MOLECULAR MECHANISMS OF SIGNAL TRANSDUCTION THROUGH CYCLIC NUCLEOTIDES 
Jackie D. Corbin, Ph.D., Investigator 
Dr. Corbin's main interest is intracellular cyclic 
nucleotide receptors. Recent investigations have fo- 
cused on three projects: 1) cAMP-dependent pro- 
tein kinase, 2) cGMP-dependent protein kinase, and 
3) negative feedback control of cyclic nucleotide 
levels. 
I. cAMP-Dependent Protein Kinase (cAK). 
cAMP sites of the cAK from the fungus Mucor 
rouxii have been characterized through the study 
of the effects of cAMP and cAMP analogues on the 
phosphotransferase activity and through binding ki- 
netics. The tetrameric holoenzyme, w^hich contains 
two regulatory (R) and two catalytic (C) subunits, 
exhibited positive cooperativity in activation by 
cAMP, suggesting multiple cAMP-binding sites. Sev- 
eral other results indicated that the Mucor kinase 
contained two different cooperative cAMP-binding 
sites on each R subunit, with properties similar to 
those of the mammalian cAMP-dependent protein 
kinase. Under optimum binding conditions, the 
[^H]cAMP dissociation behavior indicated equal 
amounts of two components that had dissociation 
rate constants of 0.09 min"^ (site 1) and 0.90 min"^ 
(site 2) at 30°C. Two cAMP-binding sites could also 
be distinguished by C-8 cAMP analogues (site 1-se- 
lective) and C-6 cAMP analogues (site 2-selective); 
combinations of site 1- and site 2-selective ana- 
logues were synergistic in protein kinase activation. 
The two different cooperative binding sites were 
probably located on the same R subunit, since the 
proteolytically derived dimeric form of the enzyme, 
which contained one R and one C component, re- 
tained the salient properties of the untreated tetra- 
meric enzyme. Unlike any of the mammalian cyclic 
nucleotide-dependent isozymes described thus far, 
the Mucor kinase was much more potently acti- 
vated by C-6 cAMP analogues than by C-8 cAMP an- 
alogues. In the ternary complex formed by the na- 
tive Mucor tetramer and cAMP, only the two sites 1 
contained bound cAMI^ a feature that has also not 
yet been demonstrated for the mammalian cAK. 
II. cGMP-Dependent Protein Kinase (cGK). 
A. Structure of the cGMP-binding site(s). Mammal- 
ian cAK and cGK show considerable similarity in 
amino acid sequence, although they specifically bind 
different cyclic nucleotides. Results of cGMP ana- 
logue-binding experiments, combined with model- 
ing of the cGMP-binding sites by analogy to the struc- 
ture of the homologous catabolite gene activator 
protein, suggest that a threonine residue forms a hy- 
drogen bond with the of cGMP This threo- 
nine is invariant in all cGMP-binding domains, but 
the corresponding residue in 23 out of 24 cAMP- 
binding sites of protein kinases is alanine, which can- 
not form the same hydrogen bond. This alanine/thre- 
onine difference has the potential for discriminating 
between cAMP and cGMP and may be important in 
the evolutionary divergence of cyclic nucleotide- 
binding sites. 
B. Monomeric cGK. A form of cGK that was differ- 
ent from previously described cGK was purified 
from bovine aorta smooth muscle. The partial 
amino- terminal sequencing of this enzyme indi- 
cated that it was derived by endogenous proteolysis 
of the type ip isozyme of cGK. On SDS-PAGE, this 
form migrated as a smaller protein {M^ = 70,000) 
than the parent cGK {M^ = 80,000), and since the 
calculated nondenatured M^ was —89,000 com- 
pared with M^ = 170,000 for the dimeric native en- 
zyme, it represented a monomeric form of cGK. The 
monomer bound ~2 mol of [^HJcGMP per mole of 
monomer, although it had only one rapid compo- 
nent in [-^HJcGMP dissociation assays, as compared 
with one rapid and one slow component for the 
native cGK. The specific catalytic activity of the ki- 
nase was similar to that of the native enzyme, sug- 
gesting that the catalytic domain was essentially in- 
tact. The monomeric cGK incorporated significant 
'^P when incubated with Mg^+ and [7-^^P]ATP in 
the presence of cGMP, although the phosphoryla- 
tion proceeded at a slower rate than that obtained 
with native cGK. In contrast to previous reports of 
monomeric forms of cGK, this monomer was highly 
cGMP dependent, although it had a slightly higher 
(0.8 |jlM) for cGMP than that of the native en- 
zyme (0.4 |jlM) and a lower Hill coefficient (1.0 ver- 
sus 1.6 for the native enzyme). The cGMP depen- 
dence of the monomer did not decrease with 
dilution, implying that the cGMP dependence was 
not due to monomer-monomer interactions in the 
assay. The results indicated that the catalytic do- 
main, cGMP-binding domain(s), and inhibitory do- 
main of cGK interact primarily within the same sub- 
unit rather than between subunits of the dimer, as 
previously hypothesized for dimeric cGK. 
Continued 
45 
