 150, 300, 600, 1200 or 1800 mM). Of the bound fraction

collected from each of the 42 columns (21 total and 21 nonspecific), an

aliquot was counted to represent the "zero hour" binding for each buffer

condition. Each of the samples representing total binding was then

further subdivided into two tubes, one containing 20 nM [3H]TA (for the

determination of destabilization) and one containing 4 uM unlabeled TA

(for the determination of destabilization plus dissociation). Each of

the samples representing nonspecific binding was also subdivided into

two tubes, one containing 20 nM [3H]TA and 4 uM unlabeled TA and the

other containing only 4 uM unlabeled TA. After an 8 hour incubation at

0 C, aliquots of each of these experimental groups was again subjected

to bound/free separation on Sephadex G-25 columns. Rates of destabil-

ization and dissociation (which rarely exceeded a few percent) were very

low for all groups, to the point that neither was likely to have been a

significant factor in any of the hydrophobic interaction chromatography

studies involving high concentrations of molybdate and/or KC1. It must

be emphasized, however, that these results cannot necessarily be applied

to the binding of this receptor to other ligands, such as DEX or

corticosterone, or to the binding of this receptor under different

temperature or pH conditions. However, since the conditions of this

experiment matched very closely the conditions of the hydrophobic

studies, any significant loss in binding seen during these studies is

likely to be the result of destabilizing effects of the hydrophobic

matrix instead of the buffer itself.

Since potassium chloride concentrations of up to 1200 mM with and

without molybdate were shown to have little effect on the stability or

rate of steroid (TA) dissociation of glucocorticoid receptors maintained

at low temperatures, and since retention of both activated and

unactivated receptors was shown to increase dramatically on alkyl

(particularly pentyl) agarose gels when ionic strength was lowered, the

effects of ionic strength and molybdate concentration on hydrophobic

interaction chromatography of both activated and unactivated receptors

was investigated in greater detail. Unactivated receptor preparation

was applied to pentyl agarose columns equilibrated in, and then eluted

with, one of the following: HEPES only, 10 mM molybdate, 50 mM

molybdate, 150 mM KC1 and 50 mM molybdate, 300 mM KC1 and 50 mM

molybdate, 600 mM KC1 and 50 mM molybdate, 900 mM KC1 and 50 mM

molybdate or 1,200 mM KC1 and 50 mM molybdate (Figure 2-3). Despite the

finding that high ionic strength had little effect on TA dissociation

from glucocorticoid receptors, the eluted fractions were subjected to

another bound/free separation step as an extra precaution. Pentyl

agarose was chosen as the gel to study the effects of ionic strength

since it was pentyl agarose that displayed total retention of both forms

of the receptor under low salt conditions. Since the effect of

molybdate (which had been present in all prior hydrophobic interaction

chromatography experiments in this study) was unknown, the concentration

of both molybdate and KC1 were varied in this experiment. Columns

eluted with HEPES only, 10 mM molybdate or 50 mM molybdate displayed

total retention by the column. The elution profiles of all the

KCl-containing buffers were very similar except that a very slight

decrease in retention was observed going from 150 to 300 to 600 mM KC1.

The differences between these three concentrations, though possibly

significant, were not very dramatic. Elution profiles resulting from

600, 900 and 1,200 mM KC1 showed virtually no changes. In a parallel

in 0
eN 41
I 0 uA
w T3 w8
1aI e0c 4
o a) Ch u o 4J 4- C

900ot E a.0
r 0 r4 ) 0 0
o 0004 .W40 WB 4
00 UO H 4 1 i

0 4 -c00 15 0 W
Wo 0 -i 0 .-40 4-4
H Hc 4O 0 "0 C0

HEC W 0 0 to W

u 0 v Co 0

0 C0 0a 0

> 0 0 Q44 04
4r r 41 0 4 d 0 c C

ooo w "-4 co 41 w -
C4-4 0 0 3- C C a
-4 w 0 C 0 u 4)

S004r-4 000$ 1-4 0
w0 000 00 o
00 u4 4 o 5 0

HH0 CO CO C 0, 0
CO Q -0 1 wo | N w l
> OO 4 0 W a
S3 4 0 o 00

44 co w
0 v0 4 C,40 4 o
00 0 w 0n 0 0a
a w wa 3o 4.

04t40 %0 '0a04
0 4 4 0 3 UC 4 a) 41

O63 Cw O w 4- > oi ODi
64 0 ) 3 O)
0 w co tB H0 0

W rC 4 4 04 -r O &
S0 w) 1w r ) u

O-4 a 4 w % O
u 3a aC 4c a o0
ca 00 0 m i 0 CO

w0 co 0d w S0 0w 0 41 0 a 0 1 -

40 0 ) w o 0 o 0
C0 ( 3 -H 3 0) 4 1 4 0

n a41 u a a -H r co
0h w w 'o 41 '.4 u co ur

V 0 0 r- W UD W r4 v-4

O u w a o O $o
o D To 0. ( H ca

0 M > w C CO 7 C
Q)41 -0 -H 0 1-4L r a 0
a j 41 z % 0 o 4

600C- 4 00 0c 4
H0 4l O i0l r0 BO a) O-a0

"0 01 O1 0 0 w0
0r p 0 a 0 0 4 0 II w
>1 3 43'-H O O c 0

04.0 4^ 0 .0 0 B 0a 0. u-.- ^u

0 0 0
W r (M






10O1j. o 0ua0jad

experiment, an activated receptor preparation was run on identical

columns under indentical conditions. Unlike the unactivated form,

however, the activated form was also very strongly (but not completely)

retained when eluted with 150 mM KC1 and again (but to a slightly lesser

degree) when eluted with 300 mM KC1. Elution profiles with 600, 900 and

1,200 mM KC1 were all rather similar.

For the purpose of comparing the hydrophobic properties of the

different forms of the glucocorticoid receptor to those of other well

studied proteins, the next two experiments examined the hydrophobicity

of two purified proteins, bovine serum albumin (BSA) and immuno-gamma-

globulin (IgG). Each of the two proteins were first radiolabeled with

[14C]formaldehyde (Rice and Means, 1971) and then run on a Shaltiel

series using 600 mM KC1 and 50 mM molybdate. All other conditions were

identical to the previously described experiments for glucocorticoid

receptors being run on a Shaltiel series under these same salt

conditions. Like the unactivated glucocorticoid receptor, BSA exhibited

some slight retention by the propyl, butyl, pentyl, hexyl and phenyl

agaroses and was totally retained by the octyl, decyl and dodecyl gels

(Figure 2-4). Like the unactivated receptor, BSA exhibited some slight

retention by the propyl, butyl, pentyl, hexyl and phenyl agaroses and

was totally retained by the octyl, decyl and dodecyl gels. The chromat-

ographic behavior of IgG was significantly different from that of BSA on

several of the gels in the series. Of particular interest were the

differences between the two proteins on pentyl, octyl, decyl and dodecyl

agaroses. The pentyl agarose exhibited a higher affinity for the IgG

than the BSA although both proteins did elute from the column. The

biggest difference occurred on the decyl agarose where the BSA was

o0 0 0
3 0 E-4 a
Lj 04-1 00

(U 4 0 1 0

ua oa 5 a0
j M II
0 1 0 0 S U
. Cn o to *0 v 0
03 w u cc
SP4 .0 0 0 0 I
u, 4 L u 0 o0
4 r-4 AJ 4 1 00
u 0 0 (n

00 $wrA 0 e .
0 r-4 0 0 0 0
o -4C o -0 11
0 0 $4 4
l 4. T3 S i i l :3 O
r. 0 l 0 0 O 00
CCO -i O C
4 )4 cc 4 0) 0 C 0 (0
- C CD 0 4
0 i C 4 U 4 0
4u a 4-4 U
o W3 4 00 0 0 D
SC a 0C c 0 C0
o Fa o 4
a) 4 u w o w o
000 O0 '0

:3 1> r Co
co r- a 0 b

0 r"A c r-4 C a 1

Q.4. ) 0 d 0-'-44- i' w
0 1 1 0 41 0
* 0 4-13 0 C' C

00.000000 0
0e A *r C CC 4 ?
O 0 4-1

0 0-40 000
CM C a a *04- a

.9 4 cd
0r c0o 0o 0
SCO 1 4 to

C43 0.0-0 0
4-1 a c0 CB *

S10 C r-4 M 6 0M
*r4 CO 0, ) 0 0

r-4 0 -.1 AJ W $4
C0i0 0 0 0mC C

SC O 0 0 4-0 B

u. >i: r O O

o Z 0044-0 0 (D
(0 u a a a 41 00

r-4 H ).0 0 011i
o 04 -- a
T I 0 o aw 0 0
0 cr- M *4C A

1M 0 5 -, L (D0

w 0 -H 4o4 w. 0-4 -4 1o
S -H -C-405-4 00
4 t d 3 0 0. a a

T 6 *r a

0 0 Q 0 U I (

Pr4 U W r l t U 0 -l--

+ +

I 4

0 0 00 0

0 0 0 0 0 0

IOtOiLjo tuaJad

0 0 0









totally retained but the IgG eluted in a pattern similar to the propyl

and butyl agarose profiles. Also of interest was the unexpected finding

that IgG exhibited a greater affinity for hexyl agarose than for pentyl

agarose and a greater affinity for decyl agarose than for octyl agarose.

The effects of changing ionic strength on these profiles was not


In an effort to improve resolution of the chromatographic profiles

of the activated and unactivated forms of the glucocorticoid receptor as

well as to increase separation between the 2 forms, these preparations

were run individually on pentyl agarose columns much longer than those

used for the experiments involving the complete Shaltiel series (7 ml as

opposed to 1.5 ml). These preparations were run on columns equilibrated

and eluted with 600 mM KC1 and 50 mM molybdate and fractions of 0.5 ml

were collected so that fraction size was smaller relative to gel volume,

thereby increasing resolution further. As the results illustrate

(Figure 2-5), the unactivated receptors were represented by the sharper,

faster eluting peak, whereas the activated receptors were represented by

a broader, more slowly eluting peak. Although a much high resolution

was attained, the two different profiles didn't change much relative to

one another when compared to their relative shapes and positions when

run under identical buffer conditions on the smaller pentyl agarose

columns. In other words, the separation between the two peaks was not

significantly increased by increasing column length and there was still

a significant amount of overlap between them. In an attempt to increase

the separation between the two peaks, the experiment was repeated under

identical conditions except that the KC1 concentration was lowered from

600 mM to 400 mM since the affinity of activated receptors for pentyl

4O 1 C
41V C
0 C

0 a 0 ) a )

w 000 Qw 04
oil :O s1
o P-4 U H m
0 3 4 4 a

4.4-1) ) 4-4 )- 0 r. 'V
W.1 4 u E o

10 0 c 1 > 0
W0 0 ca c

u a 0 0 Url C-4

Sv -4 0 -4 0 4 0 4

-I 0 4 r-4 = v 4 %.0 0 Q
.iCi$i ) 0 J C4i G

to W 0 C04-4
4 0 0 1 W
E- V3 rA 0

12 CB a a c

W C*4 :3 4 00 w0
M Os 00 l

ci (B 3 4 -4
13 0 t3 0 0- 41
Q) i-I 44 CO "0 3 W 4) C i-H
4" 0 C C4 0 o 3t
co WO r-O cc w 10 -H
--4 4J O-H 9 C
4 c4 4 o 4C a

Cnr= ca r4 4J 0-
O-i O) 005 C lC U,-0

O ( r-4 3 Cli 0) ) "-

s0$ o-i 3
0 4J 0 0. r 4

0 3 C4 Ai 0 .0 0 -4 W
0 a Ca 00

4 t> 4 0 4a O U 0-.
4- c c w co ? Q 40
O r400 0 CU0C$

4 040 O w 0

OB0dc .0 C U3 *
U 0 ,0 0 w04
u4 O >4 04 0
6u0 CU OM CU

0 W Q) > % Q0 e 0 r

0 O 4J u 0 4J C
0 C0I a a 4
w r-4 44 40J 1 3 -H 4) I3
04 3 043 0 u 41
33 C 0) cc '0
0 s0 C0 0a) 4 0 C4
0 CC 1 $4 -T = g
tn 0 OC 4 0J 4
I co o 4 e 41
0 C ) 0 0 0 1 U Co

(U) 4J 0 ) 4 "oa S -

SU 4J CO 0. .
00 u r-I 0w 3
O- 0 0 w "4 C4 w C

Sa W rSw u c 4J "o co 41



O 0



ID.OI "O ru a3lac


IDIQI 0 4U93)J9d

agarose had previously been shown to increase to a greater extent than

the affinity of pentyl agarose for unactivated receptors when lowering

the KC1 concentration from 600 mM to 300 mM. The results were that

under the 400 mM KC1 elution conditions, the activated peak was broader

than with 600 mM KC1 elution while the unactivated receptor peak did not

change significantly (data not shown). Overlap of peaks was still


To more closely investigate the relationship between column shape

and elution profile on hydrophobic interaction chromatography columns,

the next experiment examined the profiles for both the activated and

unactivated forms of the glucocorticoid receptor on two dramatically

different shaped pentyl agarose columns with the same gel volume.

Although two different sizes of columns had been used previously, the

shape difference (ratio of gel length to width) were not great and

differences in gel volume, relative sample size and relative fraction

size made comparisons of relative chromatographic resolution difficult.

In this experiment, the gel volume for both columns was 7 ml, but the

length to width ratio for one column was 18.6 (this column was identical

in dimensions to the 7 ml column described in the previous experiment)

while the length to width ratio for the second column, which was

significantly shorter and rounder, was 0.5, but, as previously

mentioned, contained an identical volumn of gel. There was therefore a

37.2-fold difference in the length to width ratio of the two columns.

The size of the sample applied to each column as well as the size of the

fractions collected from each column were identical and the elution

buffer contained 600 mM KC1 and 50 mM molybdate. The results of this

experiment (data not shown) indicate that there were no major

differences in the overall chromatographic profiles (position of peak

fraction, etc.) obtained from the two columns for either the activated

or unactivated forms of the receptor although there appeared to be a

very slight decrease in resolution (sharpness of peak) on the short wide

column for the unactivated receptor. This slight decrease in resolution

may have been related to technical problems in evenly applying such a

small sample volume to such a large gel surface.

Since most of the work to date characterizing the glucocorticoid

receptor has involved steroid labeled forms, little is known about the

physicochemical properties of the unoccupied glucocorticoid receptor.

Since hydrophobic interactions are thought to account for the high

affinity binding of glucocorticoids to their receptors (Wolff et al.,

1978; Alfsen, 1983; Eliard and Rousseau, 1984), the empty hydrophobic

pocket exposed on a receptor whose binding site is unoccupied by a

steroid ligand may have an effect on the overall hydrophobicity of the

receptor that could have functional implications in vivo. This

experiment examined the hydrophobicity of the unoccupied glucocorticoid

receptor using a Shaltiel series of alkyl agarose and phenyl agarose

columns under experimental conditions identical to those previously

described except post-, instead of pre-, labeling was required (Figure

2-6) for analysis of activated and unactivated receptors when 600 mM KC1

and 50 mM molybdate was used. This allowed for direct comparison of the

results between all three forms of the receptor. Basically, the

unoccupied receptor behaved almost exactly like the unactivated occupied

receptor. Only slight differences between the two were detected on

profiles from the pentyl agarose columns. To determine if the slight

difference seen on this column was meaningful, unoccupied receptor

S 0 00 r- u
0 a 24-4 w
0o z I (A
C 3 3 CB l C II

o *0 0 0 0
C C C1 r-4l 44 P4 un 1 U 0
w 0 Q0
4000 u o0 (O

o C d U 0 u 4 00 4a

1 O ) 4 0) U 0 0a

uU WCA 44 a

-1 o C 00 c o 0
o r 10 93. CO
.O 3 C 0 ccc0

Hr 40W-0 M U r- i -I O U -0
41 co co Q a i 0 0u C>

0. w mo ds3 L 0o
0% U W *o a ) 4--4 (i a

O' S 0 ( 0'0
COi -14 t -4 Q) o ) -
0 c$.(3U C^OCUo
ru w ; oa
1C G3 C a 4- 1 3 e.

4a C cu- a rl -40 u
0 a a w o cn ca w o

4 cao w r4 iU0 o
a. a O) 0 r a
> 4) 0 1 HW 10

0E- 1 4Ji 4-4 C:6
4 a4 w g (1 0
Sa T ^ .ri T3 u (a, II

$w :j o o c a>
U C l T^ 9l (B r iI

=0 3 4J v C.O.4
0.3 C 0 0 4 3

-Cd 0 0-4l U 0 0 0 0
W -J-. r-w4 o 0 to Au cU

uHr4Wco*W O OW
" a m *o c 0W. 0r.

1..04 01..j-O 1.JCU03 0
u00 3 0) a i w QU 4i a)

004- C -4 0 0 w iao
00 $4 4 0 II

40 0-S W 441 0 W -. C0
0 0 3 0 0 0 1o

V4 o o O- 04 0
r0 c-4 0) -40
oco q 3 -i 11 (>%o a

3 a) c C)i4 W f 0, t

$ 0 0. i 0 -r W C- U
0 U o r U-) 0 ao 0 C C4 0 w 0
CU 0 C 3 P D 0 0 r w 1 10

c4 WO OH a
3 0 0 4 04 41 6W 0
4XU H 0 W t U 4B W
0 4( Q a a o P- C 0 !
00 E4 J 41 4) l 0 31 cc

Po r4 -) s4 S 5 -9 0 0 g Cn

0 0 0
, 0C

0 0 0 0 0 0
I0 o U (D N






o i

(0 z


N Ll-

preparation was run on the long (7 ml) pentyl agarose column previously

described for higher resolution analysis of the activated and

unactivated forms. The chromatographic profile obtained for the

unoccupied receptor turned out to be virtually identical to that

obtained for the unactivated steroid-labeled form (data not shown).

The possibility of using hydrophobic interaction chromatography as

a means of separating bound from free steroid for glucocorticoid

receptor binding assays was investigated. The results of this

experiment may also prove useful in interpreting data from experiments

where some degree of dissociation was known to occur on the hydrophobic

columns. This experiment examined the chromatographic profiles of free

[3H]TA on a Shaltiel series of columns equilibrated and eluted with 600

mM KC1 and 50 mM molybdate under conditions identical to those used for

the hydrophobic interaction chromatography of various forms of the

glucocorticoid receptor. Results indicate (Fig. 2-6) that the chromato-

graphic profile of the free steroid on propyl, butyl, pentyl and hexyl

agarose was identical to that for the free steroid on unmodified

agarose. There was a slight increase in the affinity of [3H]TA for

octyl and decyl agarose and an even greater retention by dodecyl and

phenyl agarose. It should be noted that all of the steroid applied to

each of the columns was eluted by the seventh or eighth fraction. When

compared to the profiles for glucocorticoid receptors, the free steroid

profile overlaped these profiles in virtually every case. When free

steroid was run on a long 7 ml column identical to that previously

described, it eluted about midway between the peaks for the unactivated

and activated receptor and overlapped both, demonstrating that

hydrophobic interaction chromatography under the conditions used in

these experiments is less suitable for bound/free steroid separations

than chromatography on unmodified agarose.


The results of this study indicate the presence of hydrophobic

regions on the different forms of the glucocorticoid receptor in brain.

This is in agreement with the findings of Lamb and Bullock (1983) for

progesterone receptors in uterus and Bruchovsky et al. (1981) for

androgen receptors in prostate. Even more important is the fact that

the degree of surface hydrophobicity of the glucocorticoid receptor

complex-(assumed to be directly related to the degree of retention by

the hydrophobic gels) appears to be dramatically increased upon heat

activation of the receptor to the nuclear binding form. While such a

change in the surface hydrophobicity of glucocorticoid receptor

complexes associated with activation has been suggested by other workers

reporting an increase in the partitioning of these complexes after

activation into the less polar polyethylene glycol (PEG) layer in an

aqueous dextran-PEG two phase system (Andreasen, 1978 and 1982;

Andreasen and Mainwaring, 1980) and by this lab (Luttge et al., 1984d),

which reported a large increase in the fractional adsorption of

glucocorticoid receptor complexes to glass fiber (Whatman GF/C) filters

following heat activation, hydrophobic interaction chromatography has

never been used to study this phenomenon in glucocorticoid receptors.

Furthermore, although it is possible that the dramatic increase in

surface hydrophobicity associated with heat-induced activation could

represent an artefact of the nonphysiological conditions existing during

the in vitro activation process, this increased hydrophobicity is

nevertheless consistent with the presumed function of glucocorticoid

receptors in general. It is certainly possible that the increased

hydrophobicity of the activated receptor could account, at least in

part, for the increased affinity for a number of different nuclear

components including chromatin (Sakaue and Thompson, 1977; Simons et

al., 1976), nucleosomes (Climent et al., 1977), DNA (Baxter et al.,

1972; Rousseau et al., 1975; Sluyser, 1983), RNA (Tymoczko and Phillips,

1983), the nuclear matrix (Buttyan et al., 1983; Kirsch et al., 1986)

and the nuclear membrane (Smith and von Holt, 1981). If the unactivated

receptor resides in the cytosolic/cytoplasmic compartment, as originally

thought (Szego, 1974), activation-induced increases in hydrophobicity

might facilitate the permeation of the receptor through the nuclear

membrane. If, on the other hand, the current controversial theory

stating that unoccupied and occupied unactivated receptors actually

reside in the nucleus in vivo (for review see Walters, 1985) is true,

then increased hydrophobicity could decrease the likelihood of "leakage"

of activated receptors from nuclei by virtue of their greater affinity

for nuclear components.

A Shaltiel series of alkyl agarose, as well as phenyl agarose,

columns were used since the ideal gel(s) to study the surface hydropho-

bicity of the various receptor forms under different buffer conditions

was unpredictable initially. Although it has been shown that the

hydrophobicity of a linear aliphatic carbon increases linearly with

carbon chain length (Tanford, 1972), it has been argued by others

(Shanbhag and Axelsson, 1975), that the effective hydrophobicity depends

also on the flexibility of the hydrocarbon chain and consequently on the

degree of interaction within such chains, particularly for long chains.

The mechanisms by which proteins are adsorbed to hydrophobic matrices

are no doubt complex and this problem can be exemplified in the present

study by the case of the adsorption of IGG on a series of alkyl agarose

columns, in which case a decrease in adsorption between pentyl and hexyl

and between octyl and decyl agarose. This is similar to the case of the

adsorption of erthrocytes on a series of alkyl agarose derivatives,

wherein a decrease in adsorption between hexyl and octyl agarose was

noticed to occur without an obvious explanation. The present study

found that bovine serum albumen had a higher affinity for pentyl agarose

than did IGG, whereas the reverse was true for octyl, decyl and dodecyl


Another problem adding to the unpredictable nature of hydrophobic

interactions stems from the fact that some hydrophobic ligands denature

proteins through a "detergent-like" action (Hofstee, 1973). This was

evidenced by the fact that what initially appeared to be a partial

elution of both activated and unactivated receptors from octyl, decyl

and dodecyl agaroses, later proved to be dissociated free steroid

resulting from the denaturation of the receptors. The degree of

denaturation increased with alkyl chain length with dodecyl agarose

producing the greatest effect. In addition, the degree of denaturation

for activated and unactivated forms appeared to be roughly equivalent,

implying the possibility that the disruptive hydrophobic interaction

leading to loss of steroid binding involved a region or regions common

to both forms of the receptor. This finding may be relevant to the

interpretation of other studies involving hydrophobic interaction

chromatography. For instance, Lamb and Bullock (1983) reported that a

large volume of buffer was required for elution of the progesterone

receptor from decyl agarose, which they claimed was probably due to slow

equilibrium of the column with elution buffer. Another possibility is

that the progesterone receptors tightly bound to the decyl agarose may

have been undergoing a slow denaturation resulting in broad profiles of

eluted radioactivity representing free steroid only. When dissociation

was taken into account in the present study, no steroid receptor

complexes could be eluted intact from octyl decyl or dodecyl agarose

using any of a number of buffers including some with increased "hydro-

phobic character" such as glycerol or ethylene glycol. Thurow and

Geisen (1984) have reported that polypropylene glycol/polyethylene

glycol block polymers prevent both the adsorption of dissolved proteins

to hydrophobic interfaces and the resultant aggregation and denatur-

ation. However, it is not known whether or not these polymers can

successfully elute the intact receptors from highly hydrophobic

interfaces if they are already attached and the presence of such

polymers prior to any hydrophobic interactions would reduce or eliminate

any such interactions, rendering the whole procedure useless.

Overall, the most promising results were obtained with the pentyl,

hexyl and phenyl agarose gels. Phenyl-agarose has reportedly been

effective in hydrophobic studies of the progesterone receptor (Logeat et

al., 1981; Lamb and Bullock, 1983) and in differentiating between the

activated and unactivated forms of the estrogen receptor (Gschwendt and

Kittstein, 1980). The activated form of the glucocorticoid receptor

bound more tightly to phenyl agarose than to alkyl agarose (excluding

octyl or longer alkyl agaroses). When calculated as the change in

standard free energy on transfer from an aqueous solution to a solution

of pure hydrocarbon, a phenyl group is less hydrophobic than a pentyl

group (Rosengren et al., 1975; Lamb and Bullock, 1983). A portion of

the unactivated receptor was retained by phenyl agarose under high salt

conditions, while the same receptor preparation was only slightly

retarded on the pentyl and hexyl agarose columns. While the activated

form of the receptor eluted much more slowly from the pentyl and hexyl

agaroses than did the unactivated form, the activated form was

completely retained on the phenyl agarose under low and high salt

conditions. Shaltiel (1974) has proposed that adsorption requires the

interaction of an alkyl residue of sufficient length with a hydrophobic

"pocket" within the protein. It has been suggested that the aromatic

ring on phenyl agarose may be more complementary than alkyl groups to

the contours of the hydrophobic "pocket" on the progesterone receptor

(Lamb and Bullock, 1983). Because of similarities in amino acid

sequence between the two receptor types (Conneely et al., 1986), such a

pocket may also exist in or on the glucocorticoid receptor and may be

made more accessible for hydrophobic interactions by heat activation.

Alternatively, the receptor may form a charge transfer complex with the

phenyl group, thus increasing the interaction. One might therefore

conclude that phenyl agarose is the best gel for differentiating between

the activated and unactivated forms of the glucocorticoid receptor based

on differences in surface hydrophobicity. Unfortunately, the affinity

of the activated receptor for this particular matrix prevents removal of

the receptor from the matrix in an intact form. In addition, minor

changes in hydrophobic interactions between the activated receptor and

the matrix (i.e., as caused by changes in ionic strength, temp., etc.)

cannot be readily monitored by phenyl agarose and therefore alkyl

agaroses are best suited for such studies.

A number of factors in addition to the hydrophobicity of the

attached groups influence the adsorption of receptors and other proteins

to hydrophobic matrices. Jennissen (1976) has presented evidence in

favor of the idea that the adsorption of proteins to alkyl agarose

derivatives takes place at a critical alkyl group density. In other

words, this hypothesis considers that the protein needs to present

multiple attachment points in order to be adsorbed by a determinate

member of the series of alkyl agarose derivatives. Another factor

affecting the chromatographic behavior of a given protein on a

particular hydrophobic matrix is the overall protein concentration

applied to the column. Small amounts of protein bind more homogeneously

on a particular column than a large amount. This should not have been a

factor, however, in the differences in hydrophobicity exhibited by

activated and unactivated receptors in this study since great care was

taken to maintain the same protein concentration in all treatment groups

and throughout all experiments. On the other hand, this could very well

have been a factor in the chromatographic differences seen between the

unpurified activated receptor and the purified activated receptor since

there were hundreds- to thousands-fold differences in protein

concentration (discussed in more detail in a later section). Another

way in which protein concentration, especially in impure, crude

preparations such as cytosol, could potentially affect the chromato-

graphic profile of glucocorticoid receptors on these columns is through

indirect, or secondary, interactions with other proteins bound tightly

to the columns.

The influence of salt on the retention of receptors and other

proteins to hydrophobic gels is probably due to a number of factors

acting on the protein as well as on the hydrophobic matrix. Pahlman et

al. (1977) concluded that salting-out ions, such as potassium chloride,

cause conformational, but not structural, changes in biomolecules. The

"structure forming" properties of salting-out ions enhance intra-

molecular, as well as intermolecular, hydrophobic bonding as reflected

by a stabilization of the hydrophobic core of the biomolecule (Hofstee,

1975). High concentrations of salting-out ions can adversely affect the

solubility of a protein by decreasing the availability of water

molecules in the bulk and increasing the surface tension of water,

resulting in an inhancement of the hydrophobic interactions (Tanford,

1968; Melander and Horvath, 1977). It has been recently shown that the

free energy of a protein is increased by addition of salting-out ions

and this unfavorable free energy is smaller for the proteins bound to

columns because of their smaller surface area exposed to solvent

(Arakawa, 1986). The present study used, for the most part, a

combination of two salts: sodium molybdate and potassium chloride. The

50 mM concentration of molybdate was required to prevent activation and

to increase stability of all forms of the receptor under conditions

requiring high concentrations of potassium chloride since lower concen-

trations of molybdate have been shown to be inadequate in preventing

activation under high salt conditions (Weatherill and Bell, 1985). It

was therefore not feasible to run meaningful experiments in the presence

of potassium chloride only, although the possibility exists that the

presence of molybdate may have influenced the effects that potassium

chloride would have had (aside from increasing activation) on protein

conformation and therefore, possibly, hydrophobic interactions. When

the effects of salt concentration on the chromatographic profiles of

both activated and unactivated receptor complexes on pentyl agarose were

examined, it was clear that all of the gels tested except phenyl agarose

exhibited a greater affinity for both forms of the receptor under low

salt conditions than under high salt conditions. Even the unmodified

agarose exhibited a slight affinity for the activated form of the

receptor under these conditions which may be due to low level

electrostatic or other interactions with the gel matrix itself (Janson,

1967). Since this was opposite of what was expected, a more detailed

study was carried out using only pentyl agarose and increasing

concentrations of both molybdate and potassium chloride. Elution of

activated receptors from the pentyl agarose required a higher

concentration of potassium chloride than was required to elute the

unactivated receptors. Although electrostatic interactions with the

matrix (the alkyl groups are supposedly uncharged after attachment to an

agarose matrix (Rosengren et al., 1975)) probably contributes to the

higher affinity at lower salt concentrations, the effects observed are

likely to involve a combination of electrostatic and hydrophobic

interactions since the activated receptor elutes from ion-exchange

columns at a lower salt concentration than the unactivated form (Sakaue

and Thompson, 1977). Once the electrostatic interactions are overcome

by using 600 mM concentrations of potassium chloride or greater, the

chromatographic differences between the two forms are probably purely

hydrophobic. It was reported by Peale et al. (1985) that the assayed

levels of charged estrogen receptors, using both Sephadex gel filtration

or dextran-coated charcoal treatments, in buffers containing 150-200 mM

ionic strength (roughly physiologic) was approximately one half that of

estrogen receptor levels assayed in buffers either at 0-50 or 400-450 mM

ionic strength. However, molybdate was shown to eliminate this


Although Wolf et al. (1978) have shown that the high affinity

binding of a glucocorticoid by its receptor is largely due to hydro-

phobic interactions, the present study found little evidence that the

primary site of the hydrophobic interactions with alkyl and phenyl

agarose involved the steroid binding site. In fact, the chromatographic

behavior of the unoccupied receptor differed little from that of the

occupied unactivated receptor on practically all of the gels examined.

These findings are in agreement with those of Lamb and Bullock (1983)

who found that progesterone receptors bound to, and eluted from, alkyl

agarose equally well in both the free and completed form. This, in

fact, might be expected in light of the high degree of specificity

normally associated with steroid-receptor interactions. It has been

implied, however, that steroid binding might be involved with some

hydrophobic interactions between the rat estrogen receptor and Cibacron

Blue (Tenenbaum and Leclerq, 1980).

In an attempt to improve the chromatographic resolution of the

activated and unactivated glucocorticoid receptors eluting from alkyl

agarose columns, pentyl agarose columns with various shapes (but with

the same gel volume) were compared. In addition, elution flow rates

were varied. None of these changes, however, resulted in any

significant improvement in chromatographic resolution providing further

support to the early observation by Frank and Evans (1945) that the

hydrophobic interaction is a spontaneous process.




The concentration of glucocorticoid receptors is subject to

fluctuations.resulting from several factors (De Nicola et al., 1982;

Svec, 1985a), including autoregulation by the ligands themselves

(Tornello et al., 1982; Ho-Kim et al., 1983; Muramatsu et al., 1983;

Sapolsky et al., 1984; Svec, 1985b; Luttge et al., unpublished). The

glucocorticoid-induced depletion of glucocorticoid receptors is

reportedly due to an increased degradation of receptors and not due to a

decreased rate of receptor synthesis (Svec and Rudis, 1981; McIntyre and

Samuels, 1985). Androgens, on the other hand, have been associated with

an augmentation of their receptors, apparently resulting from an

increase both in receptor half-life and in rates of receptor synthesis

(Syms et al., 1985). Glucocorticoid receptors may require an

energy-dependent up-regulation or transformation before they are able to

bind the steroid. Evidence for such an up-regulation in brain was

presented by Turner and McEwen (1980), who reported a lack of depletion

of hippocampal cytosol glucocorticoid receptors after multiple

injections of [3H]corticosterone. These results could not be explained

entirely by receptor synthesis and suggested that in the hippocampus an

"excess" of glucocorticoid binding proteins exist, and that the

availability of functional cytosol receptors may be regulated to

maintain a relatively constant cellular level. Evidence of a rapid

down-regulation has also been observed. Recently, the loss of a

biological response associated with such a down-regulation of

glucocorticoid receptor levels was demonstrated in a mouse cell line

(Danielsen and Stallcup, 1984). In addition to circulating levels of

steroids, other factors may play a role in the regulation of steroid

receptor concentration. Of particular relevance to the role of

glucocorticoids in the nervous system is recent evidence that neural

input influences significantly the levels of both androgen (Bernard et

al., 1984) and glucocorticoid (DuBois and Almon, 1981) receptor levels

in striated muscle as determined from denervation studies. Similarly,

Cardinal et al. (1983) reported that pineal cytoplasmic and nuclear

estrogen and androgen receptors are modulated by norepinephrine released

from nerve endings at the pinealocyte level. These findings suggest that

target cells may utilize a variety of homeostatic mechanisms to regulate

the amount of receptor capable of interaction with the free steroid.

Little is known, however, about these regulatory mechanisms and even

less is known about the quantification, production and turnover of these

cryptic (down-regulated) receptors. In addition, factors which

stabilize or destabilize unoccupied receptors under in vitro conditions

may also provide valuable information concerning the receptor and its


One potential mechanism of receptor up- and down-regulation for

which there is fairly substantial evidence involves the reduction and

oxidation of sulfhydryl groups on the receptor. Oxidation of receptor

sulfhydryl groups (probably located in or near the steroid binding site)

has been shown to reversibly inactivate glucocorticoid receptors in a

number of tissues (Rees and Bell, 1975; Granberg and Ballard, 1977;

Sando et al., 1979; McBlain and Shyamala, 1980; Housley et al., 1982;

Harrison et al., 1983; Densmore et al., 1984d; Izawa et al., 1984).

These sulfhydryl groups are probably located in or near the steroid

binding site as evidenced by the fact that mesylated derivatives of

glucocorticoids have been shown to bind covalently to the receptor,

presumably by means of an interaction between the mesylate group and a

binding-site sulfhydryl group (Simons et al., 1980). The affinity

labeling of the estrogen receptor has similarly indicated the presence

of a thiol group at the binding site that appears to be directly

involved in estradiol-receptor binding (Ikeda, 1982). Addition of

sulfhydryl reducing agents such as dithiothreitol (DTT) or

2-mercaptoethanol leads to a total recovery of lost binding under

appropriate conditions. Further evidence that the sulfhydryl-disulfide

balance may be important in the regulation of glucocorticoid binding

capacity was presented recently when it was shown that endogenous

thioredoxin was responsible for up-regulation of glucocorticoid

receptors in liver cytosol (Grippo et al., 1983). In addition to

effects on binding capacity, oxidation of yet other sulfhydryl groups on

the activated glucocorticoid-receptor complex by various sulfhydryl

reactive agents has been shown to interfere with the DNA-binding ability

of receptors from thymus (Bodwell et al., 1984) and most recently

Kaufmann et al. (1986) has presented evidence suggesting that disulfide

bonds mediate the association between the activated glucocorticoid

receptor and the nuclear matrix, representing yet another possibility

for the regulation of glucocorticoid action. Much work in

this area remains to be done, however, before it can be determined

whether these mechanisms are likely to be of significance in vivo.

Despite the growing body of evidence that ligand binding to the

glucocorticoid receptor as well as binding of the activated gluco-

corticoid-receptor complex to nuclei or DNA are dependent upon the

reduction of certain sulfhydryl groups on the receptor surface, there

has been surprisingly little speculation concerning the possibility of

thiol/disulfide exchange as a control mechanism for glucocorticoid

action in vivo. The specific modulation of enzyme activity by post-

translational modification of specific amino acid side chains is well

established as a mechanism of metabolic regulation. Essentially all the

amino acids possessing chemically reactive side groups are capable of

undergoing some sort of posttranslational modification: the carboxyl

groups of glutamate and aspartate, the imidazole side chain of

histidine, the hydroxyl groups of serine, threonine and tyrosine, the

amine group of lysine and even the guanidino group of arginine.

However, the sulfhydryl group of cysteine is potentially the most

reactive nucleophile in proteins. The process of thiol/disulfide

exchange could provide a mechanism for the equilibration of the

sulfhydryl oxidation state of proteins with the thiol/disulfide status

of the surrounding environment. Although the regulation of enzyme

activity by thiol/disulfide exchange was proposed some time ago (Barron,

1951), it has received less attention than other posttranslational

modifications. According to Gilbert (1984), thiol/disulfide exchange as

a biological control mechanism must exhibit at least the following six


1. The intracellular thiol/disulfide status should vary in vivo in

response to some metabolic signal. Isaacs and Binkley (1977a) reported

that the disulfide-sulfhydryl ratios of subcellular (nuclear, microsomal

and cytosolic) fractions of rat hepatic tissue varied diurnally with the

ratio lowest in the early morning (after feeding) and highest in the

early evening (after fasting). The primary reaction was the reversible

formation of mixed disulfides of glutathione with proteins and the

change in the overall disulfide-sulfhydryl ratio was approximately

4-fold. These same workers (Isaacs and Binkley, 1977b) also reported

changes in the disulfide-sulfhydryl ratio associated with cyclic AMP

levels. Intaperitoneal injections of dibutyryl cyclic AMP induced an

increase in hepatic glutathione protein mixed disulfides combined with a

corresponding decrease in reduced glutathione and protein sulfhydryl.

2. Oxidation of protein sulfhydryl groups by thiol/disulfide exchange

should activate some enzymes or receptors, inactivate others and not

affect the activity of still others. Gilbert (1984) has compiled a

large number of enzyme activities that have been suggested to be

modulated by reversible thiol/disulfide exchange. Like some steroid

receptors, many enzymes are reversibly inactivated by oxidation with a

number of biologically occurring disulfides. One example is

phosphofructokinase from either rabbit muscle (Gilbert, 1982) or mouse

liver (Binkley and Richardson, 1982). Fructose-1,6-biphosphatase,

however, undergoes a 2- to 4-fold activation in the presence of such

disulfides (Gilbert, 1984).

Many enzymes have sulfydryl groups that can be chemically modified

by a variety of sulfhydryl-specific reagents such as iodoacetamide,

N-ethylmaleimide or Ellman's reagent with a resultant effect on enzyme

activity. Most of these reagents are highly reactive toward sulfhydryl

groups and somewhat less specific as a result. In addition, these

reactive disulfides are much better oxidizing agents than most

biologically occurring disulfides. The enzymes aldolase and thiolase,

for example, both have sulfhydryl groups that react with iodoacetamide,

N-ethylmaleimide or reactive disulfides, however, incubation with 10mM

glutathione disulfide at pH 8.0 for 24 hr (very harsh oxidizing

conditions) had only a minimal effect on their activity (Gilbert, 1984).

A similar situation has been observed with steroid receptors. Tashima

et al. (1984) reported that monoiodoacetamide, N-ethylmaleimide and

5,5'-dithiobis(2-nitrobenzoate) all inactivated the binding capacity of

unoccupied mineralocorticoid ([3H]aldosterone) receptors, while the

binding capacity of these receptors was unaffected by endogenous

oxidizing factors that lead to the temperature-dependent, reversible

inactivation of glucocorticoid ([3H]dexamethasone) receptors under

identical in vitro conditions (Emadian et al., 1985). Therefore, the

presence of a chemically reactive sulfhydryl group is not in itself

sufficient evidence to suggest regulation of any enzyme or receptor by

thiol/disulfide exchange.

3. The thiol/disulfide redox potential of a regulated protein

should be near the observed thiol/disulfide ratio in vivo. If the

thiol/disulfide status of a regulated protein is in equilibrium with the

thiol/disulfide status of the cell, the equilibrium constant for the

thiol/disulfide exchange reaction

Ered + RSSR = Eox + RSH

Keq = [Eox][RSH]/[Ered][RSSR]

must lie near the intracellular thiol/disulfide ratio (Gilbert, 1984).

According to this equilibrium model, changes in the ratio of oxidized to

reduced receptor ([Eox]/[Ered]), and hence, the binding capacity, would

result from changes in the thiol/disulfide ratio ([RSH]/[RSSR]). Such a

relationship has been observed, although the literature concerning the

thiol/disulfide redox potentials of individual proteins is very limited.

Two enzymes that do equilibrate with glutathione/glutathione-disulfide

(GSH/GSSG) redox buffers in vitro are rabbit muscle phosphofructokinase

(Gilbert, 1982) and chicken liver fatty acid synthetase (Walters and

Gilbert, 1984). Unfortunately, there are no reports concerning redox

potential measurement in steroid receptor systems. However,

intracellular ratios of GSH/GSSG, the most predominant low molecular

weight thiol/disulfide pair, typically decrease during in vitro

procedures if precautions are not taken to "trap" thiols and disulfides

at their original in vivo levels. This is because, at neutral pH, GSH

is oxidized in air to GSSG; and due to the large excess of GSH over

GSSG, postmortem oxidation of a small amount of the total GSH will

result in a large artifactual increase in the observed GSSG

concentration. One study reported an increase in GSSG concentration of

about 50% in 100 min in acid-soluble extracts of rat liver adjusted to

pH 7.4 (Vina et al., 1978). Such an increase would lead to changes in

the GSH/GSSG ratio that would still be within the range of change

observed to occur diurnally in rats (Isaacs and Binkley, 1977),

indicating that changes seen in receptor binding activity due to

postmortem oxidation of intracellular thiols might possibly reflect a

receptor thiol/disulfide redox potential reasonably near the

physiologically significant range. However, much work remains to be

done before such a question can be answered.

4. Thiol/disulfide exchange reactions must be kinetically competent

under physiological conditions. In order for the thiol/disulfide

exchange between protein sulfhydryl groups and biological disulfides to

be considered a viable regulatory mechanism, the rate of this process in

vivo must be at least as fast as the changes in the intracellular

thiol/disulfide ratio. With some exceptions, most enzymes thus far

studied would probably require catalysis of in vivo thiol/disulfide

exchange in order for the process to be considered a viable mechanism of

regulation (Gilbert, 1984). Since precise rates of thiol/disulfide

exchange for steroid receptors have not been determined under defined

conditions, it is currently unknown whether such a system would be

kinetically competent in the absence of catalysis. However, recent

reports have provided direct evidence for the existence of an endogenous

thioredoxin-mediated reducing system in rat liver cytosol that maintains

the glucocorticoid receptor in a reduced state in vitro in the presence

of NADPH (Grippo et al., 1983, 1985). As for kinetic competence of this

reaction, it should be noted that E. coli thioredoxin-(SH)2 reduces

bovine insulin at least 10,000 times more rapidly than dithiotreitol at

pH 7 (Holmgren, 1979). Endogenous mammalian thioredoxin activity has

probably been undetected in many glucocorticoid receptor studies because

of its easy inactivation by oxidation of structural SH groups (Luthman

and Holmgren, 1982). Therefore, the same in vitro conditions (oxidation

of reactive thiols) that inactivate unoccupied glucocorticoid receptors

also simultaneously inactivate the enzymes that catalyze rapid reduction

of receptor disulfides. The possible role of other enzymes that

catalyze the reverse reaction disulfidee formation) have not been

investigated with regards to steroid receptor systems. One such

possibility, protein disulfide-isomerase, has been shown to have a

rather broad substrate specificity, is widely distributed and has been

detected in most vertebrate tissues (Hillson et al., 1984).

5. For regulated proteins, the oxidized and reduced forms should

both be observable in vivo. The relative levels of the oxidized and

reduced forms must change in response to changes in the cellular

thiol/disulfide ratio. Although there is evidence that some enzymes,

such as phosphofructokinase, exist in vivo in both an oxidized and

reduced form, it is currently unknown whether both forms of the

glucocorticoid, or any other steroid, receptor exist in vivo. This is

for several reasons. First, in vitro preparation of cytosol for steroid

receptor binding studies is not an instantaneous process and oxidation,

as stated earlier, is most likely occurring if sulfhydryl group reducing

agents are not present. However, in the presence of such reducing

agents (such as dithiothreitol or 2-mercaptoethanol), the overall

thiol/disulfide ratio of the cytosol is being shifted in favor of thiol

formation. Even though nonenzymatic reduction of disulfides is

relatively slow, dithiothreitol and other exogenously added reductants

may reactivating or "up-regulating" enzymes such as thioredoxin that

rapidly catalyze dithiol-disulfide oxidoreductions (Holmgren, 1977,

1979). These problems are magnified by the fact that, like cytosol

preparation, the measurement of steroid binding capacity itself is very

time consuming. At saturating concentrations of DEX or TA, low

temperature incubations of 24 hours or longer are often required to

reach equilibrium (or near equilibrium) binding. Finally, ratios of

oxidized versus reduced receptors can not be assumed from receptor

binding studies alone. The possibility exists that the oxidation or

reduction of nonreceptor cytosolic components may indirectly modulate

receptor binding capacity via nonthiol/disulfide exchange mechanisms.

6. The response of particular enzyme or receptor activities to

oxidation by thiol/disulfide exchange should be consistent with the

metabolic function of the enzyme or receptor. If changes in the

thiol/disulfide ratio are coupled to changes in the activities of some

regulated enzymes and receptors, and if one assumes that such ratios are

based entirely on the feeding state of the animal (fasting leads to a

decrease in the ratio (Isaacs and Binkley, 1977a)), three general

consequences might follow (Gilbert, 1984). (1) Enzymes or receptors

that should be active only in the fed state should lose activity on

oxidation or gain activity on reduction. (2) Enzymes or receptors that

should be active only in the fasted state should either gain activity on

oxidation or lose activity on reduction. (3) Enzymes or receptors that

must function in both the fed and fasted state should either not undergo

thiol/disulfide exchange or the thiol/disulfide exchange should not

affect the enzyme activity. Although, as stated previously, it has been

reported that a decrease in the cellular thiol/disulfide ratio can be

induced by starvation or the administration of agents that increase the

intracellular concentration of cAMP, such as glucagon or epinephrine

(Isaacs and Binkley, 1977a,b), the effects of many other factors

(disease, stress, other hormonal systems) on cellular thiol/disulfide

ratios simply has not been investigated in great detail. Therefore,

while the relationship between some enzymes and changes in the cellular

thiol/disulfide ratio might appear to be consistent with the metabolic

function of the enzyme, the apparent absence or inconsistency of such a

relationship based merely on the feeding status of an animal should not

be taken as an absolute disqualification of redox control of a receptor

system whose functions are as vastly complex as that of the

glucocorticoid receptor. In addition, tissue differences in cellular

thiol/disulfide ratio, as well as levels of enzymes catalyzing

thiol/disulfide exchange, are known to exist. This may, in part,

explain the tissue differences reported in the dithiothreitol-inhibited

instability of unoccupied glucocorticoid receptors (Granberg and

Ballard, 1977), and may indicate tissue differences in the degree of

redox control.

Materials and Methods

Chemicals, Steroids and Isotopes

[6,7-3H]Triamcinolone acetonide, or 9a-fluoro-llb,16a,17a,21-tetra-

hydroxy-1,4-pregnadiene-3,20-dione-16,17-acetonide, ([3H]TA, specific

activity = 37 Ci/mmol) and [6,7-3H]dexamethasone, or 9a-fluoro-16a-

methylprednisolone ([3H]DEX, specific activity = 44.1-48.9 Ci/mmol) were

purchased from New England Nuclear (Boston, MA). Sephadex G-25 (fine)

and Dextran T70 were obtained from Pharmacia Fine Chemicals (Piscataway,

NJ). Dinitrobenzoic acid (DTNB), N-ethylmaleimide NEM), iodoacetamide,

iodoacetic acid, cystamine, glutathione disulfide (GSSG) and para-

chloro-mercurisulfonate (PCMS), sodium-m-periodate, sodium-p-periodate,

6,6'-dithionicotinic acid (DTNT), methyl methane thiosulfonate (MMTS),

mercuric chloride, sodium molybdate (Na2Mo04), calf thymus DNA-

cellulose, glycerol, sucrose, activated charcoal, PPO (2,5-diphenyl-

oxazole) and dimethyl POPOP (1,4-bis[2(4-methyl-5-phenyloxazoyl)]-

benzene) and pentyl agarose were all purchased from Sigma Chemical Co.

(St. Louis, MO). Dithiothreitol (DTT) and

4-(2-hydroxyethyl)-l-piperazineethane-sulfonic acid (HEPES) were

courtesy of Research Organics (Cleveland, OH). Scinti Verse II was

purchased from Fisher, Inc. (Fair Lawn, NJ). All other chemicals and

solvents were reagent grade.


All studies used female CD-1 mice (Charles River Laboratories,

Wilmington, MA) that were subjected to combined ovariectomy and

adrenalectomy approximately 1 week prior to each experiment in order to

remove known sources of endogenous steroids. Both operations were

performed bilaterally via a lateral, subcostal approach under barbi-

turate anesthesia, and mice were given 0.9% NaC1 (w:v) in place of

drinking water. On the day of the experiment mice were anesthetized

with ether and perfused slowly through the heart with ice-cold

HEPES-buffered saline (20-30 ml, isotonic, pH 7.6).

Cytosol Preparation and Steroid Binding

Brains were removed from the perfused animals and homogenized (2x10

strokes at 1000 rpm) in 4 volumes of ice cold buffer containing 20 mM

HEPES and 20 mM Na2MoO4, pH 7.6 at 0 C) in a glass homogenizer with a

Teflon pestle milled to a clearance between the pestle and homogeniz-

ation tube of 0.125 mm on the radius (to minimize rupture of the brain

cell nuclei (McEwen and Zigmond, 1972)). The crude homogenate was

centrifuged at 100,000 g for 20 min and the supernatant recentrifuged at

100,000 g for an additional 60 min to yield cytosol. During these

centrifuge runs, and during all other procedures, unless otherwise

indicated, careful attention was paid to maintaining the cytosol at 0-2

C. Final protein concentrations were typically in the 6-8 mg/ml range.

After the appropriate experimental manipulations, cytosol samples were

incubated with 20 nM [3H]TA or [3H]DEX for 24 to 40 hours at 0 C with or

without a 200-fold excess of unlabeled steroid.

Charcoal Pretreatment

The dextran-coated charcoal (DCC) mixture was prepared overnight by

adding 1.25% (w/v) activated charcoal and 0.625% dextran T70 to buffers

identical in composition to that in which cytosol was prepared. Just

prior to use, the desired volume of DCC suspension was centrifuged at

2,230 g for 10 min, the supernatant discarded and an equivalent volume

of cytosol added to the pellet. The DCC was resuspended by gentle

vortexing and the mixture incubated for 20 min at 0 C with 30 sec

intermittent vortexing every 5 min. At the end of the incubation

period, the dispersed DCC was pelletted by centrifugation at 2,230 g for

10 min and the resulting supernatant (i.e. the DCC-pretreated cytosol)

carefully aspirated for subsequent steps. Control groups underwent

identical procedural steps (in the absence of DCC) in parallel with

their DCC-treated counterparts. The thermal stability of unoccupied

receptors in each of these groups was tested by incubating the cytosol

in a water bath maintained at the desired temperature for the period of

time designated for each experiment. Following this "aging", the cytosol

preparations were returned immediately to an ice bath. The protein

concentration in such preparations ranged from 3.1 to 3.9 mg/ml (Lowry

et al., 1951).

Hydrophobic Interaction Chromatography

Unlabeled cytosol was either up- or down-regulated by an

appropriate incubation in the presence or absence of DTT, respectively

(described in greater detail later). The different cytosolic

preparations were then run on Sephadex G-25 columns equilibrated and

eluted with HEPES buffer containing 600 mM KC1 and 50 mM molybdate at pH

7.6. Of the macromolecular fraction collected from these columns, 0.5 ml

was then run on a 7 ml pentyl agarose column equilibrated and eluted

with HEPES buffer containing 600 mM KC1 and 50 mM molybdate at pH 7.6.

Fractions (0.5 ml) were collected and incubated with 20 nM [3H]DEX +/- 4

uM [1H]DEX for 40 hr at 0 C. The fractions were then assayed for

specific binding.

Sucrose Density Gradient Sedimentation

Unlabeled cytosol was either up- or down-regulated by an

appropriate incubation in the presence or absence of DTT, respectively.

Aliquots (400 ul) were layered onto linear 5-20% sucrose density

gradients (5 ml; prepared with HEPES buffer containing 20 mM molybdate

and either with (for up-regulated receptors) or without (for down-

regulated receptors) 2 mM DTT and centrifuged at 0 C for 19-20 hr at

200-234,000 g (average) in a Beckman SW 50.1 rotor. The cellulose

nitrate tubes were punctured and 28-30 fractions (180 ul) collected and

incubated with 20 nM [3H]DEX +/- 4 uM [1H]DEX in the presence of 2 mM

DTT at 0 C for 24 hr. The individual fractions were then assayed for

specific binding. Sedimentation coefficients (S20,w) were calculated

from the linear regression of S20,w vs sedimentation distance (Martin

and Ames, 1961) for the following standard proteins run in parallel

tubes: chicken ovalbumin (OVALB, 3.6 S), bovine serum albumin (BSA, 4.3

S), bovine gamma globulin (IgG, 7.4 S) and catalase (CAT, 11.3 S). The

standard proteins were [14C]methylated (for detection) to low specific

activity with [14C]formaldehyde by the method of Rice and Means (1971).

Steroid Binding Determination

In all experiments, bound [3H]-steroid was separated from free on

Sephadex G-25 columns (0.6 x 14 cm) pre-equilibrated in buffer identical

to that in which the cytosol was prepared. Duplicate aliquots from each

assay tube (BT and BNS) were layered onto separate columns, allowed to

penetrate the gel and the macromolecular (bound) fraction eluted (with

homologous buffer) directly into scintillation vials for liquid

scintillation spectrometry.


The purpose of this investigation was to determine the role of

sulfhydryl oxidation-reduction in the regulation of steroid binding

capacity of the unoccupied glucocorticoid receptor. It was therefore of

importance to first characterize the time- and temperature-dependent

changes in binding capacity occurring in cytosol in the absence of

exogenously-added sulfhydryl reducing or oxidizing reagents. The first

experiment investigated the temperature-dependent reversible loss of

glucocorticoid binding capacity in the presence of molybdate, but in the

absence of sulfhydryl reducing reagents such as DTT. Unlabeled cytosol

was preincubated or "aged" at 22 C for 0, 2 and 4 hr, followed by

addition of buffer with or without DTT prior to the steroid incubation.

The experiment also examined the effects of aging of cytosol after the

addition of DTT. Results (Figure 3-1) indicate that the binding

capacity lost at 22 C in the presence of molybdate is completely

restorable upon addition of DTT. Aging of the unlabeled cytosol after

DTT addition had little effect on the binding capacity. Under these

conditions, half of the maximal binding capacity (that measured after

DTT addition) was reversibly inactivated after less than 4 hr of 22 C

aging, although approximately 25% of the maximal binding capacity

appeared to have already been lost prior to any 22 C aging whatsoever.

The next experiment was designed to more carefully investigate the

phenomenon studied in the previous experiment, except the period of

Figure 3-1. Temperature-dependent reversible loss of glucocorticoid
binding capacity of unoccupied Type II glucocorticoid receptors in the
presence of molybdate and absence of DTT. Solid circles represent brain
cytosol prepared in HEPES buffer containing 20 mM molybdate that has
been incubated at 22 C in the absence of steroid prior to a 24 hr
incubation with 20 nM [3H]DEX +/- 4 uM [1H]DEX at 0 C. Solid squares
represent the addition of DTT (2 mM final concentration) to the aged
cytosol just prior to the initiation of steroid incubations. Open
symbols represent one additional hr of 22 C aging after the addition of
HEPES buffer containing either 20 mM molybdate only (open circle) or 20
mM molybdate plus 20 mM DTT (open squares, 2 mM final DTT
concentration), but prior to steroid incubations. Bound-free steroid
separations were performed on Sephadex G-25 columns. Specific binding
is expressed as percent of the 0 hr control group (250 fmole/mg cytosol










preincubation aging was greatly extended to determine if the reversible

down-regulation of glucocorticoid binding capacity could reach 100%

under in vitro conditions. This experiment also sought to determine if

unlabeled glucocorticoid receptors in the sulfhydryl down-regulated

state were more or less stable for extended periods at 22 C than

up-regulated unlabeled receptors. Like the previous experiment,

unlabeled cytosol prepared with 20 mM molybdate was subjected to a

preincubation aging at 22 C, but this time for 0, 2, 5, 9, 21 and 30 hr

(Figure 3-2). In addition to the nonDTT-containing groups that received

HEPES buffer with or without DTT after the 22 C aging, another group was

included that contained DTT throughout the aging step. As in the

previous experiment, suboptimal binding was encountered even prior to 22

C aging when DTT was absent from the cytosol. Down regulation appeared

to reach virtual completion at 22 C, but preincubation on the order of

30 hr was required. Maximum binding capacity (i.e., after addition of

DTT) diminished only slightly (about 30%) during the 30 hr period.

There were clearly no significant differences in binding between the

groups provided with DTT after the preincubation and those provided with

DTT prior to the preincubation.

In an attempt to further characterize the endogenous factors)

involved in the reversible in vitro down-regulation of unoccupied

glucocorticoid receptors, cytosol preparations were subjected to a

dextran-coated charcoal (DCC) pretreatment with or without DTT addition

(before and/or after DCC). The effect of temperature on this

DTT-reversible binding loss was also examined under each of the possible

conditions by aging half of the cytosol from each group for an

additional 4 hr at 22 C prior to incubation with labeled steroid.

4-4 E-4

0 CO 0 CU 0 )

,.-4 3-. 1 9 41 0 0
E- >0 -- .,4 C o
, 0 0U 0
0 I 4-4 0 co u

0 .3 40 0
U ,-4 ., .,
o -4 0 4' ,
0 0 a00 w 0 4 0 0
C6 0 cn "a w
a a1 4 B

0 0 G U 0 ca

U 0 t 4I
c0 .4 CU 4 CO I
,- ,.4 "0 4i 0 A

E-M 4 U 4 Wr 0o
0 0 ~ 0 4
'4-404 o O s-

o 44 0 3 -H 4

0.,4 -o 0 to
"4 0 0 H V V44

o0 4 (4 "-1
"-I UU 0 3
40 0 C' t 0 M )
0 CUt c' 0

0U 4-w r4-4,$-
-0 Eo a r0
C) O4 0 Q)
ua MN o a0

Sco 0 44 W W .,.I
S0) 0 E4 0 0

100 A 0 0 1 (
0 4 O 0 u 0

4 r0 04 "0 004
0 4 a o0 0 0 c

n, ) O W 3 U 00
T-a4 -1 a $4 o w i s 14 -I
0 4 a 0- 00

-40 W 4)1 4 OA
I-l a c00 0
W0 0'-'
> w0 0 04

C a 0 4 -1 0 0

1.c0 *N1 o W1o a
0 U N C0 4- 0 00
w ) u wu 0 S

0 01 4) U
0 CU a W &l 0 a 0
a (AO CO C -4

4 0 0 u $40

0) 0 4WM VC CW
I 2 -H 3 0 41 0 W
0 00 0 0 a-

A4 co ca w
ca M 0 M

00 M 0) 04-1 00
4 14 41 "- 1 4-) r-4

o 00 co <- 0 3
E- U 4 Cu 43 (3 co $
1r4 4J1 0 0 -
0 4J u aX w ca 0 m
m u n M 0. 0 C
0 -4C O- W A ) U 0
0o 4 C l >4 r-4 410 43 CO

W 0 0 T t i-I W pW 0) vd 4

IO1UOO 40o uaoJgad









1 I r .

1 I


Finally, liver and kidney cytosol preparations were treated in an

identical manner for comparative purposes since these tissues reportedly

have a higher endogenous reducing potential (Granberg and Ballard,

1977). For brain, the DCC pretreatment in the absence of DTT additions

or subsequent 22 C aging led to a significant decrease (approximately

40%) in the binding capacity for [3H]DEX (Figure 3-3). There was also

an increase in the rate of loss of binding capacity at 22 C. When DTT

was added back to cytosol after DCC pretreatment there was virtually a

100% recovery of binding capacity seen in cytosol treated with DTT but

not DCC. When DTT was added to cytosol prior to DCC treatment, there

was still a loss of binding capacity to a level similar to that seen if

no DTT was present before DCC. However, there was no significant

additional loss of binding capacity after a subsequent 22 C aging as was

the case for the DTT-, then DCC-treated group. Interestingly, when DTT

was added both before and after the DCC pretreatment, there was only a

partial restoration of the total binding capacity although there was

again very little loss of binding capacity associated with aging.

Results for liver and kidney were clearly different from those obtained

for brain and surprisingly different from what was expected based on the

findings of other workers regarding the endogenous reducing potential

for these tissues. The degree of loss of binding capacity seen in

groups not pretreated with DCC, DTT or 22 C aging was greater in liver

and much greater in kidney than that observed for brain. In addition,

the rate of loss of binding capacity for these same groups during 22 C

aging was greatly enhanced for both tissues when compared to brain.

However, similar to brain cytosol, DCC pretreatment led to a significant

decrease in binding capacity that was fully reversible upon addition of

Figure 3-3. Effects of dextran-coated charcoal pretreatment and DTT on
glucocorticoid binding capacity and thermal stability of unoccupied Type
II glucocorticoid receptors from brain, liver and kidney. Concentrated
cytosol was prepared in HEPES buffer plus 20 mM molybdate and then
diluted with either additional molybdate-containing buffer or molybdate-
containing buffer plus 20 mM DTT (2 mM final concentration) prior to a
20 min incubation in the presence or absence of dextran- coated charcoal
at 0 C. Cytosols were then further diluted with either additional
molybdate-containing buffer or molybdate-containing buffer plus 20 mM
DTT (2 mM final concentration) and then incubated at 22 C for either 0
(solid bars) or 4 (hatched bars) hr in the absence of steroid prior to
incubation with 20 nM [3H]DEX +/- 4 uM [1H]DEX at 0 C for 24 hr.
Bound-free steroid separations were performed on Sephadex G-25 columns.
Specific binding is expressed as percent of the 0 hr control groups
(300, 281, and 341 fmole/mg protein for brain, liver and kidney
cytosols, respectively) and represents the mean + S.E.M. of 5
independent replications and 3 independent replications for liver and

- + + +
+ + + +
- + +


















DTT. Unlike brain cytosol, however, both liver and kidney cytosols

exhibited a significant degree of protection from DCC-induced binding

loss when pretreated with DTT. However, DTT pretreatment preceding DCC

pretreatment of liver and kidney cytosol appeared to be less effective

in preventing further binding loss during 22 C aging than was found to

be the case for brain cytosol treated with DTT before and after DCC

pretreatment was significantly higher for liver and kidney than it was

for brain.

Since the chemical nature of the sulfhydryl up- and down-regulation

process was unknown, an investigation was made to determine if the

process involved any major, easily measurable changes in the molecular

structure of the receptor (i.e. receptor subunit association or

dissociation or major conformational changes). The aim of this

experiment was to compare the sedimentation properties of up- and

down-regulated unoccupied glucocorticoid receptors from brain using

sucrose density gradient analysis. Cytosol was up-regulated by

incubating with DTT at 0 C prior to the gradient run on sucrose

gradients prepared and DTT. Cytosol was down-regulated by incubating

without DTT at 22 C for 4 hr prior to the gradient run on sucrose

gradients prepared without DTT. All cytosol groups and sucrose

gradients contained molybdate. After an 18 hr gradient run, gradients

of both treatment groups were subjected to fractionation followed by

postlabeling of the individual fractions with [3H]DEX. Steroid

incubation conditions were identical for both groups except that DTT had

to be added back to each fraction of the down-regulated group to allow

for up-regulation so that binding could be determined. The profiles

representing the specific binding determined for each fraction indicated

that both forms sediment in one major peak of approximately 9.2 S

(Figure 3-4). The results therefore provide little evidence for any

major conformational changes in the receptor associated with sulfhydryl

regulation of binding capacity under these conditions.

To determine if the changes in glucocorticoid binding capacity

associated with sulfhydryl down-regulation involve merely a decrease in

the maximal binding or, instead, a modification of the receptor binding

affinity, Scatchard analysis of DEX binding was performed for cytosol

either containing or not containing DTT (Figure 3-5). The addition of

DTT led to an apparent increase in both the affinity and maximal

binding. Equilibrium binding constants were also determined after 4 hr

of 22 C aging of the cytosol with and without DTT prior to steroid

incubation. This aging only had a significant effect on the parameters

determined for non-DTT-containing cytosol, leading to a further decrease

in both maximal binding and apparent affinity.

If the apparent reduction in affinity associated with sulfhydryl

down-regulation as determined by Scatchard analysis is real, it is

likely to be due to either a decrease in the rate of steroid

association, an increase in the rate of steroid dissociation or both.

Each of these kinetic parameters were investigated. To study the effect

of reducing reagents on the rate of steroid dissociation from the

glucocorticoid receptor, cytosol was first prelabeled with [3H]DEX in

the presence of molybdate and DTT prior to being run on a Sephadex G-25

column equilibrated in non-DTT-containing buffer to remove both the free

steroid and the DTT. DTT was then added back to half of the preparation

while the other half was appropriately diluted with non-DTT-containing

buffer. As illustrated in Figure 3-6, there was no significant

e P to o

.0 0 0 0 r- 0 w
0 N a$ C 3 O4 IV W C
Or- O
3 O 0 tO
0 0
U 0, 0 -A 0 -I -4
u :3 Qk. 3 a41 : Q V) "
0 -4 3 0 w 1 0O

-0 r 4 0 E-i 1-

n m Ea m 9 000 0 0 O
d)a 44 o a) 0 1

0N 0 3 44 0 W3 0 W
I P. w co 0 w W0
O 0 u 30 ca

0 pa 0 V4 0 0 ,0 0 0 (
I pw v -4 r. w co a, i O0

0 E-4 0 C C: (30- 0 4 0 U
O4 9Q WB ) 1- >% wo

SO 0) w0 c E-4 4-
) CI4 604 0 4 Q
C UC4 C 0 0 C0
co C 0 u 0) S Z3 co
0) 41 M0 C0 0 R C U
() i cO O w c- Aj 0 4-

wa 3 0 41 O O4JM

Coor.o o0 w0 000
3 3 4 W 3 C 0 4 4 A CC
0 O 0 H -O 4 C-O CO )
o a 40 m N H CO c o

- 0 NWq "-4 :3 a 0w
f r- 4 a) w 4) 1 0 10 43 1

4 t4 0 c0! 0 0 0 0
44 CO Q0 C4 E41 0C
> a> 11 u 3 l 3 ,C

0 0 0 0 i 0 0 a)

4-i 0 404 C 0 0 0 4
00 o z 4) 0oor C e0
0 w 0 r00 W B4 W 0
c 14 0- C W -0 4 4
0 "0 0 4 C o a r

V-4 0.w t400M- 0$-t
0 0 0 3 E-4 0 i 0 w3

0 -,4 9M U w -r4
00 0 44 0o
0 4 0 d C a r-
*r4 C "4 O 4 0 :

14N wm W
600 0) r. 0 0 04 O
41 41 c a C C m VO
IO C4 1 WU 0" 10

w l 0 U 0 0 -1 4 0 r4 C
0 0Q c0 1 4-iH 40-1
41 "a C) r 3a
S0 4 l0 ,C 0.W 0
t o o N aH 3 a= 4 mC

00 0 C I

W4- 4J1 M -4 a) C
4u $4 d 0 $ o a ,3 c
C I-r a >i .I00 x

00 g aca 0 w M 4

0 0 1- 40 0 W0 P 0
u 0 1 0 --4 0 0

c 0 0 0 60 0t 0 (0

S) %-4 =.5 0 -4 0 0 tV 0 Co
1H 0% 0 00 4 r 0 $ 0 0c
SP 0 3 .0 C4 .0 00. *4 4 0 0



Lo M

S0 0

+ 10

I -


dSg 1OJ. 3HJ dO IN33V3d

Figure 3-5. Scatchard analysis of [3H]DEX binding to Type II
glucocorticoid receptors. Brain cytosol was prepared in HEPES buffer
containing 20 mM molybdate plus (circles) or minus (triangles) 2 mM DTT
and with (solid symbols) or without (open symbols) a 4 hr incubation at
22 C in the absence of steroid prior to incubation with [3H]DEX +/-
200-fold [1H]DEX at 0 C for 24 hr. Bound-free steroid separations were
performed on Sephadex G-25 columns.

0-2 0-4 0-6 0-8 1-0 1-2 1-4 1-6
Bsp (nM)

4 0u -1

0l ;0i

o oa

: H-r-4 i -i P Q
-4 0. >
o60 nu 0
co ( 0
1- 3tIn 4c

HI a
a)E- c "A
IH x a

0 9C c 4-
w Cd 0.= 0

C co U W

-0 4-1 4J

4-4 H- f CO W

u a
o cio a

4-M 4-1- *C

C4 0 -i 0
Wo u 4 0 0'I
imQ a
0 r-1 4 zB

41 E4-4 W .C
0 4 .) 3ta U
nO :1 0
CO ,~cO
w u' -4 10 c

"O 0 41 0
ro M w 4

0 00 r 3t

0 0 a ,0
4 CO4 0 U r.
w I4 u 4)

0 0 C CO
0) r4 o

O a) ) u-
0 (4 41 w
o oou

44 >0 0 u
H ao

io uo '
cu 1 r1 1 ca


O 0 0 0 0
0 0o (D C
1O8lNOO 8H OM3Z JO iN33y3d

differences in the rate of dissociation between receptors in the

presence or absence of DTT. Although an attempt was made to study the

effect of DTT on the rate of association of [3H]DEX to glucocorticoid

receptors, the results of this experiment were confounded by the fact

that the binding capacity of cytosol lacking DTT was undergoing a

time-dependent loss that did not occur in the presence of DTT.

A series of experiments were next performed to further characterize

the sulfhydryl up and down-regulation of glucocorticoid binding capacity

by studying the reactivity of the essential sulfhydryl groups) with a

variety of known reversible and irreversible sulfhydryl-reactive agents.

The first experiment examined the effects of a 1 hr incubation of

unlabeled cytosol with 1 mM dithionitrobenzoic acid (DTNB),

n-ethylmaleimide (NEM), iodoacetamide, cystamine, glutathione disulfide

(GSSG) and para-chloromercurisulfonate (PCMS) with or without the

subsequent addition of 10 mM DTT on glucocorticoid binding capacity as

measured by [3H]DEX binding (Figure 3-7). DTNB and cystamine led to a

virtual total loss in binding capacity which was completely reversible

upon DTT addition. GSSG resulted in only a partial loss in binding

capacity, but this loss was also completely reversible upon DTT

addition. NEM, iodoacetamide and PCMS led to varying degrees of loss

which were not completely reversible with DTT. In a follow-up

experiment, conditions identical to those described in the previous

experiment were used to investigate the effects on glucocorticoid

binding capacity of several additional sulfhydryl reactive reagents

including iodoacetic acid, sodium-m-periodate, sodium-p-periodate,

6,6'-dithionicotinic acid (DTNT), methyl methane thiosulfonate (MMTS)

and mercuric chloride (Figure 3-8). Iodoacetic acid was relatively


C 4

0 0 u 1 1

41 c a 0 0
0 -4 4 U c z d 0 4

4 C P
5. 3P0 0
13 0 > -l4 U -4W 0
1-1 cdu to () a) m

r 0 CO (A r

ca C140u 0 4 C
WX r-4 a -4 4 -4 0
) co0 0 w 0J

CU -4 0 0 41
04 ca C a u t44 0 O
0-. P4 0 P1 0
C C 0 03d 0 I C u
> oi CB 0 u U O
u-e a -01 0

0 0 1 .0 0
-0 d0 L W cw +

CU OC S -0
-4CIu09 u 0 c
. 4 04 0 0 w O
'A m a4o to S

1o0 H .0w (a 0)

m04.1O 1 00 cuo
> > UM Qo a a3

00 U L 4 d 0 .-d .
O .6 r. $40-CO4-4-"
0 C. "3 W r. ( .W
S40 0 )M 0 4

Sa9) Q)-4 u S

-r C U0 *0 M J .$
oTd w p o -r4ofl
e a mie r( 0 U f

m MM 0+0
w 0 o o a a. 0
B0oi01 a) (o 0

4O 0 Ai 0 1. w U 4U
3 4- J C0 U 5

4-4CU < U ao 'o

40 0 40 ,4 0 0-
4. 4-1 4 d 01 4- 0. M1
4-4-44 w 0 to 4 i r 0

>H wC 34 co V
O1 C OO 3 0

E4> a) 0a c- 0 4 o-

S-0 3- i C4 0-
co rc r *
4I1 0 (0 co C
0o '0 0 4- a C s
O4 CO C (0 C; Q ^

0w 0 00 0 O 0 0
0 U ,T 1 0 4 "01 C
U0 > U 4 0 44 04 c o
4 i V4 w r 3t a)
o o ao 0 "0 3 cn 0
S)0 1) I 0 0E 4
I o d-4 4 4)1 -r4 1 C
rr-4(U ai r. co "( A -4
1 C 0% 0 0 0 C1 .
cn 0 4)rr W + Q"o "0
=E = 8 ZSa ox s
U XQ 4) O r-4 0 3

00 I4 U 0 m8 w) C CI.
rM0 W> 0 00 W -W l 0 W

IO4luo j.o l0uoJad










(1 O

4-4 4- -10
00 C 0 00

[ ,Co ) 4.
0 w v a4 0 o
O c w 41 0 U w0

w acO c c
0 80 C 0 C CU 0 O0C
00 r-40 -o

'4 > t D4 I U I 1
*^ 4 rCO 4 C O
4M "4E a a)
iU z. 4S44 W-
d> o. w a r 0 (0 *.
WU B ) z -O 0 U -a 4
60 4 -4OU 0

0 I 0 -
C 0 O i CO 4E-4o "0 a o co
A o .- 0d 4)
t Po h 0 )

C $4 P. 'O o Mc

-4 C 41 ) )4
4.. 0. 0 0 0
01a CO- u XC a
0t O "0 O) 41

B 3 Cu ai # e (n
00 4-4 CU w r.
> a 0 C c 0 a0 EI
*j i- 0 o lu OM a
S0 o M 0 4 0 W- 0
-4l 9 0 WC)
r0 4J O 0 )
0 O4 0 4 0u

0 0u 0 0 a)4 44C
4 0 04 0 0 -
:U 4 d) %--- 0. C
rA r4 0 =

0 1-4 W 0 0= C
0o m W i 4-1 1 c.
4-1 = 0 -4l 1 0
O E-4 0 I I)W 0-

r-4 a ) 1-1 u 0X i-i r-4
0 0 0 0 0 0 9
wa 41 .i C A 4 44
0s 's P> W QW 0 ^-N 0 -N 4
44 O M I CU C
0 Q) 4-4 ,- CC0 4 r-4
4 0 M 0 m E-4
W i a14 E-
wu 0 '* 3 a 0 =
tu .61 41 U 41 0 4a 1 0. 0
SU H 0UO 1I "S w
M o 0 4u 0 3 03

0 I 4 .0 0 1 01 -0 0

V d u) ) 4i o "A 0 4J
0) z 4 a 0 w a co

00 -1 o 10 C3. w 4 01 t.
1H 3 3 0 :3 0 4) t44 Q)
M0 w go o 0 (0 w i l


IO4UOO 1o 4UOJead

ineffective in reducing the binding capacity, possibly due to the pH of

the system or the low concentration used, but it did produce some

irreversible losses. There was no difference between the two isomers of

sodium periodate which were intermediate in reducing binding capacity

which was not completely DTT-reversible. DTNT resulted in nearly a

total loss of binding capacity which was completely reversible while

MMTS was much less effective in reducing binding capacity. However,

total reversibility with DTT in the case of MMTS was less certain than

for DTNT. Finally, exposure to mercuric chloride resulted in a complete

loss of binding capacity which was partially reversible.

The next step was to investigate the ability of DTT-reversible

sulfhydryl reactive reagents to protect unoccupied receptors from

irreversible inactivation due to non-DTT-reversible sulfhydryl reactive

reagents. This experiment was designed to determine if unoccupied

receptors, down-regulated by endogenous sulfhydryl oxidizing factors,

were more or less protected from irreversible sulfhydryl reactive

reagents than were receptors down-regulated by the DTT-reversible

reagents used in the previous experiments. This was an effort to

possibly shed light on the nature of the down-regulated form in vivo as

well as to provide information on a possible means of "trapping" up-

and/or down-regulated forms of the receptor so that the true status of

up- and down-regulated receptors in vivo may someday be determined.

This experiment involved first, a 1 hr incubation of unlabeled cytosol

at 0 C with DTNB, cystamine, GSSG or no reagent for the DTT-reversible

down-regulation of the unoccupied receptors. This was followed by an

additional 1 hr incubation of each of the above groups with either NEM,

iodoacetamide, mercuric chloride or no reagent. This second incubation

with DTT-irreversible sulfhydryl reactive reagents was followed by

further dilution with or without DTT prior to [3H]DEX binding. Binding

data obtained for all 32 possible groups are illustrated in Figure 3-9.

DTNB and cystamine provided almost complete protection against

irreversible losses caused by either NEM or iodoacetamide. GSSG, on the

other hand, provided only a slight protection against the NEM- and

iodoacetamide-induced losses in binding capacity. Interestingly, none

of the reversible reagents provided any significant degree of protection

against the irreversible effects of mercuric chloride when compared to

the control that subsequently received mercuric chloride. These later

findings imply that either mercuric chloride is ineffective against the

endogenous down-regulated form, while being effective against the forms

down-regulated via reaction with DTNB, cystamine or GSSG, or that

mercuric chloride's effects are relatively nonspecific against both up-

and down-regulated receptors.

If one makes the assumption that mercuric chloride is acting

specifically on the sulfhydryl up-regulated form of the receptor, a

possibility implied by the results of the previous experiment, then one

would expect the incubation of cytosol containing quantitatively known

populations of both up- and down-regulated glucocorticoid receptors with

mercuric chloride, followed by addition of excess DTT (to neutralize the

reaction) at various time intervals, to result in a decreasing binding

capacity that approaches asymptotically the level of down-regulated

receptors. The results, shown in Figure 3-10, basically confirm this.

In the presence of 2 mM mercuric chloride, the binding capacity was

reduced by around 35% of maximum, a level approximating that of the

down-regulated receptors determined under similar conditions.

Figure 3-9 The ability of DTT-reversible sulfhydryl reagents to
protect unoccupied Type II glucocorticoid receptors from irreversible
inactivation due to non DTT-reversible sulfhydryl-reactive reagents.
Brain cytosol prepared with HEPES buffer plus 25 mM molybdate was
incubated at 0 C with either additional HEPES buffer (BUFF) or a 1 mM
final concentration of dithionitrobenzoic acid (DTNB), cystamine (CYS)
or glutathione dissulfide (GSSG). Cytosol samples were then incubated
at 0 C for a second hr with either additional HEPES buffer (BUFF) or a 2
mM final concentration of n-ethyl maleimide (NEM), iodoacetamide (IAM)
or mercuric chloride (HgC12). Each subset of cytosol was then incubated
at 0 C for an additional 1 hr prior to the addition of either HEPES
buffer (open bars) or HEPES buffer containing excess DTT (20 mM final
concentration, solid bars) followed by incubation at 0 C for 40 hr with
20 nM [3H]DEX +/- 4 um [1H]DEX. Bound-free steroid separations were
performed on Sephadex G-25 columns. Specific binding is expressed as
percent of the BUFF-BUFF plus DTT control group (302 fmol/mg protein)
and represents the mean +/- S.E.M. of 3 independent replications.


- 75


o 50


0- 25


- 75

3 50


0- 25



0 C

0 caE 0 )
a. -o 00
0 00 .C:

4 -u 4. 4a
0 0 0o 0

%a 0

@ .4-1 -14 0 -0
o CU r 0 l U 4

00 0"4 c
0000 00r

-c0o a004J

u 04u I I4O
0 0 4 II
o E- 0 -4 a Wa
O E-i e C) c a) V1
1O U0 0 3 0)+
= -H %-1 0 00
bo 4-i EW 0c
- 0 a
-40 4J 4-4 0.0
C C 0'4-4X
0 0 0 E 0
0.-4 U 0 q 0

0 10 r 0 M CEC

04-10 = s. 40
4) 43 C 440 C W

0 co 0) 0 I Q)
c 0 0 41 +

S3 0 0 X U

0u 0 0C 0n
wl 5. W 4 0.
O C6 0o0 c 0
w 0 0 ".4
C 08d00.

01 01 m 04 c0
co w ) 4) ,

00W .0 04-I
0 0 0w 0.0 14 0
01 4- 1 00
UO 4) -H 0 O

0 0 Ci- 0 4. C,
4 0 0 T CO
) 4-4 ) 0)

0013 0-0
w o r-q 4 4

4 001 0 0
0 C 4J 0 r-

31 U1 o e o i-i
0) $,4 4
0 4A 03 u 0 L 0
U r-4 e-% -4 0
Sa n o 4a :
a> 44 cc coC: a< 0-

o .0- c a4o

00. 41Jcc on
t- C0 4 ..0.0 0r-4 0
I w 0j 0 0 "-
0 p u u 31 00 0
$4 -H 5.4 0 $40 -4
0 '0 r0 w -4 0 0.

S.0 l U 4J144 0M =-W

104JU00 o0 1uaojad

The next experiment was designed to determine if DTT-reversible

sulfhydryl reactive reagents were capable of interacting with prelabeled

glucocorticoid receptors and affecting the rate of steroid dissociation,

inactivation, etc. for the steroid receptor complex. Cytosol was

prepared in the absence of any sulfhydryl reactive or reducing reagents

and prelabeled with 20 nM [3H]DEX for 40 hr at 0 C. After a bound/free

separation on Sephadex G-25 columns, cytosol was incubated with 1 mM

concentration of DTNB, cystamine, GSSG or an appropriate dilution of

HEPES buffer devoid of any sulfhydryl reagents for 4 hr at 0 C in the

presence or absence of 4 uM [1H]DEX. The results from this experiment

(data not shown) confirmed expectations that these DTT-reversible

reagents could only affect the steroid binding properties of unoccupied

glucocorticoid receptors since no enhancement of the rate of

dissociation of [3H]DEX was evident. It was evident, however, that

interactions between these reagents and receptors unoccupied by virtue

of normal steroid dissociation did reversibly prevent steroid rebinding.

When the previous experiment was replicated using mercuric

chloride, instead of known DTT-reversible reagents, very different

results were obtained. When cytosol containing both [3H]DEX-labeled

(up-regulated) receptors and down-regulated unoccupied receptors was

incubated in the presence of 1 mM mercuric chloride for 4 hr, there was

a near total loss of receptor binding. Upon addition of DTT, and after

subsequent reincubation with [3H]DEX, binding equivalent to the

population of down-regulated unoccupied receptors reappeared (data not

shown), indicating the possibility that bound steroid was not protective

of the inactivating effects of the mercury, but that the down-regulated

form was somehow protected from such inactivation.