Reprinted from Nature, Vol. 284, No. 5751, pp. 17-22, March 6 1980
© Macmillan Journals Ltd., 1980

The role of hormone receptors
and GTP-regulatory
proteins in membrane transduction

Martin Rodbell

Laboratory of Nutrition and Endocrinology, National Institute of Arthritis, Metabolic and Digestive Diseases,
National Institutes of Health, Bethesda, Maryland 20205

 

Cell membrane receptors for hormones and neurotransmitters form oligomeric complexes with GTP-
regulatory proteins and inhibit the latter from reacting with GTP. Hormones and neurotransmitters act
by releasing the inhibitory constraints imposed by the receptors, thus allowing the GTP-regulatory
proteins to interact with and control the activity of enzymes such as adenylate cyclase. This theory may
apply generally to membrane signal transduction involving surface receptors.

 

ADENYLATE CYCLASE, the enzyme that produces cyclic
AMP, is part of a complex regulatory system that mediates the
actions of hormones and neurotransmitters on their target cells.
Structured within the lipid framework of the cell membrane, the
enzyme system is composed of at least three classes of
components (Fig. 1). Located at the outer membrane surface is
the receptor (R) component containing a specific site for binding
of hormones and neurotransmitters. At the inner face of the
membrane are the catalytic unit (C) and the nucleotide regula-
tory component (N). The latter contains site(s) for binding GTP
and is responsible for mediating the effects of GTP and the
various hormones on the activity of C'. Two types of N units
have been distinguished functionally. One mediates stimulation
(termed N,), the other inhibition (Nj) of the adenylate cyclase
activity by GTP. As discussed below, each type seems to be
linked to separate classes of receptors for hormones and neuro-
transmitters.

Here I present a theoretical framework for the role of
hormone receptors and N units in regulating adenylate cyclase
activity. In essence, the theory suggests that the N and R units
normally exist separately from C as aggregates or oligomers of
an RN complex. In this complex, R inhibits interaction of N with
GTP. Hormone binding to R triggers release of the inhibitory
constraints imposed on N with resultant enhanced reaction with
GTP, followed by breakdown of the oligomers to a monomeric
RN complex. The latter reacts with C to form the holoenzyme
structure depicted in Fig. 1. Depending on the type of R and N
unit attached to C, the holoenzyme exhibits either increased or
decreased production of cyclic AMP. In developing this theory, I
review evidence for the existence of N, and N;, and their
complexes with R and C, describe the properties of the various
components, give recent evidence that RN oligomers exist and
discuss the possibility that the theory may be generalised to
include R and N units that regulate membrane transduction
processes other than adenylate cyclase.

The role of N, in activation of adenylate
cyclase

Before the discovery that GTP is the essential activator of
adenylate cyclase” and that hormones enhance the nucleotide’s
action®, it was thought that hormone receptors interacted
directly with the catalytic unit and that fluoride ion, a non-
physiological activator, affected the catalytic unit directly’. It is
now clear that the actions of both hormones (through their

specific receptors) and fluoride ion are mediated through pro-
teins that bind GTP. First identified* with GTP-photoaffinity
analogues as a heat-stable 42,000 molecular weight protein in
detergent extracts of avian erythrocytes, the protein (N,) has
been shown to exist in a variety of cell membranes. Cholera
toxin has been particularly valuable for identifying N,. Long
known to stimulate the production of cyclic AMP in animal
cells’°, the toxin potentiates the activating effects of GTP even
in the absence of hormones and affects other characteristics of
the enzyme that suggest that N, is its primary site of action'®”’*,
This was firmly established when it was found that the toxin,
which contains in its A; subunit an ADP-ribosylating activity,
preferentially labels the 42,000-MW protein in the presence of
>?P_NAD (refs 15-17). Cells deficient in N, by functional criteria
also lack the toxin-labelled protein; addition of detergent
extracts of membranes containing N, reconstitutes the ability of
hormones, guanine nucleotides, fluoride ion and cholera toxin to
stimulate cyclase activity in membranes from cells genetically
deficient in N, but otherwise containing hormone receptors and
adenylate cyclase'*'®. Thus, judged from several standpoints,
N, is an essential component in the activation of adenylate
cyclase. Although its structure remains unknown, N, must have
highly conserved recognition sites that allow it to activate C
derived from a variety of cells and to ‘couple’ with the several
types of receptors that mediate the stimulatory actions of
hormones.

Table 1 lists a few of the properties of N, and C when separate
and combined (combinations with R are discussed below). C
uses MnATP preferentially as substrate*’?', Only when
combined with N, does C use the natural substrate, MgATP, as
effectively as MnATP". Association of N, with C is reversible
and is driven by the binding of guanine nucleotides to N, (refs 6,
22).

Properties and role of RN complexes

In classical theories of hormone action, the binding event leads
the receptor to adopt a structure that is favourable for action.
The relationship between hormone binding (Kp) and action
(Kaa) on adenylate cyclase systems is complicated by the fact
that two ligands, hormone and GTP, are required for action’.
This complexity is exemplified by studies of the glucagon-
sensitive cyclase system in liver membranes***>. Direct binding
studies with labelled glucagon revealed that GTP at concen-
trations required for activation of adenylate cyclase in the
presence of hormone, converted 90% of the receptors to a state
with a higher Kp than K,..; the remaining 10% displayed both
the kinetic and thermodynamic properties commensurate with
hormonal activation of the enzyme. A similar distribution of
glucagon receptor states is seen in intact hepatocytes” which
presumably contain sufficient GTP to interact with N, at the
internal face of the cell membrane. A plausible explanation for
the effect of GTP on K,., and Kp is the following set of reactions
(modified from ref. 27)
H+R-NH-R-N* (1)
H-R:N*+GTP2H-R-N-‘GTP2H+R-N-GTP (2)
H:R-N-GTP+C2H-R-N-C
GTP
(active holoenzyme) (3)
in which hormone binding and action are initiated at step (1)
with resultant formation of an ‘activated’ state of N(N*). Reac-
tion of the latter with GTP (step 2) leads to a complex
(H:R:N:GTP) which preferentially couples with C to form the
activated holoenzyme. The negative heterotropic effects of GTP
(decreased hormone binding) derive from a lower-affinity form
of RN when occupied by GTP and not complexed with C
(step 2). Thus, association of the macromolecular components
(R,N, C) enhances the binding affinity of the small ligands
(hormone and GTP). In this ‘uncoupled’ equilibrium model, the
final concentration of the activated holoenzyme is a function of
the relative concentrations of both the macromolecular
components and the small ligands. In the overall equilibrium, all
RN complexes have equal potential to form complexes with C
but the amount of RNC formed is limited by the concentration
of C. In the case of the liver system, the latter may be 10% of R
and N.

Evidence from 8-adrenergic receptor systems supports the
above reaction scheme. For example, agonists but not
antagonists promote the negative heterotropic effects of GTP on
catecholamine binding** >’. This is consistent with an ordered
reaction in which hormones promote interaction of N in the RN
complex with GTP. Evidence that the negative heterotropic
effects derive from the RN complex stems from findings that
cells lacking C but containing N, and R display the negative
heterotropic effects of GTP on hormone binding’’. Further-
more, RN complexes have been isolated following detergent
extraction of membranes and separation from C (or NC); such
complexes show the effects of GTP on hormone binding seen
with intact membranes****. Membranes from cells genetically
deficient in N, fail to show not only responses to hormones and
guanine nucleotides, but also heterotropic effects of GTP on

 

Table 1 Properties of adenylate cyclase components

 

Selected
Components Properties refs
Cc Preferential reaction with MnATP as 20, 21
substrate
NC MgATP or MnATP as substrate when 16, 23
activated by Gpp(NH)p*, cholera toxin
and NAD in presence of GTP, and by
fluoride ion
R Low affinity for hormone agonists; no 19
heterotropic effects of GTP
RN or (RN)" GTP reduces affinity of R for hormone 24, 31, 35
agonists. Hormones form tight-binding
complex
RNC Same as NC but responds to hormones 38

and Gpp(NH)p; binds GDP tightly in
absence of hormones. Hormones
stimulate exchange of GDP and GTP at
N site

 

*Gpp(NH)p is a GTP analogue that is not hydrolysed to GDP by
GTPases in membranes.

Hormone

   

lasma
“@ membrane

        

Cyclic AMP

Fig. 1 Schematic representation of the components and organis-
ation of the adenylate cyclase holoenzyme responsible for regula-
tion by hormones and GTP. The receptor (R) is visualised as
spanning the plasma membrane and having different segments
(indicated by the shaded and cross-hatched areas) which have
functions for binding of hormone, attachment to membrane and
linkage with the nucleotide regulatory unit (N) that binds GTP. The
N unit forms a bridge between R and the catalytic component (C) at
the internal face of the membrane.

hormone binding to receptors; addition of extracts containing N,
to such membranes restores the ability of GTP to affect hormone
binding*’. Further evidence that the N, unit responsible for
activation of cyclase by guanine nucleotides, fluoride and
cholera toxin is the same as that linked to R can be deduced from
findings that cholera toxin affects the actions of GTP on both
hormone binding and cyclase activity'*. Accordingly, there is no
need to invoke different regulatory N units linked to R and C, as
has been suggested to explain differing properties of guanine
nucleotide effects on hormone binding and adenylate cyclase
activity’”**"*’. The differences can be explained by hetero-
geneous forms of N associated with the other components
(RN, NC, RNC) having different properties with respect. to
affinities and actions of guanine nucleotides (Table 1).

The evidence cited above for the existence of RN complexes is
based on the effects of GTP on hormone binding. On theoretical
grounds, reciprocal effects of hormones on guanine nucleotide
binding should also be observed. Recent findings*®*° that
hormones promote the exchange of bound and free guanine
nucleotides at the N, unit provide further evidence that the R
and N units are structurally linked. The functional consequences
of this linkage on adenylate cyclase regulation are further
discussed below.

It is evident that R units not associated with N units are
nonfunctional with respect to adenylate cyclase activation by
hormones; free R units in membranes have a lower affinity for
agonists than do RN units (when unreacted with GTP)”. Per-
haps, free R units are formed from dissociation of RN units and
are en route to endocytotic removal*®*'.

Table 2 lists receptors reported to show negative heterotropic
effects of GTP on hormone binding and which are presumed,
therefore, to represent RN complexes. Note that they fall into
three categories: those involved in stimulation of adenylate
cyclase (RN,), those known to mediate inhibition of the enzyme
(RN;), and those which either do not interact with adenylate
cyclase or whose function is unknown (RN,).

Role of RN; in the inhibition of adenylate
cyclase by neurotransmitter and GTP

In contrast to its adenylate cyclase-activating role, GTP can
inhibit some adenylate cyclase systems. First observed’? and
characterised**“*° in rat adipocyte membranes, recent
studies*’*? indicate that this process differs from the
components that mediate stimulation by several lipolytic
hormones in the same membranes. For example, adenosine
promotes inhibition of adipocyte adenylate cyclase by GTP
through processes which differ from the GTP-stimulatory
process in their differential susceptibility to effects of sulphydryl
agents, proteases, divalent cations, sodium ions and mercurials.
The close functional linkage between the adenosine receptor
(R-site*°) and N; in the adipocyte suggests that there is an RN;
complex for adenosine in fat cells. It appears that RN; and RN,
complexes in the fat cell interact with a common C unit*’.
Opposing regulation by independent types of RN complexes
(Fig. 2) may be widespread, particularly in cells that have
adenylate cyclase systems governed by the endocrine and
neuroendocrine systems. Receptors for such neurotransmitters
as opiates, dopamine, catecholamines (a-adrenergic) and
cholinergic agents (muscarinic) also mediate inhibition of
adenylate cyclase by a GTP-dependent process*!~**. In common
with the adenosine receptor linked to Nj in adipocytes, these
receptors are sensitive to sodium ions. Interestingly, in the few
cases examined, negative heterotropic effects of GTP on agonist
binding are observed with the same membranes showing the
GTP dependency of agonist inhibition of adenylate cyclase;
sodium effects on agonist binding are also observed (Table 2).
These findings suggest, by analogy with the role of RN,
complexes in the stimulation of adenylate cyclase, that RN;
complexes are involved in the inhibitory process.

The structure of RN

The technique of target size analysis has recently been used to
examine the size of adenylate cyclase systems in various condi-
tions in their membrane-bound form (for application and theory
see ref. 55). Results obtained from liver*®, adipocytes and turkey
erythrocyte membranes are shown in Table 3, The minimal size
unit expressing activity is that obtained with MnATP as sub-
strate and no activating ligands; it is presumed to represent C.
The size increases significantly as N.C is produced with pre-
activation by guanine nucleotides or fluoride ion. Activation by
hormones and GTP which should produce the holoenzyme
(RNC), is accompanied by a further increment in size com-
mensurate with this premise. In the case of the liver system, the
sizes of the regulatory complexes (RN,) were estimated from
studies of the binding of labelled glucagon in the presence of
GTP and from the ‘ground-state’ cyclase activity, that is the size
of components present before coupling. Both studies indicate
that the RN, structure in the liver has a MW in the range
6-13 x 10°, which is three to six times the estimated size of the
RN unit associated with the holoenzyme. The conclusion from
these findings is that RN, exists as oligomers when not linked to
the C unit. Note in Table 3 that the unit (presumably RN;j)
responsible for inhibition of adipocyte adenylate cyclase by
adenosine and GTP is significantly larger than the oligomeric
RN, unit that mediates the stimulatory effects of hormones and
GTP on the same enzyme system. Although target analysis

 

Stimulating agents Inhibiting agents

a-Adrenergic

B-Adrenergic Opi

piates .
Glucagon, ACTH, etc. Cholinergic (muscarinic) Sodium 4
Adenosine \ / Adenosine regulate

a (MG,

 

 

 

Fig. 2. Schematic representation of dual regulation of adenylate
cyclase systems by stimulatory and inhibitory hormones and
neurotransmitters. Depicted in the model are two classes of recep-
tor (R) one (R,) mediating hormone effects through stimulatory
nucleotide regulatory units (N,) and another (R;) mediating inhibi-
tory effects through linkage with an N; unit that binds GTP and
inhibits adenylate cyclase activity. The R\N; units require Na” for
the effects of the various inhibitory agents on adenylate cyclase (C)
activity.

cannot give the structure and composition of the target, it can be
inferred from the size differences that the stimulatory and
inhibitory processes reflect different structures that are
multiples of the RN units comprising the holoenzyme.

Although the sizes of the adenylate cyclase system in turkey
erythrocytes display increments with increasing regulatory
complexity, this system does not show the oligomeric structures
of RN,; the functional size remains identical before and after
activation with catecholamines and guanine nucleotides.
Possibly, R, N and C are already assembled in the holoenzyme
structure due to prior activation of the enzyme system during
isolation of the membranes. Pre-assembly may explain some
notable differences in kinetic behaviour of the turkey and liver
adenylate cyclase systems (discussed below). A pre-assembled
unit is also consistent with the report*’ of a linear relationship
(rather than the usual hyperbolic relationship) between
hormonal activation and receptor occupation in the turkey
system.

A model for hormone and GTP action on
adenylate cyclase

The existence of oligomeric complexes of RN provides a struc-
tural basis for the uncoupled equilibrium reactions described

above. For illustrative purposes, the model depicted in Fig. 3
shows a tetrameric structure of RN units existing in two

 

Table 2 Hormone receptors regulated by GTP (the RN complex)

 

Type of Selected
Receptor type Source N unit* Comments refs
Glucagon Rat liver N, GTP = GDP; Gpp(NH)p and Gpp(CH)sp less potent 24
Catecholamines Several cell types N, Divalent cations promote binding of agonists; only agonist 28-32
(B-receptors) binding affected
Prostaglandin E Thyroid, frog erythrocyte Ng Binding promoted by Ca** ions 98,99
Dopamine Corpus striatum N,,N, GTP=GDP>Gpp(NH)p; agonist specific 100, 101
Muscarinic Canine and rat myocardium N; Methacholine inhibits GTP effects on 8-receptor; Na~ 102, 103
probably affects binding
Neuroblastoma x glioma cells N; Mg?” enhances agonist binding 105
Catecholamines Brain N, GTP = Gpp(NH)p > GDP; Na” affects agonist binding 106, 107
(a@-receptors) Rat liver N, Agonist specific 83
Angiotensin Adrenal cortex N, Agonist binding affected by Na” in same manner as by GTP 84
Opiates Brain N, Na* affects binding of agonists and antagonists. Two 52, 104,
distinct opiate receptors 108

 

In all cases, addition of guanine nucleotides decreases binding of hormone or neurotransmitter to specific receptors in isolated membrane

preparations.

*N, is N unit linked to stimulation of adenylate cyclase; N;, affects inhibition of the enzyme; N, isan N unit that is either not related to cyclase activity
or has an undetermined relationship to a specific signal-processing system in the cell membrane.
configurations: an unoccupied structure (A) which favours
binding of hormones but not GTP, and a hormone-induced or
stabilised structure (B) which favours reaction of GTP with the
N component (equivalent to N* in the previous reaction
scheme), Reaction with GTP results in dispersion of the
oligomer to monomers (at this stage the negative heterotropic
effects of GTP occur) which uniquely react with C to form the
holoenzyme that converts MgATP to cyclic AMP. In broad
terms, this ‘disaggregation-coupling’ model for hormone and
GTP action can be likened to the manner by which cyclic AMP
controls through its receptor the activity of protein kinase; the
latter involves a change in the association of dimeric regulatory
and catalytic subunits’*. In the case of hormone receptors
associated with adenylate cyclase systems, their interaction with
N units in the oligomeric state constrains the ability of N to react
with GTP, thus preventing the formation of the ‘active’ form of
the regulatory N unit (N, or N;). It follows that cyclase systems
devoid of receptors linked to N should display high reactivity
with GTP. This has been reported” recently with a strain of
HeLa cells deficient in 6-adrenergic receptors; when the same
cells become enriched with receptors, catecholamines are
required to restore the level of activity seen with GTP alone in
receptor-deficient cells.

The disaggregation-coupling model has several other
important features that serve to distinguish it from previous
theories of hormone action®**?. First, assuming that the oli-
gomers of RN are unreactive with C, it provides a means for
functional compartmentalisation of RN units from C and
explains the observed dependency of the reaction on both
hormone and GTP. Second, an oligomeric structure of RN
allows for homotropic subunit interactions such that minimal
occupation by hormones may cause near maximal production of
the functional activating ‘signal’ (the RN-GTP monomer) in the
presence of saturating concentrations of GTP; marked activa-
tion of adenylate cyclase with minimal occupation of receptors
has been observed with several cyclase systems?**”***. Thus,
although full occupation of receptor subunits may drive
complete coupling of RN with C, coupling may occur even when
these units are not occupied with hormone. Third, because the
oligomeric RN structure is inert with respect to activation of
adenylate cyclase, aggregation of uncoupled regulatory units is a
means of ‘turning off’ the reaction cycle. How ‘uncoupling-
aggregation’ is accomplished and what factors control this part
of the cycle remain unknown. Possible candidates are cytosolic
factors reported to affect adenylate cyclase activity". Addi-
tionally, factors intrinsic to or associated with the plasma
membrane may be essential in the aggregation-disaggregation
cycle; these factors could include cytoskeletal elements” which
may stabilise the aggregated units in the membrane; interes-
tingly, agents that disrupt cytoskeletal structures have been
reported’'”? to enhance cellular production of cyclic AMP in
response to hormones.

Emphasis has been placed recently on regulation of hydrolysis
of GTP by a GTPase putatively associated with the N, unit !*7*”*

Uncoupling-aggregation
‘turn-off cycle

Disaggregation-coupling
“turn-on” cycle

 

GTP

 

: MgATP
Cyclic AMP Key:®)

Fig. 3. A model for coupling of the receptor-nucleotide regula-

tory units (RN) to the catalytic unit (C) of adenylate cyclase and the

role of hormone and GTP in this process. See text for detailed
description.

 

Table 3 Functional sizes of adenylate cyclase components

 

Components (MW x 10°)

Enzyme source RN, Cc N,C RN,C_ RN;
Rat liver: R = glucagon 6-13* 1.5 2.3 3.5 _—

Fat cell (rat)

R =catecholamine, ACTH 13 + 2.3 + >13
R; = adenosine
Turkey erythrocyte ND 09 1.8 2.5 _

R= catecholamine

 

Sizes were determined by target size analysis using high energy
irradiation. See ref. 56 for methods for determining size and rationale
for assigning components to size. Data for fat cell and turkey erythrocyte
data are unpublished. ND, Not detected.

* Estimated both from GTP-sensitive glucagon binding and adenylate
cyclase activity of ‘ground-state’ enzyme.

+ Not determined.

and the role of this regulation in turning-off the GTP-activation
process. In turkey erythrocyte membranes, catecholamines and
cholera toxin affect GTPase activity and GDP is a potent
inhibitor that binds tightly to N, (refs 38, 39). The differences in
the kinetic characteristics of the liver cyclase system in its
response to GTP and Gpp(NH)p are consistent with a GTPase
being involved in its dynamic characteristics’*. However, GDP
has been shown to stimulate liver cyclase activity in the presence
of glucagon'’**”*, stimulatory effects of GDP have been
reported also in the presence of ACTH on the adipocyte
adenylate cyclase system’’. In view of these findings it is not
clear that a GTPase turn-off mechanism”®”° (which necessarily
requires that GDP binds tightly and acts only as inhibitor) can
adequately explain the dynamic properties of all adenylate
cyclase systems. Perhaps the difference in the properties of the
turkey, liver and adipocyte systems is related to the lack of.an
oligomeric structure of RN in isolated turkey erythrocyte
membranes.

The topographical relationship between the RN oligomers
and C in the plasma membrane and the role of membrane lipids
in this relationship remain unknown. Studies with phos-
pholipases have shown that hormone action on adenylate
cyclase remains intact even after 85% of the membrane phos-
pholipids have been digested; substantial loss in action occurs
only when a fraction of the remaining phospholipids is
hydrolysed®°*?, These findings could mean that RN oligomers
and C are in close proximity and possibly in selective domains of
interacting phospholipids; the bulk of the phospholipids do not
seem to be involved in assembly of the holoenzyme.

Generalisations and problems

If the notion is accepted that the effects of GTP on hormone
binding to receptors are due to disaggregation of oligomers of
receptors linked to N units, then the theory cited above may
apply generally to all hormone-regulated systems that illustrate
these effects, even those N units (N,) not linked to adenylate
cyclase. The latter include such GTP-affected receptors as
angiotensin receptors in the adrenal medulla® and a-adrenergic
receptors in liver®* (see Table 2). Recent findings®* that insulin
activates a cyclic AMP-independent protein kinase in
sarcolemma membranes by a process regulated by GTP is an
interesting example of a potential N,-mediated process
unrelated to adenylate cyclase.

The possibility that different N units mediate the actions of
hormones raises interesting new questions. Can the properties
of the same receptor be modified by interaction with different
types of N units? What determines the interactions of receptors
with a particular type of N unit? The pertinence of these
questions is exemplified by reports that catecholamines, acting
through a-adrenergic receptors (by pharmacological criteria),
both stimulate®* and inhibit*’** cyclic AMP production, and
induce effects unrelated to adenylate cyclase activity®®. Other
hormones exerting multiple effects on membrane processes
48-50 §3,87-89 90-92
’ ’

include adenosine dopamine , Opiates vaso-
pressin?*** and serotonin’. Pharmacological studies (speci-
ficity, potency, antagonist or hormone binding to receptors)
alone may not identify the type of receptor mediating these
processes. Heterotropic effects of GTP on agonist binding
indicate linkage of receptors to an N unit but do not identify the
type of N unit. I have noted in Table 2 that certain RN
complexes are affected by sodium ions, others by divalent
cations, and that guanine nucleotides have different potencies
on agonist binding. Perhaps such differences can be used to
classify the N unit associated with the receptor. Clearly, what is
necessary in future research in this area is to develop assay
methods that identify unequivocally each type of N unit. Testing
of biological effects is not necessarily the means of determining
the type of RN unit. A bizarre example of problems encountered
is the action of cholecystokinin on pancreatic acinar cells°®. In
the cell the hormone stimulates zymogen secretion, calcium
release and cyclic GMP production, but not the production of
cyclic AMP; after breakage of the cell, the hormone stimulates
adenylate cyclase in a GTP-dependent fashion.

Although other explanations are possible, these findings raise
the possibility that ‘redistribution’ of receptors and N units
occurs on cell breakage owing to the breakdown of stationary
domains that normally segregate these units in the cell. ‘Lateral
domain redistribution’ might be a physiologically regulated
process due either to changes in the structural relationship of the
cytoskeleton to the ordered domains or tq propagated dis-
turbances in the membrane structure. A possible example of the

. Rodbell, M. in Molecular Biology and Pharmacology of Cyclic Nucleotides (eds Foleo, G. &
Paoletti, R.} 1-12 (Elsevier, Amsterdam, 1978).
. Rodbell, M., Birnbaumer, L., Pohl, S. L. & Krans, H. M. J. J. biol. Chem. 246, 1877-1882
(1971).

. Rodbell, M. et al. Adv. Cyclic Nucleotide Res. 5, 3-29 (1975).

. Perkins, J. P. Adv. Cyclic Nucleotide Res. 3, 1~64 (1973).

. Pfeuffer, T. J. biol. Chem. 252, 7224-7234 (1977).

. Pfeuffer, T. FEBS Lett. 101, 85-89 (1979).

. Finkelstein, R. A. Crit. Rev. Microbiol. 2, 553-623 (1973).

. Gill, D. M. Adv. Cyclic Nucleotide Res. 8, 85-118 (1977).

- Johnson, G. L. & Bourne, H. R. Biochem. biophys. Res. Commun. 78, 792-798 (1977).

. Fischer, J. & Sharp, G. W. G. Biochem. J. 176, 505-510 (1978).

11. Cassel, D. & Selinger, Z. Proc. nain. Acad. Sci. U.S.A, 74, 3307-3311 (1977).

12. Levinson, S. L. & Blume, A. J. J. biol. Chem. 252, 3766-3774 (1977).

13. Lin, M. C., Welton, A. F. & Berman, M. F. J. Cyclic Nucleotide Res. 4, 159-168 (1978).

14. Cassel, D. & Selinger, Z. Biochim. biophys. Acta 452, 538-551 (1978).

1S. Gill, D. M. & Meren, R. Proc. natn. Acad. Sci. U.S.A. 75, 3050-3054 (1978).

16. Cassel, D. & Pfeuffer, T. Proc. natn. Acad. Sci. U.S.A. 75, 2669-2673 (1978).

17. Johnson, G. L., Kaslow, H. R. & Bourne, H. R. Proc. nam. Acad. Sci. U.S.A. 75,
3113-3117 (1978).

18. Johnson, G. L., Kaslow, H. R. & Bourne, H. R. J. biol. Chem. 253, 7120-7123 (1978).

19. Ross, E. M., Howlett, A. C., Ferguson, K. M. & Gilman, A. G. J. biol. Chem. 253,
6401-6412 (1978).

20. Howlett, A. C., Sternweiss, P.C., Macik, B. A., Van Arsdale, P.M. & Gilman, A. G. J. biol.
Chem. 254, 2287-2295 (1979).

21. Londos, C., Lad, P. M., Nielsen, T. B. & Rodbell, M. J. supramolec. Struct. 10, 1031-1037
(1979).

22. Nielsen, T., Lad, P. M., Preston, M. S. & Rodbell, M. Biochim. biophys. Acta (in the press).

23. Rodbell, M., Krans, H. M. J., Pohl, S. L. & Birnbaumer, L. J. biol. Chem. 246, 1872-1876
(1971).

24, Rodbell, M., Lin, M. C. & Salomon, Y. J. biol. Chem. 249, 59-65 (1974).

25. Lin, M. C., Nicosia, S., Lad, P. M. & Rodbell, M. J. biol. Chem. 252, 2790-2792 (1977).

26. Sonne, O., Berg, T. & Christoffersen, T. J. biol. Chem. 253, 3203-3210 (1978).

27. Lad, P. M., Welton, A. F. & Rodbell, M. J. biel. Chem. 252, 5942-5946 (1977).

28. Maguire, M. E., Van Arsdale, P. M. & Gilman, A. G. Molec. Pharmac. 12, 335-339
(1976).

29. Ross, E. M., Maguire, M. E., Sturgill, T. W., Biltonen, R. L. & Gilman, A. G. J. biol. Chem.
252, 5761-5775 (1977).

30, Williams, L. T. & Lefkowitz, R. J. J. biol. Chem. 252, 7207-7213 (1977).

31. Lefkowitz, R. J. & Williams, L. T. Ade. Cyclic nucleotide Res. 9, 1-17 (1978).

32. Sternweiss, P. C. & Gilman, A. G. J. biol. Chem. 254, 3333-3340 (1979).

33. Welton, A. F. etal. J. biol. Chem. 282, 5947-5950 (1977).

34. Limbird, L. E. & Lefkowitz, R. J. Proc. natn. Acad. Sci. U.S.A. 75, 228-232 (1978).

35. Maguire, M. E., Ross, E. M. & Gilman, A. G. Adv. Cyclic Nucleotide Res. 8, 1-83 (1977).

36. Iyengar, R. & Birnbaumer, L. Proc. natn. Acad. Sci. U.S.A. 76, 3189-3193 (1979).

37. Iyengar, R., Swartz, T. L. & Birnbaumer, L. J. biol. Chem. 254, 1119-1123 (1979).

38. Cassel, D. & Selinger, Z. Proc. natn. Acad. Sci. U.S.A. 75, 4155-4159 (1978),

39. Lad, P. M., Nielsen, T., Preston, M.S. & Rodbell, M. J. biol. Chem. (in the press).

40. Catt, K. J., Harwood, J. P., Aquilera, G. & Dufau, M. L. Nature 280, 109-116 (1979).

41. Goldstein, J. L., Anderson, R. G. W. & Brown, M. S. Nature 279, 679-685 (1979).

42. Cryer, P. E., Jarett, L. & Kipnis, D. Biochim. biophys. Acta 177, 586-590 (1969).

43. Harwood, J. P., Low, H. & Rodbell, M. J. biol. Chem. 248, 6239-6245 (1973).

44, Rodbell, M. J. biol. Chem. 250, 5826-5834 (1975).

45. Pairault, J. Eur. . Biochem. 62, 323-334 (1976).

46, Ebert, R. & Schwabe, U. Naunyn-Schmiedebergs Archs Pharmak. 286, 297-313 (1974).

47. Yamamura, H., Lad, P. M. & Rodbell, M. J. biol. Chem. 252, 7964-7966 (1977).

48. Londos, C., Cooper, D. M. F., Schlegel, W. & Rodbell, M. Prec. nam. Acad. Sci. U.S.A. 75,

$362-5366 (1978).

N

Deen danaw

latter is the report’’ that catecholamines, acting through a
8-adrenergic receptor, stimulate a phospholipid methylating
enzyme by a process which is affected by GTP but which does
not involve the production of cyclic AMP; associated with
methylation is an apparent ‘flip’ of the internally methylated
lipids to the outer face of the membrane and a change in lipid
microviscosity. Pleiotropic effects of the hormone could thus be
generated by localised changes in lipid structure being pro-
pagated laterally and modifying the postulated domain struc-
tures of the membrane. In any event, these findings question the
commonly held view that B-adrenergic receptors are singularly
linked to stimulation of adenylate cyclase. Moreover, the
dependency of the hormone effect on GTP suggests again the
versatility of the N units in mediating the actions of hormones.

In conclusion, the classical notion of receptor alone control-
ling the events related to hormone and neurotransmitter action
on adenylate cyclase is no longer valid. The constraining role of
hormone receptors postulated here differs from the role of the
receptor postulated in other theories of hormone action. More
importantly, the theory places in perspective the role of another
set of regulatory proteins (the N units) hitherto given relatively
scant attention, particularly with regard to their apparent mul-
tiple and fundamental roles in the regulation of membrane-
associated processes. I hope this rather brief article will stimu-
late investigation of the new problems to be faced in ascertaining
both the structures of the GTP-regulatory proteins and how they
function in the transduction of hormone binding at cell
membrane receptors into physiological action.

49. Cooper, D. M. F., Schlegel, W., Lin. M. C. & Rodbell, M. J. biol. Chem. 254, 8927-8931
(1979).

50. Londos, C. & Wolff, J. Proc. nam. Acad. Sci. U.S.A. 74, 5482-5486 (1977).

$1. Sabol, S. L. & Nirenberg, M. J. biol. Chem. 254, 1913-1920 (1979).

52. Blume, A. J., Lichtshtein, D. & Boone, G. Proc. nain. Acad. Sci. U.S.A. 16, 5626-5630
(1979).

53. DeCamilli, P., Macconi, D. & Spada, A. Nature 278, 252-254 (1979).

54. Jacobs, K. H. in Molecular Biology and Pharmacology of Cyclic Nucleotides (eds Folco, G.
& Paoletti, R.) 265-278 (Elsevier, Amsterdam, 1978).

55. Kempner, E. S. & Schlegel, W. Analyt. Biochem. 92, 2-10 (1979).

56. Schlegel, W., Kempner, E. S. & Rodbell, M. J. biol. Chem. 254, 5168-5176 (1979),

57. Brown, E, M., Fedak, 8. A., Woodward, C. J., Aurbach, G. D. & Rodbard, D. J. diol.
Chem. 251, 1239-1246 (1976).

58. Nimmo, H. G. & Cohen, P. Adv. Cyclic Nucleotide Res. 8, 145-266 (1977).

59, Lin, M. C., Lin, C-S & Whitlock, J. P. Jr J. biol. Chem. 254, 4684-4688 (1979).

60. Bergman, R. N. & Hechter, O. J. biol. Chem. 283, 3238-3250 (1978).

61. Tolkovsky, A. M. & Levitzki, A. Biochemistry 17, 3795-3810 (1978).

62. Jacobs, S. & Cuatrecasas, P. Biochim. biophys. Acta 433, 482-495 (1976).

63. Hechter, O., Terana, S., Nakahara, T., Fluoret, G. & Bergman, R. N. J. biol. Chem. 253,
3219-3229 (1978).

64. Jard, S., Roy, C., Barth, T., Rajerson, R. & Bockaert, J. Adv. Cyclic Nucleotide Res. 5,
31-52 (1975).

65. Brostrom, C. O., Huang, Y., Breckenridge, B. McL. & Wolff, D. J. Proc. natn. Acad. Sci.
U.S.A. 72, 64-68 (1975).

66. Moss, J. & Vaughan, M. Proc. natn. Acad. Sci. U.S.A. 74, 4396-4400 (1977).

67. Pecker, F. & Hanoune, J. J. biol. Chem. 252, 2784-2786 (1977).

68. Egan, J. J., Majeska, R. J. & Rodan, G. A. Biochem. biophys. Res. Commun, 80, 176-182
(1978).

69. Katz, M. S., Kelly, T. M., Pineyro, M. A. & Gregerman, R. IL. J. Cyclic Nucleotide Res. 5,
389-407 (1978).

70, Edelman, G. M. Science 197, 218-226 (1976).

71. Insel, P. A. & Kennedy, M. J. Nature 273, 471-472 (1978).

72. Rudolf, S. A., Greengard, P. & Malawista, §. E. Proc. natn. Acad. Sci. U.S.A. 74,
3404-3408 (1977).

73. Cassel, D., Levkovitz, H. & Selinger, Z. J. Cyclic Nucleotide Res. 3, 393-406 (1977).

74. Cassel, D. & Selinger, Z. Biochem. biophys. Res. Commun. 77, 868-873 (1977).

75, Rendell, M., Rodbell, M. & Berman, M. J. biol. Chem. 252, 7909-7912 (1977).

76. Salomon, Y., Lin, M. C., Londos, C., Rendell, M. & Rodbell, M. J. bial. Chem. 250,
4239-4245 (1975).

77, Harwood, J. P., Low, H. & Rodbell, M. J. biol. Chem. 248, 6239-6245 (1973).

78. Eckstein, F., Cassel, D., Levkovitz, H., Lowe, M. & Selinger, Z. J. biol. Chem. 254,
9829-9834 (1979). :

79. Cassel, D., Eckstein, F., Lowe, M. & Selinger, Z. J. biol. Chem. 254, 9835-9838 (1979).

80. Rubalcava, B. & Rodbell, M. J. biol. Chem. 248, 3831-3839 (1973).

81. Lad, P.M., Preston, S., Welton, A. F., Nielsen, T. B. & Rodbell, M. Biochim. biophys. Acta
§§1, 368-391 (1979).

82. Welton, A. F. et al. Biochim. biophys. Acta 522, 625-639 (1978).

83. Blackmore, P. F., Brumby, F. T., Marks, J. L. & Exton, J.H.J. biol. Chem. 253, 4851-4858
(1978).

84. Glossman, H., Baukal, A. & Catt, K. J. J. biol. Chem. 249, 664-666 (1974).

85. Walaas, O., Walaas, E., Lystad, E., Alertsen, A. R. & Harn, R. S. Molec. ceil. Endocr. 16,
45-55 (1979).

86. Chan, T. M. & Exton, J. H. J. biol. Chem. 252, 8645-8651 (1977).

87. Kebabian, J. W. & Caln, D. B. Nature 277, 93-96 (1979).

88. Kebabian, J. W., Petzold, G. L. & Greengard, P. Proc. natn, Acad. Sci. U.S.A. 69,
2145-2149 (1972).

89. Spano, P. F., Govoni, S. & Trabucchi, M. Adv. Biochem. Psychopharmac. 19, 155-166
(1977).
90.
91.
92.
93.
94.
95.
96.
97.

98.
99.

100.

VanInwegen, R. G., Strada, $. J. & Robison, G. A. Life Sci. 16, 1875-1876 (1975).

Puri, §. K., Cochin, J. & Volicer, L. Life Sci. 16, 759-768 (1975).

Clouet, D. H. & Iwatsubo, H. Life Sci. 17, 35-40 (1975).

Jard, S. & Bockaert, J. Physiol. Rev. §5, 489-536 (1975).

Kirk, C. S. & Hems, D. A. FEBS Lett. 47, 128-131 (1974).

Macdermot, J. et al. Proc. natn. Acad. Sci. U.S.A. 76, 1135-1139 (1979).

Gardner, J. D. A. Rev. Physiol. 41, 55-66 (1979).

Hirata, F., Strittmatter, W. J. & Axelrod, J. Proc. natn. Acad. Sci. U.S.A. 76, 368-372
(1979).

Moore, W. V. & Wolff, J. J. biol. Chem. 248, 5705-5711 (1973).

Lefkowitz, R. J., Mullikin, D., Wood, C. L., Gorder, T. B. & Mukherjee, C. J. biol. Chem.
252, 5295-5303 (1977).

Zahniser, N. R. & Molinoff, P. B. Nature 275, 453-454 (1978).

101.
102.
103.
104.
105.
106.
107.

108.

Creese, I., Usdin, T. & Snyder, §. H. Nature 278, 577-578 (1978).

Watanabe, A. M. etal. J. biol. Chem. 253, 4833-4836 (1978).

Berrie, C. P., Birdsall, N. J. M., Burgen, A. S. V. & Hulme, S. C. Biochem. biophys. Res.
Commun. 87, 1000-1005 (1979).

Pert, C. B. & Taylor, D. in Endogenous and Exogenous Opiate Agonists and Antagonists
(ed. Way, E. L.) (Pergamon, Oxford, in the press).

Blume, A. J., Boone, G. & Lichtshtein, D. in Modulators, Mediators, and Specifiers in Brain
Function (eds Erlich, Y. H., Volavka, J., Davis, L. F. & Brunngreben, E.) 163-173
(Plenum, New York, 1979),

U'Prichard, D. C. & Snyder, S. H. J. biol. Chem. 283, 3444-3452 (1978).

Glossman, H. & Presek, P. Naunyn-Schmiedebergs Archs Pharmak. 306, 67-73
(1979).

Childers, S. R. & Snyder, S. H. Life Sci. 23, 759-762 (1978).

Printed in Great Britain by Macmillan Production Ltd.. Basingstoke. Hampshire