Laboratory of Molecular Biophysics
Laboratory Journal 2002
Louise N. Johnson

Return to: Contents.

Louise N. Johnson

Structural studies on protein kinases and regulatory molecules of the cell cycle

Introduction

We are interested in the structural basis of the recognition processes that control biological interactions, especially those that mediate control by phosphorylation. Our work in the year 2001/2002 has focussed on the regulatory proteins of the cell cycle, that include CDK2/cyclin A, CDK2/cyclin E, CDK7/cyclin H, polo-like kinase and those proteins that mediate interactions of ubiquitinated proteins with the proteasome. We continue our special interest with phosphorylase kinase and its regulation by calcium/calmodulin. Our major technique is protein crystallography, which for some problems is combined with electron microscopy (see Section 11 by Catherine Vénien Bryan). We have also been involved in a programme of structure based drug design for CDK2/cyclin A (see section 2 by Jane Endicott and section 8 by Martin Noble) that has led to several publications in collaboration with colleagues at Newcastle University and AstraZeneca (1, 2). Our work on the radio-sensitising anti-cancer compound UCN-01 (7-hydroxystaurosporine) in complex with CDK2/cyclin A has also been published (3)

1. Cell cycle proteins

The eukaryotic cell cycle involves replication of the genetic material and the ancillary components of the genome and cell's biomass to yield two duplicate daughter cells. Progression through the cycle is driven by the sequential activities of different cyclin dependent protein kinases that require association with their respective cyclins and phosphorylation for activation. Selection of the right substrates is a key event in the organisation of both the temporal and spatial events that characterise the complex processes of cell growth and cell division. In response to a mitogenic signal, cells move from their quiescent phase, G0, to G1 through activities that are initiated by CDK4 and CDK6 whose association with their D type cyclins, synthesised in response to mitogens, is promoted by members of the p21^Cip1 class of CDK inhibitors. The major target of these CDKs is the tumour suppressor protein pRb whose phosphorylation results in relief of its transcriptional repression of the E2F family of transcription factors. Relief of E2F repression by pRb leads to transcription from promoters containing E2F sites, which include the promoter for cyclin E. Cyclin E in association with CDK2 initiates events that drive cells from G1 to S phase though phosphorylation of selected substrates. These substrates also include pRb whose hypophosphorylation by CDK2/cyclin E results in dissociation of pRb from the transcription factor E2F complexes and subsequent transcriptional activation of many of the genes involved in DNA synthesis and that of cyclin A. At the G1/S boundary, levels of cyclin A rise. As cells enter S phase, cyclin E becomes phosphorylated and is abruptly degraded by the Skp1/cullin/F-box complex ubiquitin mediated targeting to the proteasome. CDK2 in association with cyclin A drives cells through S phase, phosphorylating a wider range of substrates than CDK2/cyclin E, that include the tumour suppressor protein pRb, transcription factors such as MybB, ID2, ID3 and p53, and CDC6, the factor involved in the formation of the pre-replication complex that is required for initiation of DNA replication. CDK2/cyclin A substrates also include the transcription factor E2F itself, whose phosphorylation by CDK2/cyclin A results in inactivation of transcription, an important signal that blocks re-initiation of DNA replication and indicates exit from S phase. Cyclin A levels remain high through G2. As cells enter mitosis cyclin A associates with CDK1 but is then degraded through the anaphase promoting complex ubiquitination and targeting to the proteasome. Cyclin B then becomes active and in association with CDK1 drives cells through M phase, together with the activation of other kinases such as polo-like kinase, in processes for which some of the down stream substrates have been identified but by no means all.

1.1 CDK2/cyclin A recruitment site specificity

Nick R. Brown, Kin-Yip Cheng, Ed D. Lowe, and Martin E. M. Noble, in collaboration with Sheraz Gul, Ivo Tews and Stephen J. Gamblin (National Institute for Medical Research, Mill Hill).

During each stage of the cycle, the CDKs recognise specific substrates. The obligatory local epitope that is recognised by CDK2/cyclin A is Ser/Thr.Pro.X.Lys/Arg where Ser or Thr are the residues phosphorylated at site P0, X is any amino acid, and Lys/Arg denote a preference for basic residues in the residues C-terminal to the site of phosphorylation (P+3 site). A decade ago the substrates p107 and p130 were found to have an interaction domain that promoted their stoichiometric association with CDK2/cyclin E or CDK2/cyclin A and which is shared with the CDK2 inhibitors p21 and p27, although the functional significance of the association was not understood. It is now recognised that many other CDK2 substrates, such as pRb, p53, E2F1, human papilloma virus (HPV) replication factor E1, p220^NPAT, CDC6 and endomexin, contain a site that is remote from the site of phosphorylation and which contains the signature motif RXL or KXL. Such sites that are recognised by a specific hydrophobic binding site on the cyclin and confer selectivity. They provide an additional specificity that in part explains the selective phosphorylation of certain substrates by CDK2.

Structural studies from Nicola Pavletich's Laboratory with the p27^Cip1 inhibitor (4) and by us with a peptide from p107 (5) have shown that the RXL recognition site is located on the cyclin A molecule, involving recognition from residues located mostly on the helix H1 of the cyclin. These provide an extensive exposed non-polar site that is conserved in cyclins A, B, D and E. The site is some 40 Å from the catalytic site located by a substrate peptide in complex with phospho-CDK2/cyclin A. Experiments from Wade Harper's Laboratory with the p107 substrate of CDK2/cyclin A have shown that one of the roles of the RXL motif is to make an otherwise poor substrate a good substrate (6).

Two proposals have been put forward for the mechanism of action of the RXL motif. Firstly it is proposed that the RXL motif serves solely as an entropic localising effector, localising the CDK/cyclin on the substrate and serving to increase the local substrate concentration (7). Alternatively, it has been proposed that there is a direct trajectory between the catalytic site on CDK2 and the RXL recognition motif on the cyclin (8). In support of this Dutta and colleagues in a kinetic analysis have shown that with a synthetic substrate from CDC6 the phosphorylation site Ser⁷⁴.Pro.X.Lys is dependent on the integrity of R⁹⁴.X.L motif. Peptides with synthetic linkers that placed too short a distance between the RXL and the site of phosphorylation were not efficiently phosphorylated. Only when the synthesis linker mimicked the natural substrate length did phosphorylation occur efficiently.

RXL containing peptides effectively inhibit CDK activity, presumably by blocking access to the cyclin A recruitment site by the natural substrate. This property has been exploited in the design of anti-cancer drugs. Peptides derived from E2F were shown to preferentially induce transformed cells, whose E2F transcriptional machinery had been deregulated, to undergo apoptosis relative to non-transformed cells (9). These studies have provided a rational for the development of CDK2/cyclin A antagonists based on the RXL motif as antineoplastic agents.

Examination of the sequences around the RXL motif from a number of key substrates shows variation in sequence and in position of the RXL motif with respect to the site of phosphorylation. Most RXL motifs are C-terminal to the site of phosphorylation with a minimum separation observed to date of about 20 residues. In order to understand further the factors involved in the recognition of the RXL motives by cyclin A and the relative affinities of the different sites on different substrates, we have co-crystallised phospho-CDK2/cyclin A in complex with a peptide from E2F (residues 84-97). We have further carried out a systematic study of five 11-mer peptides from pRb, p107, E2F, P27 and p53, in crystallographic (2.3 Å resolution) and thermodynamic isothermal calorimetry studies, in order to aid recognition of common features and those factors that govern affinity (Figure 4.1).

Figure 1. Recruitment peptide sequences used in crystallographic binding studies with pCDK2/cyclin A. The RXL motif is shaded.

The results (which have been published recently (10)) show that the cyclin A recruitment site is dominated by recognition of the leucine from the RXL motif. In the five peptide complexes studied the position of the leucine and its contacts are conserved. The leucine docks into a non-polar pocket on the surface of the cyclin. The pocket is readily identified by display of the hydrophobic potential on the van der Waals surface with the combined programmes of Peter Goodford's GRID and with Martin Noble's Aesop. In the previously determined structures of p27^kip1 and p107 peptides bound to pCDK2/cyclin A, the phenylalanine, which follows the leucine in the RXLF motif, is involved in significant van der Waals contacts to the cyclin and the contacts also exploit this significant non-polar pocket. In E2F, p53 and pRb the corresponding residues adjacent C-terminal to the leucine are aspartic acid, methionine and arginine, respectively (Figure 1). We were interested therefore to compare the structure of the bound E2F, p53 and pRb peptides with those from p107 and p27 in order to understand how the site accommodates such a marked difference in sequence. In the E2F, p53 and pRb complexes, the peptide is oriented so that the respective Asp, Met or Arg is directed away from the cyclin and the hydrophobic pocket is filled by the next residue which is strictly non-polar (Leu in E2F, Phe in p53 and Phe in pRb) (Figure 2). The fact that the recruitment site accommodates not only the arginine and leucine of the RXL motif but also has a strong preference for a non-polar residue that may be directly adjacent to the leucine, (for example the phenylalanine in the p107 or p27 peptides RXLF) or next but one to the crucial leucine (for example the leucine in the E2F peptide RXLDL or phenylalanine in the p53 KXLMF or pRb KXLRF peptides) is a new observation. This hydrophobic residue makes important contributions to specificity.

Figure 4.2. A schematic diagram showing important polar and non-polar contacts for the complexes of pCDK2/cyclin A with the p107 and E2F peptides. The peptide bonds are black and those of cyclin A are grey.

Figure 2. ...more

For the E2F, p27 and p107 peptide complexes, the arginine of the RXL makes contact to a glutamate, Glu220, on the cyclin A, in crystal structures where glycerol was used as the cryoprotectant but in those crystals where 8 M formate was used the ionic interaction is less strong, presumably weakened by the high ionic strength formate. In those substrates that contain a KXL motif (e.g. p53 and pRb) there are no ionic interactions between the lysine and the cyclin and these substrates evidently bind more weakly than the RXL containing substrates as shown by the isothermal calorimetry measurements (Table 1). The sequences N-terminal to the RXL or KXL motifs have no conservation (Figure 4.1) yet the structural results show that each of the five peptides binds with a common conformation in main chain atoms. The different side chains are accommodated with different interactions. Thus the recruitment site is able to recognise diverse but conformationally constrained target sequences. These observations have implications for the further identification of physiological substrates of CDK2/cyclin A and the design of specific inhibitors.

Table 1. Isothermal calorimetry results for recruitment peptides bound to pCDK2/cyclin A.
Protein	Sequence	Binding constant (mM)
		11-mer	6-mer
p53	STSRHKKLMFK	2.9	2.9
pRb	PPKPLKKLRFD	1.8	2.3
p27	KPSACRNLFGP	0.9	2.5
E2F *	PVKRRLDLE	0.5	7.9
p107	AGSAKRRLFGE	0.3	2.4
Residues in bold are the sequences for the short peptide versions. The 'RXL' motif is underlined for each sequence * For the E2F peptide the longer version of the peptide contained 9 instead of 11 residues.

1.2 CDK2/cyclin E

Elena Dubinina and John Sinclair

Cyclin E is a G1 cyclin, which forms complexes with CDK2 and is essential for the S phase entry. The expression of cyclin E, its activity, substrate specificity, as well as the cellular localisation of CDK2/cyclin E complexes play a crucial role in the cell proliferation. The recruitment substrate-binding site on cyclin E, which is important in the recognition of the RXL containing substrates, is identical in amino acid sequence to that of cyclin A. Since CDK2 is the same in the CDK2/cyclin A and CDK2/cyclin E complexes why does the cell need
both cyclin A and cyclin E? The answer appears to be that each operates on different substrates and yet despite our extensive knowledge of CDK2/cyclin A substrate specificity, the differential selection of substrates by the two CDK2 complexes is hard to understand without a structure. There is considerable interest in solving the crystal structure of CDK2/cyclin E complex. This year we have made significant progress in expression of cyclin E for crystallographic studies.

Figure 3. ...more

Cyclin E has been expressed in Sf9 insect cells using a baculoviral construct kindly given by David Morgan (UCSF). The E. coli expression system based on the p-GEX polycistronic construct encoding the GST-CDK2 and Cak1 kinase was used for the expression of the phosphorylated GST-CDK2 protein (5). Purification of the pCDK2/cyclin E complex was achieved by the loading the GST-phospho-CDK2 containing E. coli cell lysate onto the glutathione-Sepharose 4B column. Following extensive washing of the column, the Sf9 insect cell lysate containing cyclin E was loaded to form the complex. The pCDK2/cyclin E complex was released from the column by glutathione elution. The complex was further purified by gel-filtration on a Superdex 75 column. A typical yield of pCDK2/cyclinE was 7-10 mg from 1.5 l of the CDK2 and 1 l of the cyclin E cultures respectively (Figure 3). This method provides a rapid purification procedure that yields suitable quantities of the protein complex for crystallization. Our initial attempts to crystallize purified pCDK2/cyclinE complex were performed using various crystallization screens.

Crystals of the intact CDK2/cyclin E complex did not readily appear. Therefore we have turned our attention to 'trimming' the cyclin E with the hope of removing more flexible regions. Through limited proteolysis with trypsin and subtilisin of the pCDK2/cyclinE complex, we have identified several proteolytically sensitive sites in the amino terminal end of human cyclin E. Based on the proteolytic sensitivity data we designed two baculoviral constructs for the expression of the truncated cyclin E variants, containing deletion of 16 and 35 N-terminal amino acids respectively. Each construct was used to express the corresponding truncated cyclin E protein in Sf9 insect cells. Currently, we are focused on purification of the pCDK2/truncated cyclin E complexes, which we plan to use in our further crystallization studies.

1.3 CDK Activating Kinase (CAK)

Nick R. Brown and John Sinclair

CAK is a positive regulator of CDK1, CDK2 and CDK4. The kinase phosphorylates CDKs on the threonine residue in the activation segment (Thr160 in CDK2). CAK is itself a member of the CDK family and is composed of CDK7, cyclin H and an assembly protein Mat1. In the presence of Mat1, activation of the CDK7/cyclin H complex does not require phosphorylation of the activation segment for activity. In the absence of Mat1, phosphorylation of two residues in the activation segment (Ser170 and Thr176 in Xenopus CDK7) is required for activity. CDK7 raises intriguing questions concerning substrate recognition. CDK7/cyclin H does not promote its own phosphorylation but is able to phosphorylate CDK2/cyclin A. Likewise CDK2/cyclin A is able to phosphorylate CDK7/cyclin H but does not phosphorylate itself. Garrett et al. (11) have produced a hybrid CDK in which the activation segment of CDK2 was replaced with the activation segment from CDK7. It was found that the CDK2 (CDK7-activation segment) hybrid was a substrate for CDK7/cyclin H but not for CDK2/cyclin A and the CDK2 (CDK7 activation segment) hybrid could phosphorylate CDK7 but not CDK2. The results demonstrate that sites elsewhere from the catalytic site on the CDK appear to be important for substrate recognition.
Two years ago, we reported the production in baculoviral infected insect cells of quantities of Xenopus CDK7/cyclin H for crystallisation trials. We were able to show by mass spectrometry that in the CDK7/cyclin H complex, CDK7 was fully phosphorylated on its two sites in the activation segment presumably by kinases present in the insect cells (12). This year we have continued studies with expression of human CDK7/cyclin H by co-infection of insect cells with the viruses encoding the respective proteins and with the third protein Mat1. We are grateful to Dr David O. Morgan (UCSF) for the generous gift of the viruses. However we found that, although the virus encoding the His-tagged cyclin H was stable and expressed well in insect cells, the virus encoding the CDK7 protein tended to be lost from the insect cells. After some work to confirm that this was the case, we find that other laboratories have experienced a similar loss of CDK7, possibly because the active kinase is toxic to cells. With generous gifts of new viruses we are developing alternative strategies to engineer inactive CDK7 for co-expression with cyclin H or for expression, purification and crystallisation trials of CDK7 by itself.

1.4 Polo-like kinase

John Sinclair in collaboration with Erich A. Nigg (Max Planck Institute for biochemistry, Martinsried)

Polo-like kinases (Plks) are important regulators of cell cycle progression at the initiation of and during mitosis. Named after the polo gene of Drosophila melanogaster, Plks are involved in the assembly and dynamics of the mitotic spindle apparatus and in the activation and inactivation of CDK/cyclin complexes. In mammalian cells, Plk1 protein levels increase as cells approach M phase, with the peak of phosphorylation activity reached during mitosis. Known substrates include Cdc25C phosphatase, cyclin B, a cohesin subunit of the mitotic spindle, subunits of the anaphase promoting complex, and mammalian kinesin-like protein 1 MKLP-1 and other kinesin related motor proteins. These substrates demonstrate the multiple roles of Plk1 in promoting mitosis. During M-phase exit, Plk1 appears to be an important up-regulator of the ubiquitin dependent proteolytic degradation machinery that controls passage through mitosis. The mechanism for this is not clear but Plk1 can directly phosphorylate three subunits of the anaphase-promoting complex (APC), provided that Plk1 has been pre-activated by CDK1/cyclin B. In Xenopus, Plx1 is found localised on centrosomes at prophase consistent with its role in Cdc25C activation and organisation of the bipolar spindle assembly, it is localised on spindles at metaphase consistent with its role in APC activation, and at the mid-body during cytokinesis. While other physiological substrates remain to be identified, the existing roles and functions of Plk1 in cell cycle control make it an exciting target for structural studies.

Plk1 is a 601 amino acid protein that comprises an N-terminal kinase domain and a C-terminal regulatory domain that contains a sequence of termed the Polo box. The motif is only observed in the Polo-like kinases. Its function remains unknown. It may be involved in an auto-regulatory mechanism or in targeting the kinase to substrates. Deletion of the C-terminal regulatory domain results in activation of the kinase. Plks have a threonine in the activation segment whose phosphorylation is essential for activity.

Following the gift of bacterial constructs and viruses encoding the Plk1 gene from Professor Erich Nigg, we have begun expression trials to produce material for crystallisation studies. Although initially we were encouraged by high level expression of the full length Plk1 polo-kinase in a baculoviral infected insect cells, efforts to produce crystals of this protein were hampered by its proteolytic instability. The his-tagged protein could be purified with yields of around 10mg/litre culture using a two-step metal affinity/gel filtration strategy. However, degradation of the protein was evident over the course of one week in both purified material stored at 4/degC and in crystal trials at 20/degC. Various fragments were observed as partially stable intermediates in the course of this proteolysis. These have been identified by N-terminal sequencing. The sequence analysis and size estimates from SDS-PAGE, have allowed truncated protein targets to be deduced for both the kinase and polo-box domains of Plk1. The polo-box target has been used as the focus of design for both bacterial and baculoviral expression systems. Early results involving expression of the polo-box as a GST fusion in bacteria have generated encouraging results. A soluble, purified fragment has been produced, albeit with low yield (~100µg/litre). Work is now underway at producing similar bacterial and baculoviral expression systems for the kinase domain.

1.5 Studies with Ubiquitin Domain Proteins (UDPs)

Nick Brown, Laura Fonzo, and Ed Lowe with Martin Noble and Jane Endicott in collaboration with Hideki Kobayashi (Kyushu University, Japan) and Colin Gordon (Edinburgh, UK)

Ubiquitin is used as a signal molecule for controlled targeting of proteins to the proteasome and in other signalling processes such as DNA repair and splicing. Ubiquitin is attached to the target protein by an isopeptide linkage between the carboxy terminus of ubiquitin and an epsilon-amino group of a lysine side chain in the target protein. The linkage is promoted by a specific series of enzymes. For the regulatory proteins of the cell cycle, the final ubiquitin ligase transfer enzyme is contained in multi-protein complexes the Skp1-Cullin-F box (SCF) complex (for S phase regulatory molecules) and the anaphase promoting complex (APC) (for mitotic regulatory proteins). Recognition for degradation involves poly-ubiquitin chains (at least four ubiquitin subunits) in which each ubiquitin is linked via an isopeptide bond from the carboxy terminus of one ubiquitin to Lys48 on the adjacent ubiquitin. Saccharomyces cerevisiae Dsk2 is a member of a family of proteins that are involved in recognition of ubiquitin and targeting poly-ubiquitin labelled proteins to the proteasome (13). DSK2 was originally isolated as a suppressor of kar1, which is defective in spindle pole duplication (14).
Dsk2 comprises an N-terminal Ubiquitin-Like domain (UBL) (residues 1-77) and a C-terminal Ubiquitin-Associated domain (UBA) (residues 328-373) separated by a region that contains repeat sequences (residues 78-327). Recent evidence has shown that Dsk2, and its homologues in higher eukaryotes, participate in protein degradation by the ubiquitin-proteasome pathway (15, 16). Current models indicate that these proteins may help deliver poly-ubiquitinated substrates to the proteasome through binding of the UBA domain to poly-ubiquitin and binding by the UBL domain to subunit(s) of the 19S proteasomal lid. We wish to uncover the molecular basis of these recognition processes important for controlled protein degradation. As a first step, we have determined the structure of the UBL domain of Dsk2.

Hideki Kobayashi has supplied us with expression constructs encoding full-length budding yeast Dsk2 and 3 fragments: Full-length DSK2: residues 1-373; UBL domain: residues 1-77; Repeat sequence: residues 78-335; UBA domain: residues 328-373. These are in the expression plasmid pGEX-KG that produces GST fusion protein bearing a thrombin cleavage site. Following transformation into B834(DE3)pLysS, expression was induced with 0.1mM IPTG and allowed to proceed for 3-4 hours at 37/degC. The fusion proteins were purified from clarified lysates by glutathione-Sepharose chromatography and the protein of interest released by thrombin cleavage. Superdex 75 chromatography was then used to purify both UBL and UBA fragments with yields of 5 and 10 mg per litre of culture, respectively.

GST-UBA was shown to bind polyubiquitin in a pull down assay followed by immunoblotting with an ant-ubiquitin serum, indicating the protein was correctly folded. N-terminal sequencing revealed that the UBA domain had 14 additional non-native amino acids at its N-terminus that were derived from the linker sequence encoded by the pGEX-KG vector. Mass spectrometry of the UBA protein gave a mass of 6318 Da, consistent with the predicted mass of 6319 Da. Following concentration to 10-12 mg/ml, the purified UBA domain was crystallised at 20/degC using sodium acetate as precipitant. Single crystals were rod shaped and readily grew overnight with dimensions up to 250 x 50 x 50 mm. Following cryoprotection with 20% glycerol in mother liquor, datasets have been collected to 3 Å resolution at ESRF, Grenoble and at SRS, Daresbury. Attempts to obtain phases by molecular replacement using coordinates from an NMR structure of a UBA domain from the human RAD23 protein were unsuccessful so seleno-methionine protein has been produced for SAD experiments. Small crystals of SeMet UBA have been grown and an initial experiment at ESRF 14.2 showed that they diffracted to higher resolution (2.2Å) than the native crystals. This initial dataset however had twinned lattices so further data collection is planned and a structure should soon follow.

Figure 4. ...more

The UBL domain protein has been crystallised using sodium citrate as precipitant. Data to high resolution (1.15Å) were collected at ESRF and the structure solved by molecular replacement using the UBL domain from the plant protein Rub1 (36% identical) as the search molecule. The structure (Figure 4) shows a typical UBL fold, comprising a 5 stranded antiparallel

sheet, a single a helix and a 3₁₀ helix. The fold is similar to other UBL domains and to ubiquitin (rms difference in C

positions 1.4 Å). These proteins show a conserved hydrophobic patch located on the surface that, in some proteins, has been implicated in interactions with a specified proteasomal subunit. Future experiments will centre on the production of a complex of UBL with an active binding fragment of Rpn1, the subunit of the 19S proteasomal particle that has been shown to bind Dsk2.

Purification of the full length Dsk2 and its repeat region have proved more difficult. Although soluble protein expression is good, thrombin digests have lead to non-specific cleavage of the target proteins. We are in the process of subcloning these 2 clones into pGEX6P-1 that utilises the highly specific 3C protease. This should circumvent non-specific protein cleavage.

Colin Gordon has provided us with a number of clones encoding 5 different UDP's from Schizosaccharomyces pombe, namely Rhp23, Dph1, Mud1, Bag1 and Ubp6. Work is underway to clone and express these proteins and their fragments to enable a comparative structural analysis.

2. Regulation of phosphorylase kinase

Atlanta Cook

Phosphorylase kinase plays a central role in coordinating cellular signals in the regulation of glycogen metabolism, from hormone responses via cAPK and through calcium signalling. The kinase converts inactive glycogen phosphorylase b to the active glycogen phosphorylase a through specific phosphorylation of residue Ser14. Activation of glycogen phosphorylase leads to glycogenolyis and the production of ATP to sustain muscle contraction. The kinase is a large hexadecameric holoenzyme made up of four different subunits:

and

. The

chain contains the kinase domain and a C-terminal regulatory region. The

subunit is an intrinsic molecule of calmodulin (CaM) that binds to the regulatory region of the

chain and confers calcium sensitivity on the enzyme. The

and

subunits also regulate the

subunit by incorporating signals from phosphorylation, the binding of extrinsic Ca²⁺/calmodulin and metabolite binding.

Figure 5. ...more

A structure of the holoenzyme to 22Å using negative staining electron microscopy methods has been determined with the enzyme in complex with phosphorylase, giving a low resolution model of holoenzyme structure (17, 18). At higher resolution two structures are available of the kinase domain of the enzyme in the absence and presence of a substrate peptide (19, 20). The kinase domain consists of residues 1-296 of the

chain and is constitutively active suggesting that for kinase activity to be controlled, the kinase domain has to be inhibited by other parts of the structure. The C-terminal region of the

chain (residues 297-386) is known to play a role in regulating the kinase activity, possibly through an autoinhibitory mechanism. The C-terminal domain forms a complex with the

subunit, the intrinsic CaM molecule, and activation by Ca²⁺ can, perhaps, be explained by relief of the autoinhibition through conformational changes produced by the calmodulin. Calcium calmodulin regulated kinases form an important class of kinases responsible for a number of processes that include muscle contraction, neuronal mechanisms and apoptosis, and yet there is no detailed structural model for the way in which Ca²⁺/CaM controls these kinases.
In order to gain further insights in to the regulation of the

subunit by CaM it is necessary to examine the C-terminal region. To address this aspect of kinase regulation various constructs of the

chain were made as N-terminal GST fusion proteins using the pGex6p1 vector. The advantages of this approach are two-fold: not only is GST a useful aid to purification but, when used as an N-terminal fusion, it can also assist the folding of proteins that are relatively unstable.

Expression trials of the full-length

construct, GSTgFL failed to yield soluble protein under a variety of conditions and indicated that the protein had difficulty in folding correctly. This was also observed for non-fusion forms of the full-length gamma chain and is perhaps due to the absence of CaM, which forms intimate associations with the regulatory region. Indeed, the

and

can be isolated as a stable complex from the holoenzyme. To overcome this we have elected to follow a coexpression strategy for with the CaM gene to allow the CaM to rescue the solubility of the

subunit. A biscistronic pGex6p construct containing GST

FL and CaM has been constructed. However both the construction of the vector and expression trials were hampered by several difficulties that led us to believe that the active full-length kinase in combination with calmodulin was toxic to E. coli cells. This phenomenon had also been observed with the kinase domain. We have therefore adopted an alternative strategy in which an inactive kinase, achieved through site directed mutagenesis, has been engineered into the biscistronic construct. We are also investigating a truncated version of the full-length kinase subunit that still retains the calmodulin-binding domain.

Figure 6. ...more

More success in expression trials was met with the GST

trnc and GST

CA constructs, both of which can be expressed as soluble protein. In previous purification protocols for the untagged

trnc the protein has been purified by refolding from inclusion bodies and it has not been possible to obtain reproducible crystals from one preparation to another. A more robust system for obtaining crystals would allow structural studies of the kinase in complex with a variety of different ligands to be undertaken. A purification strategy for the GSTgtrnc protein has been achieved, involving glutathione-Sepharose chromatography followed by cleavage of the fusion and purification on a Cibacron blue matrix. The purified catalytic domain using this protocol is active and crystallises under the same conditions as the refolded protein (Figure 6). This method of purification has the advantage that it is faster and easier than previous preparations and is more reproducible. Crystallographic studies of the truncated gamma subunit in complex with ATP-analogue inhibitors, such as KT5720 a specific inhibitor of phosphorylase kinase (21), are in progress.

3. Calmodulin binding to PhK peptides

In collaboration with J. McDonnell

HSQC spectra of CaM in the absence of PhK5 (red) and at 100% saturation of PhK5 (blue)

Figure 7. ...more

The

subunit of phosphorylase kinase is an intrinsic molecule of CaM that forms a complex with the

subunit. Two peptides of 25 amino acids in length from the C-terminal region were identified as CaM binding peptides through their ability to inhibit the Ca²⁺/calmodulin-dependent activation of myosin light chain kinase (22). The peptides, PhK5 (residues 342-366) and PhK13 (residues 302-326) bind simultaneously to CaM in different binding modes (23, 24). The expression system for calmodulin that utilises the PROTet vector produces significant amounts of calmodulin. Analytical gel filtration experiments indicated that CaM/PhK5 complexes were stable but CaM/PhK13 complexes were not. Attempts to co-crystallise the purified calmodulin with the peptides PhK5 and PhK13 yielded only empty calmodulin crystals. We have turned to NMR in order to characterise the interactions of the peptides with calmodulin. E. coli cells expressing CaM were grown on minimal medium in the presence of ₁₅N labelled NH₄Cl to produce ₁₅N labelled protein. HSQC experiments were then carried out, first comparing spectra in the absence and presence of Ca²⁺ and then titrating in the PhK5 peptide in 10% to 20% increments until the CaM was saturated. A comparison of the chemical shifts for the protein in the absence of peptide and at the 100% saturation point is shown in (Figure 7.) It is clear from these spectra that there are many changes in the structure on binding, which is likely to arise from conformation changes as well as binding interactions in the protein. Analysis of the residues that are involved in the binding is ongoing.

References

1.   Davies, T. G., Pratt, D. J., Endicott, J. A., Johnson, L. N., and Noble, M. E. M. (2002) Structure -based design of cyclin-dependent kinase inhibitors, Pharmacology & Therapeutics 93, 125-133.
2.   Davies, T. G., Bentley, J., Arris, C. E., Boyle, F. T., Curtin, N. J., Endicott, J. A., Gibson, A. E., Golding, B. T., Griffin, R. J., Hardcastle, I. R., Jewsbury, P., Johnson, L. N., Mesguiche, V., Newell, D. R., Noble, M. E. M., Tucker, J. A., Wang, L., and Whitfield, H. J. (2002) Structure-based design of a potent purine-based cyclin dependent kinase inhibitor, Nature Structural Biology 9, 745-749.
3.   Johnson, L. N., De Moliner, E., brown, N. R., Song, H., Barford, D., Endicott, J. A., and Noble, M. E. M. (2002) Structural studies with inhibitors of the cell cycle regulatory kinase cyclin-dependent protein kinase 2, Pharmacology & Therapeutics 93, 1-12.
4.   Russo, A. A., Jeffrey, P. D., Patten, A. K., Massague, J., and Pavletich, N. P. (1996) Crystal structure of the p27Kip1 cyclin-dependent-kinase inhibitor bound to the cyclin A-CDK2 complex, Nature 382, 325-331.
5.   Brown, N. R., Noble, M. E. M., Endicott, J. A., and Johnson, L. N. (1999) The structural basis for specificity of substrate and recruitment peptides for cyclin-dependent kinases, Nature Cell Biology 1, 438-443.
6.   Leng, X., Noble, M. E. M., Adams, P. D., Qin, J., and Harper, J. W. (2002) Reversal of growth supression by p107 via direct phosphorylation by cyclin D1/cyclin dependent kinase 4, Mol. Cell. Biol. 22, 2242-2254.
7.   Schulman, B., Lindstrom, D. L., and Harlow, E. (1998) Substrate recruitment to cyclin-dependent kinase 2 by a multipurpose docking site on cyclin A, Proc. Natl. Acad. Sci. USA 95, 10453-10458.
8.   Takeda, D. Y., Wohlschlegel, J. A., and Dutta, A. (2001) A bipartite substrate recognition motif for cyclin-dependent kinases, J. Biol. Chem. 276, 1993-1997.
9.   Chen, Y.-N. P., Sharma, S. K., Ramsey, T. M., Liang, L., Martin, M. S., Baker, K., Adams, P. D., Bair., K. W., and Kaelin, W. G. (1999) Selective killing of transformed cells by cyclin/cyclin dependent kinase 2 antagonists, Proc. Natl. Acad. Sci. USA 96, 4325-4329.
10.   Lowe, E. D., Tews, I., Cheng, K.-Y., Brown, N. R., Gul, S., Noble, M. E. M., Gamblin, S. J., and N., J. L. (2002) Specificity determinants of recruitment peptides bound to phospho-CDK2/cyclin A, Biochemistry In press.
11.   Garrett, S., Barton, W. A., Knights, R., Jin, P., Morgan, D. O., and Fisher, R. P. (2001) Reciprocal activation by cyclin dependent kinases 2 and 7 is directed by substrate specificity determinants outside the T loop, Mol. Cell. Biol. 21, 88-99.
12.   Lawrie, A. M., Tito, P., Hernandez, H., Brown, N. R., Robinson, C. V., Endicott, J. A., Noble, M. E. M., and Johnson, L. N. (2001) Xenopus phospho-CDK7/cyclin H expressed in baculoviral-infected insect cells, Protein Expression and Purification 23, 252-260.
13.   Jentsch, S., and Pryrowlakis, G. (2000) Trends in Cell Biol. 10, 335-342.
14.   Biggins, S., Ivanoska, I., and Rose, M. D. (1996) J. Cell Biol. 133, 1331-1346.
15.   Funakoshi, M., Sasaki, T., Takeharu, N., and Kobayashi, H. (2002) Budding yeast Dsk2p is a polyubiquitin-binding protein that can interact with the proteasome, Proc. Natl. Acad Sci USA 99, 745-750.
16.   Rao, H., and Sastry, A. (2002) Recognition of specific ubiquitin conjugates is important for the proteolytic functions of the ubiquitin-associated domain proteins Dsk2 and Rad23, J. Biol. Chem. 277, 11691-11695.
17.   Nadeau, O. W., Carlson, G. M., and Gogol, E. P. (2002) A Ca²⁺-dependentglobal conformational change in the 3D structure of phosphorylase kinase obtained from electron microscopy, Structure 10, 23-32.
18.   Venien-Bryan, C., Lowe, E. D., Boisset, N., Traxler, K. W., Johnson, L. N., and Carlson, G. M. (2002) Three-dimensional structure of phosphorylase kinase at 22 Å resolution and its complex with glycogen phosphorylase, Structure 10, 33-41.
19.   Owen, D. J., Noble, M. E., Garman, E. F., Papageorgiou, A. C., and Johnson, L. N. (1995) Two structures of the catalytic domain of phosphorylase kinase: an active protein kinase complexed with substrate analogue and product, Structure 3, 467-482.
20.   Lowe, E. D., Noble, M. E. M., Skamnaki, V. T., Oikonomakos, N. G., Owen, D. J., and Johnson, L. N. (1997) The crystal structure of a phosphorylase kinase peptide substrate complex: kinase substrate recognition, EMBO J. 16, 6646-6658.
21.   Davies, S. P., Reddy, H., Caivano, M., and Cohen, P. (2000) Specificity and mechanism of action of some commonly used protein kinase inhibitors, Biochem. J. 351, 95-105.
22.   Dasgupta, M., Honeycutt, T., and Blumenthal, D. K. (1989) The g subunit of skeletal muscle phosphorylase kinase contains two non-contiguous domains that act in concert to bind calmodulin., J. Biol. Chem. 264, 17156-17163.
23.   Trewella, J., Blumenthal, D. K., Rokop, S. E., and Seeger, P. A. (1990) Small-angle scattering studies show distinct conformations of calmodulin in its complexes with two peptides based on the regulatory domain of the catalytic subunit phosphorylase kinase, Biochemistry 29, 9316-9324.
24.   Juminaga, D., Albaugh, S. A., and Steiner, R. F. (1994) The interaction of calmodulin with regulatory peptides of phosphorylase kinase, J. Biol. Chem. 269, 1660-1667.