historical perspective The origins of protein phosphorylation Philip Cohen The reversible phosphorylation of proteins is central to the regulation of most aspects of cell func- tion but, even after the first protein kinase was identified, the general significance of this discovery was slow to be appreciated. Here I review the discovery of protein phosphorylation and give a per- sonal view of the key findings that have helped to shape the field as we know it today. he days when protein phosphorylation was an abstruse backwater, best talked Tabout between consenting adults in private, are over. My colleagues no longer cringe on hearing that “phosphorylase kinase phosphorylates phosphorylase” and their eyes no longer glaze over when a “”kinase kinase kinase” is mentioned. This is because protein phosphorylation has gradu- ally become an integral part of all the sys- tems they are studying themselves. Indeed it would be difficult to find anyone today who would disagree with the statement that “the reversible phosphorylation of proteins regu- lates nearly every aspect of cell life”. Phosphorylation and dephosphorylation, catalysed by protein kinases and protein phosphatases, can modify the function of a protein in almost every conceivable way; for Carl and Gerty Cori, the 1947 Nobel Laureates. Picture: Science Photo Library. example by increasing or decreasing its bio- logical activity, by stabilizing it or marking it for destruction, by facilitating or inhibiting movement between subcellular compart- so long before its general significance liver enzyme that catalysed the phosphory- ments, or by initiating or disrupting pro- was appreciated? lation of casein3. Soon after, Fischer and tein–protein interactions. The simplicity, Krebs4,5, as well as Wosilait and Sutherland6, flexibility and reversibility of phosphoryla- found that the interconversion of phospho- tion, coupled with the ready availability of Regulating by phosphorylation rylase b to phosphorylase a involved a ATP as a phosphoryl donor, explains its In the late 1930s Carl and Gerty Cori dis- phosphorylation/dephosphorylation mech- selection as the most general regulatory covered that there were two forms of glyco- anism. In particular, Fischer and Krebs4,5 device adopted by eukaryotic cells. gen phosphorylase (called b and a), the demonstrated that the b form could be con- It is thought that perhaps 30% of the enzyme that catalyses the rate-limiting step verted to the a form in the presence of Mg- proteins encoded by the human genome of glycogenolysis. Phosphorylase b was only ATP and an enzyme they termed phospho- contain covalently bound phosphate, and active in the presence of 5′ AMP, whereas rylase kinase4,5. Phosphorylase kinase was abnormal phosphorylation is now recog- phosphorylase a was active in the absence of subsequently shown to catalyse the transfer nized as a cause or consequence of many this nucleotide. They reasoned (incorrectly) of the γ-phosphoryl group of ATP to a spe- human diseases. A number of naturally that phosphorylase a must contain tightly cific serine residue on phosphorylase b 7. occurring toxins and tumour promoters bound 5′ AMP, and that the enzyme that The reconversion of phosphorylase a to exert their effects by targeting particular converts phosphorylase a to phosphorylase phosphorylase b was therefore catalysed by protein kinases and phosphatases. A topical b, discovered in 1943 (ref. 2), must catalyse a ‘phosphate-releasing’ (or PR!) enzyme, example is the cyclic heptapeptide micro- the removal of 5′ AMP.The effect of 5′ AMP today called protein phosphatase 1 to reflect cystin, which has just been listed as a “noti- on phosphorylase b was the first example of its much wider use in cell regulation8. fiable dangerous substrance”, along with allosteric activation, but, because this term In 1950, Earl Sutherland showed that anthrax, in the Anti-terrorism, Crime and would not be coined for another 20 years, glycogenolysis could be stimulated if liver Security Act of 2001 recently approved by they called the a-to-b converting enzyme slices were incubated with adrenalin or the British parliament. Microcystin, pro- ‘prosthetic-group-removing’ (or PR) glucagon; he subsequently showed that the duced by toxic blue-green algae, is a potent enzyme2. But the Coris’ never demonstrated activity of phosphorylase a was increased hepatotoxin and liver carcinogen that that PR enzyme released 5′ AMP from under these conditions (reviewed in ref. 9). inhibits members of one of the major fam- phosphorylase a and, although they received This was the first demonstration that a hor- ilies of protein phosphatases1. a Nobel Prize in 1947 for “discovering the mone could influence the activity of a spe- In view of these developments, it seems course of the catalytic conversion of glyco- cific enzyme, although the response was lost timely to reflect on the early days of gen”, many years passed before the true if the liver slices were homogenized. But, research on protein phosphorylation. How nature of the reaction was discovered. when the activation mechanism of phos- was this phenomenon originally discovered Protein kinase activity was first observed phorylase was discovered, it became obvi- as a control mechanism and why did it take in 1954 when Gene Kennedy described a ous that Mg-ATP would be necessary for NATURE CELL BIOLOGY VOL 4 MAY 2002 http://cellbio.nature.com E127 © 2002 Nature Publishing Group historical perspective adipose tissue and how glucagon inhibits Adrenalin Electrical glycolysis in the liver. But the widespread excitation distribution of PKA in animal tissues and other organisms suggested an even wider range of functions19. More substrates were cAMP identified, such as cardiac troponin I (ref. 20) and phospholamban21, which explained how adrenalin regulates the rate and force of heart-muscle contractility. This cAMP-dependent protein kinase extended the involvement of phosphoryla- tion to proteins that are not enzymes, although the demonstration by Tom Langan in 1969 that PKA phosphorylates histone 22 Phosphorylase Phosphorylase H1 at a specific serine residue had already kinase Kinase hinted at this possibility. (inactive) (active) The first calmodulin-dependent protein kinases were identified in the late 1970s and included myosin light-chain kinase23, phos- 15 Ca2+ phorylase kinase and calmodulin-depend- ent protein kinases I and II in the brain24. The subsequent realization that calmod- Phosphorylase b Phosphorylase a ulin-dependent protein kinase II has multi- 2+ (active) (active) ple functions in Ca -signalling akin to PKA (ref. 25), and especially the discovery that protein kinase C (ref. 26) is activated by the second messenger diacylglycerol, broad- ened the concept of second-messenger- Glycogenolysis dependent protein kinases. Some of the major serine/threonine- specific protein phosphatases were classi- Figure 1 The glycogenolytic cascade in mammalian skeletal muscle. Adrenalin stimulates fied during the late 1970s and early 1980s the production of 3′ 5′ cyclic adenosine monophosphate (cAMP) leading the sequential (ref. 8), and mechanisms by which they are activation of cAMP-dependent protein kinase and phosphorylase kinase. The latter con- regulated began to be identified. Prominent verts glycogen phosphorylase from the inactive dephosphorylated b form to the active among these was the characterization in phosphorylated a form, stimulating glycogenolysis in advance of an increased energy 1981 of the calmodulin-dependent protein demand. The activity of phosphorylase kinase also depends on calcium ions and is phosphatase 2B (also termed calcineurin)27, therefore also switched on during muscle contraction. This provides energy (via the which 10 years later was shown to be the 28 breakdown of glycogen) to sustain muscle contraction. target for cyclosporin , the immunosup- pressant drug that made organ transplants possible. In 1975, PKA was shown to phosphory- activation; addition of Mg-ATP did indeed Ca2+ receptors of eukaryotic cells) was one late peptides in proteolytic digests of myelin restore the response to hormones. The of its subunits15. These findings explained basic protein29, and this led to the realization reconstruction of a hormone response in a how glycogenolysis and muscle contraction that PKA phosphorylates serine residues in cell-free system was a major breakthrough were synchronized (Fig. 1). But by the end specific amino-acid sequence motifs30,31. that led to the discovery that adrenalin of the 1960s, 15 years after phosphorylase These studies paved the way for the devel- exerted its effects by generating a small, kinase had been discovered, phosphoryla- opment of synthetic peptide substrates that heat-stable factor later identified as 3′5′ tion was still thought of as a rather special- have been a key technical advance in the cyclic adenosine monophosphate (or cyclic ized control mechanism largely confined to study of protein phosphorylation. AMP). The remarkable story of how the first the regulation of one metabolic pathway In retrospect, the determination of the ‘second messenger’ was identified is beauti- (glycogen metabolism). amino-acid sequence of the first protein fully described in the first chapter of Cyclic kinase (PKA) in the early 1980s (ref. 32) AMP (ref. 9) published in 1970, the year was more significant than it seemed at the before Sutherland received a Nobel Prize. Phosphorylation develops time (at least to me!), because it allowed It took much longer before other impor- It was through the 1970s