Inhibition of pRb phosphorylation and cell-cycle progression by a 20-residue peptide derived from p16CDKN2/INK4A.

BACKGROUND: The CDKN2/INK4A tumour suppressor gene is deleted or mutated in a large number of human cancers. Overexpression of its product, p16, has been shown to block the transition through the G1/S phase of the cell cycle in a pRb-dependent fashion by inhibiting the cyclin D-dependent kinases cdk4 and cdk6. Reconstitution of p16 function in transformed cells is therefore an attractive target for anti-cancer drug design. RESULTS: We have identified a 20-residue synthetic peptide--corresponding to amino acids 84-103 of p16--that interacts with cdk4 and cdk6, and inhibits the in vitro phosphorylation of pRb mediated by cdk4-cyclin D1. The amino-acid residues of p16 important for its interaction with cdk4 and cdk6 and for the inhibition of pRb phosphorylation were defined by an alanine substitution series of peptides. In normal proliferating human HaCaT cells and in cells released from serum starvation, entry into S phase was blocked by the p16-derived peptide when it was coupled to a small peptide carrier molecule and applied directly to the tissue culture medium. This cell-cycle block was associated with an inhibition of pRb phosphorylation in vivo. CONCLUSIONS: These results demonstrate that a p16-derived peptide can mediate three of the known functions of p16: firstly, it interacts with cdk4 and cdk6; secondly, it inhibits pRb phosphorylation in vitro and in vivo; and thirdly, it blocks entry into S phase. The fact that one small synthetic peptide can enter the cells directly from the tissue culture medium to inhibit pRb phosphorylation and block cell-cycle progression makes this an attractive approach for future peptidometic drug design. Our results suggest a novel and exciting means by which the function of the p16 suppressor gene can be restored in human tumours.

Cell-cycle arrest and inhibition of Cdk4 activity by small peptides based on the carboxy-terminal domain of p21WAF1.

BACKGROUND: A common event in the development of human neoplasia is the inactivation of a damage-inducible cell-cycle checkpoint pathway regulated by p53. One approach to the restoration of this pathway is to mimic the activity of key downstream effectors. The cyclin-dependent kinase (Cdk) inhibitor p21(WAF1) is one such molecule, as it is a major mediator of the p53-dependent growth-arrest pathway, and can, by itself, mediate growth suppression. The primary function of the p21(WAF1) protein appears to be the inhibition of G1 cyclin-Cdk complexes. Thus, if we can identify the region(s) of p21(WAF1) that contain its inhibitor activity they may provide a template from which to develop novel anti-proliferative drugs for use in tumours with a defective p53 pathway. RESULTS: We report on the discovery of small synthetic peptides based on the sequence of p21(WAF1) that bind to and inhibit cyclin D1-Cdk4. The peptides and the full-length protein are inhibitory at similar concentrations. A 20 amino-acid peptide based on the carboxy-terminal domain of p21(WAF1) inhibits Cdk4 activity with a concentration for half-maximal inhibition (l0.5) of 46 nM, and it is only four-fold less active than the full-length protein. The length of the peptide has been minimized and key hydrophobic residues forming the inhibitory domain have been defined. When introduced into cells, both a 20 amino-acid and truncated eight amino-acid peptide blocked phosphorylation of the retinoblastoma protein (pRb) and induced a potent G1/S growth arrest. These data support a physiological role for the carboxyl terminus of p21(WAF1) in the inhibition of Cdk4 activity. CONCLUSIONS: We have discovered that a small peptide is sufficient to mimic p21(WAF1) function and inhibit the activity of a critical G1 cyclin-Cdk complex, preventing pRb phosphorylation and producing a G1 cell-cycle arrest in tissue culture cell systems. This makes cyclin D1-Cdk4 a realistic and exciting target for the design of novel synthetic compounds that can act as anti-proliferative agents in human cells.

Characterization of the cyclin-dependent kinase inhibitory domain of the INK4 family as a model for a synthetic tumour suppressor molecule.

We have previously shown that a 20 amino acid peptide derived from the third ankyrin-like repeat of the p16CDKN2/INK4a (p16) tumour suppressor protein (residues 84-103 of the human p16 protein) can bind to cdk4 and cdk6 and inhibit cdk4-cyclin D1 kinase activity in vitro as well as block cell cycle progression through G1. Substitution of two valine residues corresponding to amino acids 95 and 96 (V95A and V96A) of the p16 peptide reduces the binding to cdk4 and cdk6 and increases its IC0.5 for kinase inhibition approximately threefold when linked to the Antennapedia homeodomain carrier sequence. The same mutations increase the IC0.5 approximately fivefold in the p16 protein. Substitution of aspartic acid 92 by alanine instead increases the binding of the peptide to cdk4 and cdk6 and the kinase inhibitory activity. The p16 peptide blocks S-phase entry in non-synchronized human HaCaT cells by approximately 90% at a 24 microM concentration. The V95A and V96A double substitution minimizes the cell cycle inhibitory capacity of the peptide whereas the D92A substitution increases its capacity to block cell cycle progression. A deletion series of the p16 derived peptide shows that a 10 residue peptide still retains cdk4-cyclin D1 kinase and cell cycle inhibitory activity. The p16 peptide inhibited S-phase entry in five cell lines tested, varying between 47-75%, but had only a limited (11%) inhibitory effect in the pRb negative Saos-2 cells at a concentration of 24 microM. Like the full length p16 protein, the p16 peptide does not inhibit cyclin E dependent cdk2 kinase activity in vitro. These data suggest that acute inhibition of CDK-cyclin D activity by a peptide derived from the INK4 family will stop cells in late G1 in a pRb dependent fashion.

Differential post-translational modification of the tumour suppressor proteins Rb and p53 modulate the rates of radiation-induced apoptosis in vivo.

Ionizing radiation induces p53-dependent apoptosis in the spleen, providing a model system to study p53 regulated events in a normal cell type. We have developed an in vivo model that identifies genetic differences in the regulation of p53-mediated apoptosis and addresses whether altered post-translational events in the p53-p21/Rb axis modulate the sensitivity of cells to radiation-induced cell death in vivo. Splenocytes from mice with distinct genetic backgrounds (DBA/2 and C57BL/6) exhibit differences in the rate of apoptosis. Whilst no obvious strain differences in protein levels of Bcl-2 or the cyclin-CDKs were observed, early post-translational regulatory events in the p53-p21/Rb axis showed striking differences in the two mouse strains. Cells from C57BL/6 animals undergo more rapid apoptosis after irradiation resulting from elevated levels and rapid induction of p53, pronounced Rb-cleavage, and the absence of a sustained induction of p21. In contrast, cells from DBA/2 animals have a reduced rate of apoptosis following irradiation with elevated levels of hyperphosphorylated Rb and a sustained induction of the p21 protein that is coincident with the C-terminal phosphorylation of p53. These data suggest that quantitative differences in the level of p21 protein can affect the rate of apoptosis in vivo, consistent with the view that p21 is an anti-apoptotic effector of p53. However, striking differences in the Rb protein-caspase cleavage or hyperphosphorylation-in the same cell type, but in different genetic backgrounds, demonstrates that p53-dependent apoptosis can be modulated in vivo by genetic factors that impinge upon the pro- or anti-apoptotic potential of Rb. In addition, we show that Rb cleavage is p53-dependent and that its phosphorylation status can be uncoupled from p21 expression. This study highlights the possibility that genetic factors can be identified that affect differential sensitivity of cells to ionizing radiation in vivo.

Effects of adrenaline on the turnover of lipoprotein lipase in rat adipose tissue.

The mechanisms by which adrenaline brings about a reduction in the lipoprotein lipase activity of adipose tissue in vitro were investigated. The incorporation of [3H]leucine into lipoprotein lipase was measured during 1-h pulse incubations of rat epididymal fat bodies that had been preincubated for 4 h in the presence of glucose, insulin and dexamethasone. When adrenaline was added to the incubation medium at the start of the pulse, the incorporation of [3H]leucine was markedly reduced, suggesting that the rate of the enzyme's synthesis had decreased. On the other hand, the degradation of lipoprotein lipase, as measured by the loss of 3H-labelled enzyme protein during pulse-chase incubations of the epididymal fat bodies, was found to be significantly increased by the addition of adrenaline to the incubation medium at the start of the chase period. It is concluded that adrenaline is able both to inhibit the synthesis of lipoprotein lipase and to stimulate its degradation.

Immunological evidence that HMG-CoA reductase kinase-A is the cauliflower homologue of the RKIN1 subfamily of plant protein kinases.

Three different antibodies against the RKIN1 and BKIN12 gene products from rye and barley recognized the 58 kDa subunit of HMG-CoA reductase kinase-A (HRK-A) from Brassica oleracea on Western blots. HRK-A was also detected by an antipeptide antibody in enzyme-linked immunoassays, and this was competed by the peptide antigen. HRK-A was not recognized by antibodies against plant, mammalian and Saccharomyces cerevisiae relatives of RKIN1, i.e. wheat PKABA1, rat AMP-activated protein kinase and S. cerevisiae Snf1p. RKIN1/HMG-CoA reductase kinase-A are now among the first protein kinases in plants to be well characterized at both the molecular and biochemical levels.

Specificity determinants for the AMP-activated protein kinase and its plant homologue analysed using synthetic peptides.

Inspection of sequences around sites phosphorylated by the AMP-activated protein kinase (AMP-PK), and homologous sequences from other species, indicates conserved features. There are hydrophobic residues (M, V, L, I) at P-5 and P+4, and at least one basic residue (R, K, H) at P-2, P-3 or P-4. The importance of these residues has been established for AMP-PK and its putative plant homologue using a series of synthetic peptides. These results confirm the functional similarity of the animal and plant kinases, and suggest that the required motif for recognition of substrate by either kinase is M/V/L/I-(R/K/H,X,X)-X-S/T-X-X-X-M/V/L/I.

Growth hormone deficiency in Sturge-Weber syndrome.

Sturge-Weber syndrome (SWS) is a disorder involving central nervous system abnormalities that may increase the risk of hypothalamic-pituitary dysfunction. Records of 19 patients with suspected growth hormone deficiency (GHD), identified from a registry of 1653 patients with SWS, were reviewed; nine patients with GHD were found.

p21: structure and functions associated with cyclin-CDK binding.

The cyclin dependent kinase inhibitor, p21, is a multifunctional protein involved in coordinating the cellular response to negative growth signals. Induced by cellular damage under the transcriptional control of the p53 tumour suppressor protein, p21 interfaces with a number of cellular proteins involved in growth control. Although p21 has a diverse range of activities, from assembly factor to transcriptional modulator, its ability to interact with and regulate the activity of the cyclin dependent protein kinases is paramount to many of these functions.

Discoordinate regulation of contractile protein gene expression in the senescent rat myocardium.

The myocardium is a highly adaptive tissue, as evidenced by phenotypic alterations throughout development and under conditions of altered hemodynamic load. With pressure overload, the myocardium displays adult-to-fetal transitions in expression of contractile and non-contractile proteins. Most intriguing is the fact that many of these transitions are also observed in the senescent heart. The purpose of this work was to establish if the thin filament regulatory proteins, troponin I and troponin T, exhibit reexpression of early developmental isoforms, suggestive of coordinate reprogramming of contractile protein isoform expression. As a functional index of reexpression of the early isoform of troponin I, slow skeletal troponin I, myofibrils were isolated from 12 and 24-month-old Fischer 344 rat ventricles and assayed for myofibrillar ATPase activity at pH 7.0 and 6.5. Both preparations displayed rightward shifts in Ca-ATPase relationships with no differences between groups. SDS-PAGE and Western blot analysis showed that whereas myosin heavy chain expression underwent a transition to predominance of the early development isoform, beta-myosin heavy chain, there was no reexpression of the fetal isoforms of either troponin I or troponin T in the rat heart at 24 months of age. Northern blot analysis using cDNA probes specific for cardiac or slow skeletal troponin I also confirmed the lack of slow skeletal reexpression in the 24-month ventricle. These results are significant in that they demonstrate a lack of coordinate expression of contractile protein isoforms under myocardial adaptation to the aging process.

Evidence for an arginine residue at the allosteric sites of spinach leaf ADPglucose pyrophosphorylase.

The covalent modification of spinach leaf ADPglucose pyrophosphorylase leads to inactivation of both activator-stimulated and -unstimulated activity. Inactivation can be prevented if either the activator 3PGA or the inhibitor Pi are present during the modification. Pi proved to be more effective at protecting the enzyme from inactivation as it afforded 50% protection at 51 microM compared to 50% protection by 405 microM 3PGA. Partial modification of the enzyme using [14C]-phenylglyoxal leads to a decrease in both Vmax, A0.5 and a decrease in the ability of the 3PGA to stimulate the enzyme's activity. Modification increased the enzyme's susceptibility to inhibition by Pi and completely abolished the cooperative binding of Pi seen in the unmodified enzyme in the presence of 3PGA. Thus, phenylglyoxal appears to interfere, with the normal allosteric regulation of ADPglucose pyrophosphorylase from spinach leaf. Greater than 90% of the enzyme's activity is lost when 7.2 mol [14C]-phenylglyoxal are bound per mole of tetramer and this label is present in both the larger and small subunits. In addition, inactivation appears to involve two different arginine residues having different rates of modification.

Fructose 2,6-bisphosphate and the climacteric in bananas.

This work was done to test the view that there is a marked rise in the content of fructose 2,6-bisphosphate during the climacteric of the fruit of banana (Musa cavendishii Lamb ex. Paxton). Bananas were ripened in the dark in a continuous stream of air in the absence of exogenous ethylene. CO2 production and the contents of fructose 2,6-bisphosphate and sucrose were monitored over a 15-day period. A range of extraction procedures for fructose 2,6-bisphosphate were compared. Recovery of fructose 2,6-bisphosphate added to samples of unripe fruit varied from poor to unmeasurable. Recoveries from samples of ripe fruit were high. It is argued that this differential recovery of fructose 2,6-bisphosphate undermines claims that the amount of this compound increases at the climacteric. When recoveries are taken into account, our data suggest that there is no major change in fructose 2,6-bisphosphate content during the onset of the climacteric in bananas.

Human and plant proliferating-cell nuclear antigen have a highly conserved binding site for the p53-inducible gene product p21WAF1.

The mechanism(s) whereby higher plants respond to environmental agents that damage their DNA, which leads to the arrest of cell division, is poorly understood. In mammalian cells, the tumour-suppressor protein p53 plays a central role in a DNA-damage-induced cell-cycle-checkpoint pathway by induction of transcription of a set of gene products that have a direct role in a DNA-damage-induced cell-cycle growth arrest. One such protein, p21WAF1, has been shown to be essential for radiation-induced growth arrest. There appear to be at least two cellular targets of p21WAF1 during checkpoint control, the G1-cyclin-dependent kinases (CDK) and proliferating-cell nuclear antigen (PCNA). The aim of the research reported here was to determine whether the interactions between the human growth inhibitor p21WAF1 and PCNA from plants and humans are conserved. If so, this would suggest that modulation of PCNA activity may play an important role in plant responses to DNA damage and would imply that functional homologue(s) of p21WAF1 exist in plants. We show that the p21WAF1-interaction domain of PCNA is conserved between humans and plants. A peptide that contains the site of human p21WAF1 that binds human PCNA has been used to precipitate PCNA from crude pea (Pisum savitum) extracts. We used the p21WAF1 peptide as an affinity matrix and showed that pea PCNA bound in a specific high-affinity manner. This finding was used to develop a purification protocol that allowed PCNA from plant tissue to be purified to homogeneity. Pure pea PCNA forms a stable complex with full-length human p21WAF1 and the specific amino acids of p21WAF1 required for the interaction have been identified. The critical residues were identical to those required for binding to human PCNA, which indicates that the interaction of human p21WAF1 with PCNA is highly conserved at each amino acid position between pea and human.

Isoform specific interactions of troponin I and troponin C determine pH sensitivity of myofibrillar Ca2+ activation.

We investigated whether differences in isoforms of troponin I (TnI) and troponin C (TnC) can account for the greater inhibition of Ca(2+)-dependent MgATPase activity by acidic pH in cardiac (c) than in fast skeletal (fs) myofilaments. We studied fast skeletal myofibrils from which whole Tn was extracted by displacement with excess fsTnT (the tropomyosin binding subunit of Tn) followed by reconstitution with TnC-TnI. Exchange of fsTnI with cTnI did not alter the effect of a drop in pH from 7.0 to 6.5 on the relation between pCa (-log[Ca2+]) and MgATPase activity of fast skeletal myofibrils. Exchange of fsTnC with cTnC did, however, induce an increase in the effect of this same pH change on Ca2+ activation. Yet, the pH sensitivity of Ca2+ activation of fast skeletal myofibrils containing cTnC was not as great as that of native cardiac myofibrils. However, when both fsTnC and fsTnI of fast skeletal myofibrils were replaced by cTnC-cTnI, there was a pH-induced shift in Ca2+ sensitivity similar to that of cardiac myofibrils. In studies using fluorescent probes, both pure fsTnC and pure cTnC showed decreased Ca2+ binding as pH was lowered. This decrease was potentiated in the fsTnC-fsTnI and cTnC-cTnI complexes. However, the effect of acidic pH was the same in fsTnC and the hybrid complex, fsTnC-cTnI, and in cTnC and the hybrid complex, cTnC-fsTnI. Thus, isoform specific interactions between TnI and TnC appear important in the differential response of skeletal and cardiac myofilaments to acidosis.

Glycolysis at the climacteric of bananas.

This work was carried out to investigate the relative roles of phosphofructokinase and pyrophosphate-fructose-6-phosphate 1-phosphotransferase during the increased glycolysis at the climacteric in ripening bananas (Musa cavendishii Lamb ex Paxton). Fruit were ripened in the dark in a continuous stream of air in the absence of ethylene. CO2 production, the contents of glucose 6-phosphate, fructose 6-phosphate, fructose 1,6-bisphosphate, phosphoenolpyruvate and PPi; and the maximum catalytic activities of pyrophosphate-fructose-6-phosphate 1-phosphotransferase, 6-phosphofructokinase, pyruvate kinase and phosphoenolpyruvate carboxylase were measured over a 12-day period that included the climacteric. Cytosolic fructose-1,6- bisphosphatase could not be detected in extracts of climacteric fruit. The peak of CO2 production was preceded by a threefold rise in phosphofructokinase, and accompanied by falls in fructose 6-phosphate and glucose 6-phosphate, and a rise in fructose 1,6-bisphosphate. No change in pyrophosphate-fructose-6-phosphate 1-phosphotransferase or pyrophosphate was found. It is argued that phosphofructokinase is primarily responsible for the increased entry of fructose 6-phosphate into glycolysis at the climacteric.

Adverse effects of anabolic steroids.

Anabolic steroids are used therapeutically for various disorders and as ergogenic aids by athletes to augment strength, muscular development, and to enhance performance. There is a wide range of concomitant temporary and permanent adverse effects with steroid administration. Several well-documented adverse actions of these hormones may develop rapidly within several weeks or less (i.e. altered reproductive function) or require up to several years of steroid intake (i.e. liver carcinoma). More recent studies indicate that glucose intolerance, insulin resistance, increased cardiovascular disease risk profiles, cerebral dangers, musculoskeletal injuries, prostate cancer, psychosis and schizophrenic episodes, among others, accompany anabolic steroid intake. There is, at present, no evidence to support the claim that athletes are less susceptible to adverse effects than those individuals receiving hormone treatment in a clinical setting. Based on the available information which has accumulated primarily from cross-sectional, short term longitudinal, and case studies, there is a need: (a) to develop a comprehensive battery of specific and sensitive markers of adverse effects, particularly those that would be able to detect the onset of adverse actions; and (b) to conduct controlled long term longitudinal studies in order to fully understand the extensiveness and mechanisms involved in the occurrence of adverse effects.

Reversible phosphorylation at the C-terminal regulatory domain of p21(Waf1/Cip1) modulates proliferating cell nuclear antigen binding.

The p53-inducible gene product p21(WAF1/CIP1) plays a critical role in regulating the rate of tumor incidence, and identifying mechanisms of its post-translational regulation will define key pathways that link growth control to p21-dependent tumor suppression. A eukaryotic cell model system has been developed to determine whether protein kinase signaling pathways that phosphorylate human p21 exist in vivo and whether such pathways regulate the binding of p21 to one of its key target proteins, proliferating cell nuclear antigen (PCNA). Although human p21 expressed in Sf9 cells is able to form a complex with human PCNA, the inclusion of cell-permeable phosphatase inhibitors renders p21 protein inactive for PCNA binding. The treatment of this inactive isoform of p21 with alkaline phosphatase restores its binding to PCNA, suggesting that p21 expressed in Sf9 cells is subject to reversible phosphorylation at a key regulatory site(s). A biochemical approach was subsequently used to map the phosphorylation sites within p21, whose modification in vitro can inhibit p21-PCNA complex formation, to the C-terminal domain at residues Thr(145) or Ser(146). A phospho-specific antibody was developed that only bound to full-length p21 protein after phosphorylation in vitro at Ser(146), and this reagent was further used to demonstrate that the inactive isoform of p21 recovered from Sf9 cells treated with phosphatase inhibitors had been phosphorylated in vivo at Ser(146). These data identify the first phosphorylation site within the C-terminal regulatory domain of p21 whose modification in vivo modulates p21-PCNA interactions and define a eukaryotic cell model that can be used to study post-translational signaling pathways that regulate p21.

Strategies for manipulating the p53 pathway in the treatment of human cancer.

Human cancer progression is driven in part by the mutation of oncogenes and tumour-suppressor genes which, under selective environmental pressures, give rise to evolving populations of biochemically altered cells with enhanced tumorigenic and metastatic potential. Given that human cancers are biologically and pathologically quite distinct, it has been quite surprising that a common event, perturbation of the p53 pathway, occurs in most if not all types of human cancers. The central role of p53 as a tumour-suppressor protein has fuelled interest in defining its mechanism of function and regulation, determining how its inactivation facilitates cancer progression, and exploring the possibility of restoring p53 function for therapeutic benefit. This review will highlight the key biochemical properties of p53 protein that affect its tumour-suppressor function and the experimental strategies that have been developed for the re-activation of the p53 pathway in cancers.

v-Abl protein-tyrosine kinase up-regulates p21WAF-1 in cell cycle arrested and proliferating myeloid cells.

v-Abl protein-tyrosine kinase (PTK) promotes cell survival without cell proliferation in interleukin (IL)-3-deprived IC.DP premast cells (1). We now show that in these conditions v-Abl PTK transcriptionally up-regulated the cyclin-dependent kinase inhibitor (CDKI) p21(WAF-1) and inhibited CDK2 and CDK4. When readdition of IL-3 stimulated cell proliferation, p21(WAF-1) was inactivated as a CDKI despite maintenance of elevated protein level. p21(WAF-1) was also up-regulated yet was nonfunctional as a CDKI when v-Abl PTK was activated in cells maintained in IL-3, but this occurred without increased p21(WAF-1) transcription. Using a C-terminal epitope-specific p21(WAF-1) antibody, v-Abl PTK-mediated increase in p21(WAF-1) could be detected in intact cells only in the presence of IL-3. This indicated different binding partners of p21(WAF-1) and/or protein conformation in nondividing or proliferating cells, respectively. The binding of CDK2, CDK4, or proliferating cell nuclear antigen to p21(WAF-1) and its subcellular localization were unchanged in the presence or absence of IL-3. However, two-dimensional analysis revealed different forms of up-regulated p21(WAF-1) in IL-3-deprived, nondividing cells compared with IL-3-stimulated proliferating cells. These data demonstrate that elevation of the CDKI p21(WAF-1) is not always sufficient for cell cycle arrest and indicate an IL-3-sensitive pathway for the inactivation of p21(WAF-1) function as a CDKI.

Biochemical characterization of two forms of 3-hydroxy-3-methylglutaryl-CoA reductase kinase from cauliflower (Brassica oleracia).

We recently reported the existence of a protein kinase cascade in higher plants, of which the central component is a 3-hydroxy-3-methylglutaryl(HMG-)-CoA reductase kinase functionally related to mammalian AMP-activated protein kinase [MacKintosh, R. W., Davies, S. P., Clarke P. R., Weekes, J., Gillespie, S. G., Gibb, B. J. & Hardie, D. G. (1992) Eur. J. Biochem. 209, 923-931]. We have now purified this protein kinase 9000-fold from cauliflower inflorescences. During the course of this work we noticed a second minor form (form B) which separated from the major form (A) on ion exchange and gel filtration. Both forms phosphorylate the catalytic fragment of mammalian HMG-CoA reductase. Both forms are markedly inactivated by incubation with the reactive ATP analogue p-fluorosulphonylbenzoyl adenosine (FSO2PhCOAdo), and also by mammalian protein phosphatase 2C, indicating that form B, like form A, is activated by phosphorylation. Form A has an apparent native molecular mass of 200 kDa by gel filtration and, after labelling with [14C]FSO2PhCOAdo, of 150 kDa by electrophoresis in non-denaturing gels. The catalytic subunit was identified as a polypeptide of 58 kDa after labelling with [14C]FSO2PhCOAdo. Form B has an apparent native molecular mass of 45 kDa by gel filtration, and was identified as a polypeptide of 45 kDa after labelling with [14C]FSO2PhCOAdo and [gamma-32P]ATP. Using a series of variants of the synthetic peptide substrate, the substrate specificities of the two forms are similar but not identical. Form B does not appear to be a proteolytic fragment of form A, and we therefore propose that it represents a closely related member of the same protein kinase sub-family.