Structural Studies on PP2A and Methods in Protein Production

Structural studies on PP2A and Methods in Protein Production

Auður Magnúsdóttir

©Auður Magnúsdóttir, pp.1-53, Stockholm 2008

ISBN 978-91-7155-750-6

Printed in Sweden by Universitetsservice AB, Stockholm 2008 Distributor: Department of Biochemistry and Biophysics, Stock- holm University

All previously published papers are reprinted with permission from the publisher

To Elski Pelski

Abstract

PP2A is a major phosphatase in the cell that participates in multiple cell signaling pathways. It is a heterotrimer of a core dimer and variable regulatory subunits. Details of its structure, function and regulation are slowly emerging. Here, the structure of two regulators of PP2A are described; PTPA and B56γ. PTPA is a highly conserved enzyme that plays a crucial role in PP2A activity but whose biochemical function is still unclear. B56γ is a PP2A regulatory subunit linked to cancer and the structure presented here of B56γ in its free form is particularly valuable in light of the recent structures of the PP2A holoenzyme and core dimer. Protein production is a major bottleneck in structural genomic projects. Here, we describe two novel methods for improved protein production. The first is a colony based screening method where any DNA library can be screened for soluble expression of recombinant proteins in E.coli. The second method involves improvements of the well established IMAC purification method. We have seen that a low molecular weight component of E.coli lysate decreases the binding capacity of IMAC columns and by removing the low molecular weight components, recombinant proteins only present at low levels in E.coli lysate can be purified, which has previously been believed to be unfeasible.

Contents

Introduction ...... 11 Macromolecular Structures...... 11 Scope of the Thesis ...... 12

Protein phosphatase 2A ...... 13 Ser/thr dephosphorylation...... 13 Biochemical properties...... 15 Cellular role of PP2A ...... 19 Structural studies of PTPA ...... 21 Biological role of PTPA...... 21 Structure of PTPA ...... 22 The structure of the PP2A regulatory subunit B56γ ...... 25 Structure of the free B56γ ...... 26 B56γ in its free and holoenzyme forms ...... 28 Holoenzyme assembly...... 28

Methods in Protein Production...... 31 Escherichia coli as an expression host...... 31 Structural genomics and protein production ...... 32 The CoFi blot...... 32 Procedure of the CoFi blot...... 33 Screening of an DNA library with the CoFi blot ...... 34 Limitations to IMAC...... 34 IMAC and the 6xhis tag ...... 34 Migration of a target protein on IMAC columns...... 35 Desalting of lysate before loading on IMAC ...... 35 Who is to blame?...... 38

Summary...... 40

Future perspective ...... 42

References ...... 45

List of Publications

Magnusdottir, A., Stenmark, P., Flodin, S., Nyman, T., Kotenyova, T., Gräslund, S., Ogg, D., and Nordlund, P. (2008) The structure of the PP2A regulatory subunit B56gamma: The remaining piece of the PP2A jigsaw puzzle. Proteins: Structure, Function, and Bioinformatics, In press (available online as of 10th July 2008)

Magnusdottir, A. Johansson, I.,Dalhgren, L-G., Bakali, A., Nordlund, P. and Berglund, H. (2008) IMAC purification has severe and serious limitations. Means for improvement. Manuscript

Magnusdottir, A., Stenmark, P., Flodin, S., Nyman, T., Hammarstrom, M., Ehn, M., Bakali-H, M.A., Berglund, H. and Nordlund, P. (2006) The crystal structure of a human PP2A phosphatase activator reveals a novel fold and highly conserved cleft implicated in protein-protein interactions. Journal of Biological Chemistry, 281: p. 22434-22438

Cornvik, T., Dahlroth, S.L., Magnusdottir, A., Flodin, S., Engvall, B., Lieu, V., Ekberg, M., and Nordlund P. (2006) An efficient and generic strategy for producing soluble human proteins and domains in E. coli by screening construct libraries. Proteins-Structure Function and Bioinformat- ics, 65: p. 266-273

Cornvik, T., Dahlroth, S.L., Magnusdottir, A., Herman, M.D., Knaust, R., Ekberg, M. and Nordlund P. (2005) Colony filtration blot: a new screening method for soluble protein expression in Escherichia coli. Nature Methods, 2: p. 507-9

Additional publications

S. El-Andaloussi, H. Johansson, A. Magnusdottir, P. Jarver, P. Lundberg, and U. Langel. (2005) TP10, a delivery vector for decoy oligonucleotides targeting the Myc protein. Journal of Controlled Release, 110: p.189-201.

A. Magnusdottir, M. Masson, T. Loftsson. (2002) Self Association and Cyclodextrin Solibilization of NSAIDs. Journal of inclusion phenomena and Macrocyclic Chemistry, 44: p. 213-218

T. Loftsson, A. Magnusdottir, M. Masson and J.F. Sigurjonsdottir. (2002) Self-Association and Cyclodextrin Solubilization of Drugs. Journal of Pharmaceutical Sciences 91 (11): p. 2307-2316

H.H. Sigurdsson, A. Magnusdottir, M. Masson, T.Loftsson. (2002) The Effects of Cyclodextrins on Hydrocortisone Permeability Through Semi- Permeable Membranes. Journal of inclusion phenomena and Macrocyclic Chemistry, 44: p. 163-167

Abbreviations

6xhis Hexahistidine fusion tag E. coli Escherichia coli GFP Green Fluorescent Protein IMAC Immobilized Metal Affinity Chromatography IRL Intra repeat loop PP2A Protein Phosphatase 2A PP2Aa Structural subunit of PP2A PP2Ab Regulatory subunit of PP2A PP2Ac Catalytic subunit of PP2A PP2Ad PP2A core dimer of PP2Aa and PP2Ac PTPA PP2A phosphatase activator PPIase Peptidyl prolyl cis/trans isomerase PME-1 PP2A methyl esterase 1 SGP Structural genomics project

Introduction

Macromolecular Structures Information about the three dimensional structure of biological macromolecules is vital for understanding their function at the molecular level. Although the role of structural biology has not been as dominating in drug design as first expected it has become increasingly important in other fields of the life sciences and has now regained ground in drug discovery. The large number of structures now available of different biologically relevant macromolecules; proteins, nucleic acids and viruses, and complexes thereof has brought this about. Structural biology is no longer a narrow field of a few eccentrics with generous amount of time (it took 22 years to determine the first protein structure); rather it is an indispensable tool in the understanding of the molecular mechanisms of life. Despite the difficulties in obtaining protein crystals and other chal- lenges such as missing phase information, twinning and weak diffracting crystals, X-ray crystallography continues to be the dominating method for high-resolution structure determination of macromolecules. On the 50th an- niversary of the first protein structure, X-ray crystallography is still used for the ab initio determination of the majority of macromolecular structures and 85%, of all structures in the PDB are determined with X-ray crystallography (figure 1). Other methods used to obtain 3D structures of macromolecules, NMR and electron microscopy, are limited to lower molecular mass proteins and low resolution, respectively. However, the three methods are to some extent complementary; NMR can provide dynamic and kinetic information and electron microscopy is well suited for the study of large macromolecular assemblies (Branden and Tooze, 1999; Manjasetty et al., 2008; McPherson, 2003).

Figure 1: The number of entries in the protein data bank each year of this millennium in total and by experimental method. In 1990, 508 structures had been submit- ted to the PDB, on August 19th, 2008 this number was 52535.

Scope of the Thesis The thesis describes the structures of two regulators of the protein phosphatase PP2A. These structural studies address the role of these regulators in assembly and activation of PP2A. This thesis also describes two novel methods for producing challenging proteins for structural studies. The first is the Colony filtration (CoFi) blot, which is a colony-based method for screening solubility of DNA libraries of protein variants. The second method addresses some severe limitations to the widely used IMAC purification methods and means of improving it.

12 Protein phosphatase 2A

Ser/thr dephosphorylation Regulation of virtually all cellular processes involves rapid and re- versible phosphorylation - some 30% of intracellular human proteins are believed to be reversibly phosphorylated, and the genes encoding kinases and phosphatases constitute as much as 3% of the genomes of higher eukaryotes. Many human diseases can be attributed to a malfunctioning kinase or phosphatase (for reviews, see (Barford et al., 1998; Johnson and Barford, 1993)). Protein phosphatases 1 and 2A (PP1 and PP2A, respectively) are responsible for the vast majority of all ser/thr dephosphorylations in the cell. PP1 interacts with endogenous inhibitors and more than 50 divergent regulatory subunits. Although the regulatory subunits are structurally unrelated they all form stable complexes with PP1. Most of them contain an anchor- ing motif, the RVXF motif that binds to a grove on PP1’s surface and thereby tethers the regulatory subunit to PP1. However, it is mainly interactions at other contact points between the regulatory subunits and PP1 that determine the catalytic activity and substrate specificity of PP1. Each contact point on PP1 can be utilized by several regulatory subunits and each regulatory subunit usually utilizes more than one contact point (for recent reviews see (Ceulemans and Bollen, 2004; Gallego and Virshup, 2005)). Other eukaryotic ser/thr protein phosphatases are the magnesium dependant protein phosphatases (PPM) e.g. PP2C and the phosphoprotein phosphatases (PPP) PP3 (PP2B), PP5, PP4 and PP6, the two latter being type 2A phosphatases. The PPP and PPM families share no sequence similarity and are only roughly structurally similar, comprized of a central β-sheet sandwich flanked by α-helices on both sides. The relatively few catalytic subunits of ser/thr phosphatases, compared with protein kinases, obtain the required structural diversity by binding to different regulatory subunits that control their substrate specificity, as in the case of PP1 discussed above. The crucial determinant of substrate specificity is therefore not residues of the substrate protein surrounding the Ser/Thr to be dephosphorylated. This is in contrast to the protein kinases where the neighboring residues to the phosphorylation site are most often used for specificity recognition. For phos-

13 phatases the specific recognition is believed to be mediated through interactions of the substrate with the regulatory subunits, and/or sites on the catalytic subunit that have been altered through interaction with the regulatory subunits.

Figure 2: The proposed reaction mechanism of the PPP family phosphatases. This mechanism was put forward for PP1 (Egloff et al., 1995) and the corresponding conserved residues in PP2A are shown in bold The PPPs participate in many diverse regulatory processes in the cell. PP1 plays a role in controlling for example glycogen metabolism, cell cycle progression, RNA splicing and neuronal activities. PP2B is an important member of the Ca2+ signaling cascade activating T-cells and is characterized by its Ca2+ dependence. PP2C reverses stress-activated protein kinase cascades (for reviews, see (Barford et al., 1998; Cohen, 1997; Kennelly, 2001; Mitic et al., 2006)). The type 2A phosphatases, PP6 and PP4 are among other things involved in NF-κB signaling (Stefansson et al., 2008) and centrosome maturation (Cohen et al., 2005), respectively. The catalytic domain of the proteins of the PPP-family is highly conserved. They share 34% sequence similarity and all contain a bimetal center. The identity of the metals is unclear but one is most likely a Fe2+ and the other Zn2+ or Mn2+ (Barford et al., 1998; Kennelly, 2001; Mitic et al., 2006). With the recent structure determination of different PP2A complexes, the structure of the catalytic subunit of all the major PPPs has been determined. Not surprisingly, the structures of the PPP catalytic subunits are very similar. Two mixed β-sheets form a sandwich with three α-helices and a β-sheet on one side and seven α-helices on the other. The catalytic site is at the C-terminus of the two β-sheets forming the sandwich and residues on

14 loops sticking out from the sandwich coordinate the binuclear center (Cho and Xu, 2007; Das et al., 1996; Egloff et al., 1995; Goldberg et al., 1995; Kissinger et al., 1995; Swingle et al., 2004; Xu et al., 2006). The dephosphorylation is thought to be carried out in one step. The role of the bimetal center is to bind and polarize the phosphate moiety to make it more vulner- able to a nucleophilic attack from a metal coordinated water. The metal ions also play a key role in activating the water molecule for a nucleophilic attack (Egloff et al., 1995). The reaction mechanism proposed by David Barford’s group for the PP1/PP2 family is shown in figure 2.

Protein phosphatase 2A

Biochemical properties PP2A is a heterotrimer consisting of a highly conserved catalytic subunit, a regulatory subunit and a structural subunit thought to serve as a scaffold bringing the other two subunits together. The core enzyme is the dimer (PP2Ad) of the 36 kDa catalytic subunit (PP2Ac) and the 65 kDa scaffolding subunit (PP2Aa) that comes in contact with one regulatory subunit (PP2Ab) (Janssens and Goris, 2001). PP2A is an essential enzyme, as mouse embryos with the α isoform of the PP2Ac knocked out die (Götz et al., 1998) and the ablation of the PP2Ac in drosophila cell lines is lethal (Silverstein et al., 2002). It is the major phosphatase in the cell along with PP1 and participates in regulation of multiple cellular pathways such as the cell cycle, apoptosis, development, transcription, translation, metabolism, cytoskeleton stability and it is a key player in many human diseases (for reviews, see (Arroyo and Hahn, 2005; Eichhorn et al., 2008 ; Janssens and Goris, 2001; Janssens et al., 2005; Schonthal, 2001; Sontag, 2001)

The catalytic subunit In the cell, the PP2Ac is present in two isoforms, α and β, expressed from two different genes. They are non-redundant despite 97% sequence identity. The α isoform is highly over-represented (ca. 90%) but both are ubiquitously expressed. The Cα and Cβ subunits do not, however, differ in holoenzyme formation or kinetic activity (Zhou et al., 2003). The expression of PP2Ac is thoroughly controlled (Lizotte et al., 2008; Strack et al., 2004; Turowski et al., 1997).

15 PP2Ac is one of evolutions most highly conserved enzymes and 7 out of the last 9 residues of the C-terminus are invariant (Cohen et al., 1990; Ogris et al., 1997). PP2Ac is covalently modified post-translationally. The invariant C-terminus is carboxymethylated and this has been shown to regulate to some extent the PP2Ab association with the PP2Ad (Tolstykh et al., 2000; Wu et al., 2000). The PP2A methylesterase PME-1 that demethylates PP2Ac, is associated with an inactive PP2A pool that can be reactivated by PTPA with ATP/Mg2+ (Fellner et al., 2003; Longin et al., 2004). Recent biochemical and structural work on the PME-1 indicates that it plays an important role in holoenzyme assembly and regulation of basic catalytic activity (Hombauer H, 2007; Xing et al., 2008). Y307 of PP2Ac can be phosphorylated by p60v-src and P56lck and is thought to autodephosphorylate (Chen et al., 1992). PP2Ac can also be phosphorylated at some still uniden- tified threonines. The effect of the phosphorylation is generally thought to be inhibitory (Damuni et al., 1994) although PP2Ac phosphorylated at Y307 is active in vitro (Ogris et al., 1997).

The structural subunit The PP2Aa subunit, called PR65 or A-subunit, is also present in two non redundant isoforms in the cell, α and β, and the α isoform is in general more abundant (Hemmings et al., 1990). It is composed of 15 tandem repeats, each of on average 39 residues that form two antiparallel helices. The repeats arrange themselves in a slightly tilted fashion giving the protein the overall shape of the letter “C” made of an α-helical double layer. This repeated motif is known as the HEAT motif and it is known in many other proteins, such as Importin β and mTOR kinase (Groves et al., 1999). PP2Aa is a tumor suppressor and mutated forms have been found in lung and colon tumors and in breast cancer (Calin et al., 2000; Wang et al., 1998). The dissociation constant (apparent) of 100 - 85 pM for the core dimer is unusually low (Kamibayashi et al., 1992; Turowski et al., 1997). Whether or not the core dimer exists in vivo remains to be conclusively demonstrated. (Gentry et al., 2005; Kremmer et al., 1997; Silverstein et al., 2002).

The regulatory subunit The PP2Abs control PP2A’s substrate specificity and cellular localization. They are very diverse but can be divided into 3 different structural families: B, B’and B’’. They all appear to have similar binding sites on PP2Ad as the binding of one PP2Ab excludes the binding of all others. The interaction between PP2Ad and PP2Ab is not equally strong for all PP2Abs in vivo (for reviews, see (Eichhorn et al., 2008 ; Janssens and Goris, 2001; Sontag, 2001). The B56 or B’ family is the largest family of PP2Ab, with 5 gene members, some of which have several splice variants. They are expressed at high levels in muscle and brain; the α and γ in skeletal and cardiac

16 muscle and β, δ and ε in brain (McCright and Virshup, 1995) and (McCright et al., 1996a). All members are themselves targets for phosphorylation (McCright et al., 1996b; Xu and Williams, 2000), including the splice variant B56γ1 that is expressed and phosphorylated in a tissue specific manner. Phosphorylation of B56γ1 inhibits binding to the PP2Ad (Letourneux et al., 2006). α, β and ε are cytoplasmic whereas δ and γ1 and γ3 are mainly nuclear (McCright et al., 1996b). The other two families of PP2Abs are B55 or B which are composed of WD40 repeats with >80% sequence similarity (Hu et al., 2000; Strack et al., 2002) and B’’ or PR72 which contains an EF-hand sequence motif. They have 4 and 2 gene members, respectively and the largest PP2Ab, PR130 is one of two splice forms of the PPP2R3A gene in the PR72 family (Stevens et al., 2003). In the new millennium a still increasing number of studies have identified which PP2Ab proteins direct the activity of PP2A towards which substrates (Chen et al., 2004; Silverstein et al., 2002; Strack et al., 2004; Van Kanegan et al., 2005). This has recently been reviewed by Eichhorn et al. (Eichhorn et al., 2008) The B55 regulatory subunits are involved in cytoskeleton stability (Evans and Hemmings, 2000), growth factor signaling (Bα) (Arnold and Sears, 2006; Van Kanegan et al., 2005) and neuronal activities (Bβ) (Dagda et al., 2003) with substrates such as C-myc and Erk (Arnold and Sears, 2006; Dagda et al., 2003). The B56 regulatory subunits are involved in regulation of growth factor signaling (Van Kanegan et al., 2005), cell cycle check points (Margolis et al., 2006), Wnt signaling (Baek and Seeling, 2007; Seeling et al., 1999; Yang et al., 2003), transcription (Chen et al., 2005) and apoptosis (Li et al., 2002; Strack et al., 2004) with substrates such as p300, Cdc25 and Akt. B’’ family members participate in for example cell cycle control (Stevens et al., 2003; Voorhoeve et al., 1999) with substrates such as Cdc6 and the retinoblastoma protein (Magenta et al., 2008; Yan et al., 2000). Striatin and SG2NA are sometimes included as a fourth family of PP2Abs but are not accepted yet as a regulatory subunit family as the relevance of the holoenzyme in vivo is unclear (Benoist et al., 2006; Moreno et al., 2000).

PP2Ac can also form complexes outside the abundant heterotrimeric form with PP2Aa and PP2Ab. These are e.g. two PP2A inhibitors (for reviews (Janssens and Goris, 2001)), the atypical cyclins G1 and G2 (Bennin et al., 2002) and both the simian virus 40 small T-antigen and the polynoma middle T-antigen (for a review, see (Arroyo and Hahn, 2005)). The α4/Tap42 protein can form complexes with PP2Ac that are active and substrate specific (Chung et al., 1999; Prickett and Brautigan, 2006). As discussed below, the PTPA protein also regulates PP2Ac although the mode of regulation has been debated (Chao et al., 2006; Fellner et al., 2003; Hom- bauer H, 2007; Jordens et al., 2006; Leulliot et al., 2006; Zheng and Jiang, 2005)

Holoenzyme structure A quantum leap in PP2A research was made when the structures of the PP2Ad (Xing et al., 2006) and the heterotrimeric holoenzyme (Cho and Xu, 2007; Xu et al., 2006) were solved independently by two groups in 2006 and could for the first time explain the molecular basis for heterotrimer interaction (see figure 4A). Although the resolution was limited, they were able to identify key residues at the heterotrimer interface. The scaffolding subunit PP2Aa is a true scaffold upon which the PP2Ac and PP2Ab rest. All three components interact with each other and the interaction surface is large although the PP2Aa-PP2Ab interaction is relatively weak. The B56γ expose its concave surface at the PP2Ac active site like a platform which could serve for interactions with substrate proteins. The active site on PP2Ac is facing away from the other two subunits and is free for substrate access and the surface available for substrate contact on PP2Ac is highly reduced upon B-subunit binding. The holoenzyme structure conforms to the earlier suggestion that the recruitment of substrate is controlled/aided by PP2Ab interaction by creating new binding sites on the holoenzyme (Cho and Xu, 2007; Janssens and Goris, 2001). B56γ and PP2Ac interact through four motifs; the conserved surface loops on B56γ; the helical domain of PP2Ac; the invariant C-terminal tail of PP2Ac as well as the PP2Ac surface loop containing the conserved R268 (figure 3 and (Cho and Xu, 2007; Xu et al., 2006)).

Figure 3: Two views of the interaction surface PP2Ac (light gray) and B56γ (dark gray) in the PP2A holonezyme complex. The active site metals are shown as black spheres, PP2Aa has been omitted for clarity

The structure of both the heterotrimer and heterodimer can explain, on molecular level, the effect of several mutations of the PP2Aa found in cancer where these compromise interactions with the other subunits (Cho and Xu, 2007; Xing et al., 2006). Furthermore, they could shed some light

18 on the importance of methylation of the invariant C-terminus of PP2Ac for the interaction with PP2Ab. The C-terminal methylation enables binding of the C-terminus at its negatively charged binding site between the PP2Aa and B56γ by neutralizing the negative charge of the C-terminus of PP2Ac. The invariant C-terminus of PP2Ac interacts extensively with both PP2Aa and PP2Ab. A binding pocket could also be identified for the invariant Y307c of the C-terminus. As mentioned above, Y307 can be phosphorylated (Chen et al., 1992) and mutations mimicking this phosphorylation have been shown to abolish the interaction between the PP2Ad and the regulatory B55 and B56 (Ogris et al., 1997) (Gentry et al., 2005) but to increase the abundance of the Tap42(α4)/PP2Ac dimer. Phosphorylation of Y307 would result in a serious clash in the tight binding site for the C-terminal tail of PP2Ac in the holoenzyme (Cho and Xu, 2007; Xing et al., 2006).

Cellular role of PP2A PP2A regulates cell proliferation. It promotes progression through the G1/S phase but inhibits G2/M progression (Schonthal, 2001; Sontag, 2001). PP2A regulates many enzymes in cell cycle control and is among other things involved in formation of the pre-replication complex in G1 (Chou et al., 2002; Yan et al., 2000), spindle checkpoint control (Wang and Burke, 1997) and DNA damage response (Li et al., 2007; Magenta et al., 2008; Petersen et al., 2006). PP2A also regulates cell death and is a pro-apoptotic factor. It activates the pro-apoptotic Bad in collaboration with other phosphatases and inactivates the anti-apoptotic Bcl-2 (for a review, see (Klumpp and Kriegl- stein, 2002). PP2A also activates the transcription factor FOXO that regulates the transcription of many pro-apoptotic and growth inhibiting genes (Yan et al., 2008).

In addition to its involvement in the cell cycle and apoptosis PP2A has furthermore been shown to regulate development, transcription, translation, metabolism, cytoskeleton stability and more (for reviews, see (Eichhorn et al., 2008 ; Janssens and Goris, 2001; Janssens et al., 2005; Sontag, 2001)). When PP2Ac can form many different holoenzymes through binding to different partners, each with its own substrate specificity, PP2A should be con- sidered a group of enzymes rather than a single enzyme. This view subdues the confusion involving PP2A’s multiple roles and can explain the many contradicting activities attributed to it in the past such as, for example; growth inhibition and promotion; inhibition and activation of the MAP kinase cascades (reviews: (Magenta et al., 2008; Schonthal, 2001)). As discussed in the section about the PP2Ab, research is now focused on establish- ing the substrate specificity of each individual type of holoenzyme.

19 PP2A and diseases At least some of the PP2A holoenzymes have a tumor suppressor activity. The inactivation of PP2A is necessary for progression of some cancer types, either through substitutions in PP2Aa that retard or eliminate interactions with the other PP2A subunits, or through the upregulation of oncogenic PP2A inhibitors such as SET and CIP2 ((Calin et al., 2000; Junttila et al., 2007; Ruediger et al., 2001; Tamaki et al., 2004) for reviews, see (Janssens et al., 2005; Perrotti and Neviani, 2008; Sablina and Hahn, 2008)). Simi- larly, many DNA tumor viruses, such as the SV40, polyoma virus and papil- loma virus, target PP2A and its inactivation is necessary for viral transformation (for reviews, see (Arroyo and Hahn, 2005; Goldberg, 1999)). Fur- thermore, dephosphorylation of the notorious oncogenes c-Myc and Akt and the tumor suppressor p53 is mediated by B56α, B55α and B56γ, respectively. This dephosphorylation marks c-Myc for degradation and inactivates Akt but stabilizes p53 (Arnold and Sears, 2006; Kuo et al., 2008; Li et al., 2007). B56γ has been shown to have some additional couplings to human malignancies. Deletion of B56γ can replace the ST antigen in SV40 trans- formations and over expressing B56γ can partly reverse such transformation (Chen et al., 2004; Moreno et al., 2004). A truncated isoform of B56γ has been found upregulated in malignant melanomas (Ito et al., 2003). Tau is hyperphosphorylated and aggregated in the brain tissue plaques of Alzheimer’s disease patients. B55α directs PP2A towards Tau and dephosphorylates it but activity of PP2A is significantly lower in cells affected by Alzheimer’s disease, possibly due to loss of PP2Ac methyltransferase (Sontag et al., 2004; Zhou et al., 2006), but methylation of the PP2Ac C-terminus is absolutely required for formation of B55α containing holoenzyme (Yu et al., 2001). Finally, a mutation in the E3 ligase MID1 that is responsible for PP2Ac degradation, causes X-linked Optiz syndrome, characterized by developmental defects (Trockenbacher et al., 2001), and an extension in B55β causes one form of the neurodegenerative disorder spinocerebellar ataxia (O'Hearn et al., 2001). It has become evident that the PP2A is strictly regulated in a complex and elegant manner. However, the detail and molecular mechanism of this regulation have only just begun to rear its beautiful head. This thesis describes the structure of two PP2A regulators, PTPA and B56γ, which have somewhat furthered our understanding of the fascinating and elusive dephosphorylation system that is PP2A.

20 Structural studies of PTPA

Biological role of PTPA PTPA is a very conserved enzyme that regulates PP2A activity. De- leting it in yeast is lethal (Fellner et al., 2003; Rempola et al., 2000) It was once believed to function as a PP2A phosphotyrosyl activator and then PTPA stood for phosphotyrosine phosphatase activator. However, it be- came apparent a few years ago that in those studies (Cayla et al., 1990; Van- hoof et al., 1994) only a background increase in phosphotyrosyl activity was being measured and the real impact of PTPA on PP2A activity was on the ser/thr specific dephosphorylation. PTPA was therefore renamed protein- phosphatase two A activating protein. PP2Ac is reactivated by PTPA from an inactive conformation in complex with PME-1 (Fellner et al., 2003; Longin et al., 2004). It is critical for PP2A biogenesis in yeast (Fellner et al., 2003; Rempola et al., 2000), possibly as a key player in a gating mechanism where the PP2Ac is in an inactive state in complex with first PME-1, and is then activated by PTPA in complex with PP2Aa, thus avoiding free PP2Ac running amok in the cell (Fellner et al., 2003; Hombauer H, 2007). The surprising discovery of Shi’s group that human PP2Ac does not contain the active site metals when in its inactive state complex with PME-1 (Xing et al., 2008) supports these findings and makes the suggestion of Fellner et al. of PTPA as a PP2A specific metal chaperone of interest again (Fellner et al., 2003).

Fellner et al. are not alone in suggesting possible assignments of PTPAs biochemical function. PTPA has been shown to have peptidyl prolyl cis/trans isomerase (PPIase) activity in vitro, which may act specifically on P190 in PP2Ac (Jordens et al., 2006). Furthermore, the yeast PTPA ho- mologs Rrd1 and Rrd2 interact with the conserved Tap42 protein. Tap42 and PP2Ac from yeast also form a complex, participating in the TOR (target of rapamycin ) signaling pathway (Beck and Hall, 1999; Inoki et al., 2005). Rrd1 and Rrd2 may form a trimer with Tap42 and PP2Ac that is active and has specific substrate specificity (Prickett and Brautigan, 2006; Zheng and Jiang, 2005). PTPA is an essential enzyme critical for the proper function of the ubiquitous and abundant PP2A system (Fellner et al., 2003; Rempola et al., 2000). However, the biochemical activity of PTPA is still under debate and four very different suggestions have been put forth regarding how PTPA functions. It has been suggested to be a ATP dependant PPIase, a metal chaperone, phosphotyrosyl phosphatase activator and a novel PP2A regulatory

21 subunit. Although these suggestions are not mutually exclusive, the elucidation of the role(s) of PTPA in the PP2A system remains an exciting challenge.

Structure of PTPA We have determined the structure of human PTPA, published as an accelerated online publication in Journal of Biological Chemistry in June 2006 (Magnusdottir et al., 2006) and in August the same year two papers on the structure of PTPA were published in Molecular Cell (Chao et al., 2006; Leulliot et al., 2006). We solved the structure with MAD phasing to 1.6Å and refined to R and Rfree of 15% and 18%, respectively. We showed that PTPA has an all α-helical and novel protein fold. A peptide binding site was suggested (peptide shown in grey in figure 4B) and a deep and conserved cleft identified (Cleft 2 in figure 4B) as a likely site for binding interacting molecules. We were not able to find any support for the claims that PTPA binds ATP or that it is a PPIase (Magnusdottir et al., 2006). The two Molecular Cell papers could however show structure based results indicating that the base moiety of ATPγS binds into the deep and conserved cleft (Cleft 2 in figure 4B) (Chao et al., 2006) and that a synthetic peptide (suc-AAPK-pNA) used in PPIase assays, bound to a

Figure 4: A. Space filling model of the PP2A holoenzyme solved by (Cho and Xu, 2007; Xu et al., 2006) B. left: A cartoon representation of PTPA colored from the N- to the C-terminus. Right. PTPA shown as a space filling model colored by con- servation degree (red: completely conserved, blue not conserved). PTPA has a novel fold with two clear clefts; a conserved and deep cleft (cleft 2) likely to be involved in binding other molecules and cleft 1 that is less conserved. A bound peptide is shown as a grey stick model. C. A cartoon representation of two PTPA structures superimposed, one in complex with ATP (PTPA in cyan, ATP in dark blue, pdb code 2HV7) and the other in a dimeric form with a synthetic PPIase substrate suc-AAPK- pNA (substrate in yellow, PTPA in green and magenta, pdb code 2IXP)

23 dimer of PTPA from yeast (Leulliot et al., 2006). The two Molecular Cell papers are however in strong contradiction to each other. There is very little room for two PP2Ac molecules that are to be isomerized, at a PTPA dimer interface. Furthermore, according to Chao et al., a complete ATP binding site is formed at the interface of PTPA and PP2Ac, which is altogether im- possible if PTPA is to be dimerized for PPIase activity, for which ATP is necessary (Chao et al., 2006; Leulliot et al., 2006) (figure 4C). Additionaly, superimposing the two structures on each other shows that ATPγS cannot bind at the binding site identifed by Chao et al. in a PTPA dimer because this binding site is a site for dimer interaction (figure 4C). The two papers further contradict each other regarding the phosphotyrosyl phosphatases activity and ATP hydrolysis. In line with the latest work on PTPA activity Leulliot et al. ignore the phosphotyrosyl phosphatases activity of PP2A altogether while Chao et al. use it for basic biochemical explanation of PTPA activity. Chao et al. detect considerable ATP hydrolysis while Leulliot et al. do not (Chao et al., 2006; Leulliot et al., 2006). It is possible that the ATP hydrolysis, which is a dephosphorylation reaction, is performed by a poorly active and unspecific portion of the PP2Ac population, that can also be responsible for the phosphotyrosyl phosphatases activity (Fellner et al., 2003).

In the paper of Jordens et al. on the PPIase activity of PTPA, the P190 of PP2Ac is thought to be the proline isomerized. They tested peptides corresponding to all prolines in PP2Ac and surrounding residues. P190 was identified through peak broadening in an NMR spectrum. Such peak broadening can also result from molecular interactions. This peak broadening was observed in a dodecamer of L186-L198 from PP2Ac where two prolines, P190 and P194 are present. P190 was thought to be the residue isomerized as peak broadening was still observed for the P194A substituted peptide. The reverse (that is P190A) was not demonstrated (Jordens et al., 2006) therefore it is plausible that if the peak broadening is due only to interaction of the PTPA and the peptide, that P194 is not a key player in that binding. Leul- liot et al. also use similar NMR peak broadening results to show that a peptide from PP2Ac is indeed isomerized but their results could possibly also indicate merely an interaction between a long peptide and PTPA. There is some inconstancy between the NMR and PPIase assay results for the PTPA mutants presented in their paper that may supports this (Jordens et al., 2006; Leulliot et al., 2006). The structure of the PP2A core dimer and holoenzyme was published only a few months after our JBC and the Molecular Cell papers on the PTPA structure. This structure now serves as a basis for interpreting the suggestions for a PPIase activity of PTPA. The suggested substrate proline of PP2Ac, P190 is surface exposed. However, it is not a part of the active site of PP2Ac, but is situated at the base of a hydrophobic pocket that binds the inhibitors okadaic acid or microcystin-LR in the structures of PP2Ac available (Cho and Xu, 2007; Xing et al., 2006; Xu et al., 2006). This

24 pocket is also at the heart of PME-1 binding (Xing et al., 2008). Relevance of PPIase activity of PTPA and the isomerization of P190 in the activation of PP2Ac requires that the hydrophobic pocket can be assigned a direct, yet to be defined, role in catalysis. PTPA reactivates inactive PP2Ac that is in complex with PME-1 (Fellner et al., 2003; Hombauer H, 2007; Longin et al., 2004). The Shi group has now determined the structure of PME-1/PP2Ad complex where P190 of PP2Ac is in its trans conformation as in the structure of the holoenzyme (Cho and Xu, 2007; Xing et al., 2008; Xu et al., 2006), implying that no PPIase activity is required for activation of PP2Ac from the inactive PME-1 complex form, although the trans conformation in the holoenzyme might be induced by the inhibitors. Intriguingly, the deletion of PME-1 in yeast has no detectable phenotype (Wu et al., 2000) indicating a non-crucial role of this enzyme while knocking out the PME-1 gene in mice is lethal (Ortega- Gutiérrez et al., 2008). Combined these data underline the need for further investigations before accepting this activity and opens for the possibility that the PPIase activity of PTPA could be a biochemical in vitro artifact, just as the previously proposed phosphotyrosyl activity.

The construct we used in our study of PTPA was tagged with a purification tag that made an intermolecular interaction with PTPA. We suggested based on the structure and sequence comparisons that the surface where the peptide binds is indeed a hot spot for binding and could be where PTPA interacts with the catalytic subunits. This proposal is consistent with mutagenesis studies made by Chao et al. (Chao et al., 2006; Leulliot et al., 2006). It can however, not be excluded that the tag and the linker region has masked interactions in our biochemical and soaking studies, which could explain why we don’t see any evidence of ATP or PPIase substrate binding.

The structure of the PP2A regulatory subunit B56γ

B56γ is a nuclear PP2Ab participating in transcription control. It has been implicated in cancer through several mechanisms (for a recent review, see (Arnold and Sears, 2008)) and it is the only regulatory subunit whose structure has been solved in complex with PP2Ad. We have solved the structure of the free B56γ , initially as a part of a bigger effort in solving the structures of individual PP2A regulators as well as complexes with the core dimer.

25 After the structure of the holoenzyme was published, the structure of the regulatory subunit has turned out to be very informative in understanding the assembly of the heterotrimeric PP2A complex.

Structure of the free B56γ A truncated variant of the B56γ protein was crystallized using MPD as pre- cipitant. The crystal had a rather high solvent content (57.6%) and the protein molecules are quite flexible in the crystal, as indicated by high B-values. This was apparent already during the synchrotron data collection when the diffraction pattern had attributes of strong diffuse scattering characterize of large thermal motions in the crystal (Wall et al., 1997). The structure of the B56γ regulatory subunit was solved using MAD phasing, and refined to an R and Rfree of 21% and 26%, respectively at 2.6Å resolution. The protein has the general elongated bent shape of a boomerang and is comprised of 8 HEAT like repeats (7 are present in this structure) (figure 5A). HEAT repeat proteins are members of the ARM-superfamily that contain ca. 40 residue long repeats of two helices connected by a short loop, although long intra repeat loops can also be found. Many of the ARM-superfamily proteins are so called scaffolding proteins connecting two or more interacting proteins. The HEAT repeats pack against each other to form an elongated structure bent to a varying extent, with a hydrophobic core. They are then capped by special N- and C-terminal repeats to protect the hydrophobic core (Andrade et al., 2001; Conti and Izaurralde, 2001; Main et al., 2005). Interestingly, the hydrophobic core of the HEAT repeat proteins is loose/flexible and upon binding of interaction partners, the core can realign, drastically altering the relative orientation of the HEAT repeats to each other such as, for example, when Importinβ binds Ran (Conti and Izaurralde, 2001) or when PP2Aa binds PP2Ac and PP2Ad binds B56γ (Cho and Xu, 2007; Groves et al., 1999; Xing et al., 2006). The construct used in our study is truncated in the last HEAT like repeat (E380 is the last residue). In retrospect, truncating at this point was probably not optimal, as this truncation proved to be in the middle of a helix of the 8th repeat (Cho and Xu, 2007) exposing the hydrophobic core. However, it has been essential to generate good crystals since it artificially generates a dimer, and a tail-to-tail interaction between two crystallographic neighbors is seen in the crystal lattice. P371, M370 and L367 from both molecules form this dimer inducing hydrophobic core. When superimposing our structure of free B56γ to the one of B56γ from the PP2A heterotrimer, which contains an

26 intact last repeat, it appears that the last helix from the

Figure 5: The structure of the free B56γ colored by HEAT repeats. B. Superimposi- tion of the free B56γ and B56γ from the holoenzyme (pdb code 2IAE). C. The binding pocket for the invariant C-terminal tail of the PP2Ac is tighter in the holoenzyme form. Shown here is a zoom in of the binding pocket between IRL5 and IRL6. The B56γ free form in magenta has been superimposed on the B56γ in the holoenzyme complex and is shown to the left together with the PP2Ac from the holonezyme. B56γ in the holoenzyme complex in green to the right together with the PP2Ac from the holonezyme. The PP2Ac is shown as a blue stick model. PP2Aa has been omitted for clarity. The PP2Ac C-terminal tail and its binding pocket have been modeled in different ways in the two holoenzyme structures available (pdb code 2npp shown here, 2iae not shown) but this tighter confomation of the C terminal tail binding pocket symmetry related molecule substitutes the missing concave side helix in our constructs.

27 The loops connecting the helices of HEATs 1 and 2 in B56γ are unusually long for a HEAT repeat protein and the density for them is poor; parts of IRL2 are missing completely. All loops have somewhat poor side chain density but good backbone density, except IRL2.

B56γ in its free and holoenzyme forms The structure of B56γ in the holoenzyme complex was solved independently by two groups in 2006 (Cho and Xu, 2007; Xu et al., 2006) while we were in the model building process of our free form. The structure constitute the first, and so far the only PP2Abs free form, and makes it possible to discern some of the conformational changes that take place in a PP2A regulatory subunit upon holoenzyme assembly (figure 5B). Comparisons show that the B56γ from the holoenzyme has a slightly more extended conformation and the overall rot mean square deviation (rmsd) is 1.47 Å over 287 backbone Cα. Locally, 3 intra repeat loops, IRL2, 4 and 6, show significant conformational differences of the backbone and some side chain rearrangements are also observed. Interestingly, the conformational changes in backbone atoms are mainly at sites interacting with the PP2Ac in the holoenzyme, while the residues interacting with the PP2Aa show mainly conformational changes of the side chains. IRL2, which is disordered in the free B56γ, is ordered in the holoenzyme structure where a conserved acidic cluster makes a salt bridge with a conserved arginine of PP2Ac close to the active site, possibly changing catalytic activity or substrate access. IRL4 is involved in intramolecular noncovalent interactions as well as intermolecular interactions with the PP2Aa and PP2Ac. It can be noted that parts of IRL4 have been modeled in different ways in the two holoenzyme structures. Therefore, although intriguing, the significance of the conformational change of this loop is hard to decipher at this point. IRL6 however, forms a part of the binding pocket for the invariant C- terminal tail of PP2Ac. This tail is subjected to methylation and phosphory- lations that regulate the interaction with the PP2Ab (Chen et al., 1992; Da- muni et al., 1994; Ogris et al., 1997; Tolstykh et al., 2000; Wu et al., 2000). IRL6 is not close to or involved in any major crystal contacts in the free B56γ, and the conformational differences at this loop between the different forms are both relevant and interesting.

Holoenzyme assembly Analysis of the conformational changes in B56γ along with examination of the PP2A literature indicated a certain order of events in the holoenzyme

28 assembly process (figure 6). The complex is formed when the PP2Ad binds a PP2Ab. A plausible scenario is that B56γ first interacts with PP2Aa as the conformational changes at the interaction sites with PP2Aa are small indicating a lower entropic component and a lower energy for binding. Further- more, it has been shown that the PP2Aa and B56γ can form a complex without PP2Ac (Li and Virshup, 2002). We suggest that the long and flexible IRL2 is next to establish its binding on PP2Ac and that subsequently other interactions between B56γ and the helical domain of PP2Ac take place. This induces the conformational changes, such as for example the 6Å change in IRL6, necessary for the correct conformation of the binding pocket of PP2Ac C-terminal tail. As a last step of holoenzyme assembly the C- terminal tail of PP2Ac can settle in its binding pocket (figure 5C). Accord- ing to this scenario, the C-terminal tail of PP2Ac

Figure 6: Proposed mechanism for PP2A holoenzyme assembly does not initiate holoenzyme assembly, which might seem somewhat con- troversial in light of the crucial role it plays in PP2A regulation (Chen et al., 1992; Damuni et al., 1994; Ogris et al., 1997; Tolstykh et al., 2000; Wu et al., 2000). Xu et al. have solved the structure of the PP2A holoenzyme with a truncated PP2Ac missing this tail. They show that the holoenzyme structures with and without the C-terminal tail of PP2Ac are virtually identical (at 3.2 Å resolution), implying that it is not necessary for holoenzyme assembly. We instead suggest that the C-terminal tail works as a barrel bolt lock for the complex, ensuring a stable holoenzyme where the posttranslational modifications modulate this function depending on which regulatory subunit is present. Alternatively, the posttranslational modifications of the C-terminal tail of PP2Ac function primarily for other PP2A regulators, where it can serve as a flag directing the PP2Ac into distinct holoenzyme complexes. These suggestions are not mutually exclusive.

30 Methods in Protein Production

Escherichia coli as an expression host Whether the structure of a protein is determined by NMR or X-ray crystallography, large quantities of homogeneous protein are required. Purification directly from the host is seldom feasible due to high cost, very low yields, inefficient purification steps and heterogeneity, and for human proteins availability is a major challenge. The protein of interest is therefore usually produced from an recombinant expression system. E. coli is by far the most common host used for heterologous expression of recombinant proteins. It has several important advantages over other available hosts, such as low cost, rapid growth, simplicity and easy handling. Furthermore, it has been extensively studied and its physiological properties are therefore well known, and many different expression vectors and strains are available (Baneyx and Mujacic, 2004). However, many eukaryotic proteins cannot be expressed in a soluble form in E.coli, but aggregate in an un- or partly-folded state and can only be isolated from inclusion bodies (Gasser et al., 2008). Although refolding from inclusion bodies has resulted in a soluble and active biomolecule in some cases, it has several disadvantages that limit its usefulness. In a high throughput context, the biggest one is the nongeneric nature, i.e. optimization for each protein is needed, low yields and relatively high cost regarding both time and money (Sorensen and Mortensen, 2005). Expression in other systems such as cell free transcription and translation and in higher eukaryotes such as baculovirus expression in insect cells or mammalian cells, have the disadvantage of high cost and being very technically challenging. The latter are also time consuming and recovery yield is low. Expression in yeast cells is another approach which for most protein only shows low yields of expression, However, the problems of missing necessary binding or folding partners or post translational modifications that often lead to aggregation in bacterial expression hosts such as E.coli, are often circumvented in the eukaryotic systems (Endo and Sawasaki, 2006; Forstner et al., 2007).

31 Structural genomics and protein production Structural Genomics Projects (SGP) have as their ambitious aim to map the 3D biomolecular structural space. Although initial progress was somewhat slow and disappointing, SGPs now contribute to half of the novel structures deposited in the PDB (Levitt, 2007). SGPs are contingent upon successful expression and purification of large quantities of hundreds of different proteins. In protein production of such caliber the advantages of E.coli as an expression host become particularly important and it is still the dominant expression host tested (Consortium, 2008), although cell free systems are also utilized. Expression in insect or mammalian cells is not used unless other expression hosts have failed or the target protein is expected to be particularly cumbersome to express (Braun and LaBaer, 2003; Manjasetty et al., 2008). Statistics from several SGPs show that less than 25% at best of proteins from higher eukaryotes can be solubly expressed with current methods (http://targetdb.pdb.org/statistics/TargetStatistics.html as of July 10 2008). Success of expression of soluble eukaryotic proteins in E.coli is somewhat lower even when combined with minor rescue operations such as construct design, solubilization tagging, introduction of extra chaperones or codons to the host, etc . Again, at the very best less than 30% of purified proteins can then be crystallized (Dale et al., 2003; Manjasetty et al., 2008) and http://targetdb.pdb.org/statistics/TargetStatistics.html as of July 10 2008). 3D structures from SGPs are available for only 5% of successfully cloned targets. Structural genomics is slowly entering the post low hanging fruit era, where more recalcitrant proteins need to be produced and crystallized. Therefore it is of great importance for structural genomics to be able to increase the throughput of soluble protein expression, keeping the cost and scale manageable.

The CoFi blot

We have developed a novel colony based screening method, the Colony Filtration (CoFi) blot, for identification of clones expressing soluble protein from larger clone libraries. It can, for example, be used to screen any in vitro evolution library, or a library of pooled wild type proteins of interest, e.g. of an organism or of a functional pathway. An important advantage of the CoFi blot as compared to other colony based screening methods is that it relies only on a small tag, most often the 6xhis purification tag for detection of soluble expression. Therefore no recloning is required before scale up and purification of the protein. Fur- thermore, a small tag does not artificially, solubilize the protein as is the case for other methods using larger protein tags. Another main advantage compared to e.g. the GFP fusion method used by many other laboratories, is that

32 the CoFi blot also reports solubility of the target while the GFP method is proposed to only report the folding state of the target protein (Cabantous and Waldo, 2006; Savva et al., 2007). However, novel implementations of the GFP method as a split variant is a significant improvement when it also reports solubility of the target in the E.coli cell. The CoFi blot has the advantage of also reporting solubility after the cells have been lysed and cell debris has been filtered away, which means that it partially simulates the scale up process of a protein.

Procedure of the CoFi blot

Figure 7: The CoFi blot method. Any DNA library in an expression vector can be screened for soluble expression of target protein For screening with the CoFi blot, a DNA library is transformed into an expression strain and plated on an agar plate with antibiotics. The follow- ing day the colonies are lifted with a hydrophobic Durapore membrane and transferred to a new agar plate containing antibiotics and IPTG for induction. The colonies are induced on the Durapore membrane for a defined induction time. In the next step it is placed on top of a nitrocellulose membrane that has been laid on top of a Whatman paper soaked in lysis buffer. After 30 minutes, 3 cycles of freeze thawing are carried out to complete lysis. Solu- ble proteins will diffuse through the Durapore membrane and adhere to the nitrocellulose. The nitrocellulose membrane is blotted by immunochemical

33 methods against the 6xhis tag to identify colonies expressing soluble proteins. The CoFi blot is then compared to the original plate that was left to regrow over night. Positive colonies are identified and picked for further verification of soluble expression in liquid cultures (see figure 7). Inclusion bodies and soluble proteins are usually separated by centrifugation at high speed (>30,000g) but in the CoFi blot this is done by filtration. The filtration step of the CoFi blot was compared to centrifugation in a test set of 87 constructs and 73 showed good correlation. It is possible that the solubility of some of the constructs that did not correlate was below the detection limit of the CoFi blot and therefore could not be predicted as soluble. The CoFi blot was also shown to be reproducible (Cornvik et al., 2005).

Screening of an DNA library with the CoFi blot A test set of 19 proteins that could not be expressed as soluble full length proteins in E.coli, were subjected to in vitro evolution and a library of random N-terminal truncations was created. The library for each target was then screened with the CoFi blot and colonies expressing soluble protein could be detected for all but 2. 24 positive colonies were picked from each CoFi blot and used in small scale purification and dot blot analysis. Those clones that were confirmed to produce purifiable constructs from liquid cultures were analyzed on SDS-PAGE. A clear band could be observed for 14 out of the 17 proteins on the gels. The size of the purifiable constructs was determined by DNA sequencing. The translational start of the soluble constructs correlated surprisingly well with domain borders predicted for the proteins by two bioinformatics programs (Cornvik et al., 2006).

Limitations to IMAC

IMAC and the 6xhis tag IMAC is a very powerful and widely used method for the purification of recombinantly expressed proteins. A particularly common variant is the purification against a polyhistidine fusion tag that binds to a divalent cation, usually nickel or cobalt, immobilized on a resin. This is the single most used purification method in SGP due to its many important advantages; it is ro- bust and generic, the purification materials used are cheap and tolerate with both native and denaturing conditions and the small size of the tag makes it

34 compatible with most biochemical studies of the recombinant protein (Bornhorst and Falke, 2000; Braun and LaBaer, 2003; Consortium, 2008; Gaberc-Porekar and Menart, 2001; Nilsson et al., 1997; Waugh, 2005). A common setup in a SGP laboratory is a recombinant expression of the target protein fused to 6xhis in E.coli followed by IMAC purification and one more purification step, such as size exclusion chromatography. This usually yields a target protein pure enough (>95%) to start crystallization trials (Consortium, 2008). However, when the target protein cannot be recombinantly expressed over the levels of the host proteins, this setup does not yield pure target protein. Intriguingly, increasing the number of cells, that is the lysate volume, loaded on the columns does not increas yield of soluble target protein. This observation has been explained by the presence host proteins that, by low affinity binding, are able to compeat out the target protein (Bolanos-Garcia and Davies, 2006; Hengen, 1995). We studied the purification of 6xhis tagged proteins on a nickel resin and have found that this is not the case.

Migration of a target protein on IMAC columns Somewhat surprisingly, we saw that E.coli lysate without recombinant overexpression causes migration of an already bound 6xhis tagged target protein, here 6xhis tagged GFP. Intriguingly, this corresponds with leakage of metal ions from the column, where the lysate apparently not only washes 6xhis tagged proteins of the column but also detaches metal ions from it which can then be found in the flow through. Previously, metal ion leakage has been reported from IMAC resins and different resins have different susceptibility towards metal ion leakage (Schafer et al., 2000). However, in practice, resins that have been reported to be resistant to metal ion leakage also display a significant amount of metal in the samples eluted from the columns (Oswald et al., 1997). This is consistent with our observations that 6xhis tagged GFP also migrates on columns claimed to be highly resistant to metal ion leakage. These results indicated to us that an unknown component of E.coli lysate extracts or expels metal ions from IMAC columns resulting in a lower binding capacity of the column. In an effort to better understand and characterize our observations, 6xhis tagged GFP was used as a test protein throughout the study.

Desalting of lysate before loading on IMAC A “common knowledge” regarding IMAC purifications is that a 6xhis target protein present at low levels in E.coli lysate cannot be separated from the proteins of the host due to the competition with the latter

35 for the binding sites on the columns. To test this “common knowledge”, we diluted lysate of E.coli recombinatly overexpressing 6xhis tagged GFP with lysate of E.coli without recombinant overexpression to simulate the conditions of a protein present at low levels in lysate. We separated the high and low molecular weight components of this simulated low level protein by desalting and analyzed the effect the two fractions (figure 8) had on binding of a simulated low level protein in lysate on an IMAC column as compared to the same lysate containing both low and high molecular weight components. We have found that competition with other proteins is not a problem in purification of a 6xhis tagged target protein present at low levels in lysate.

Figure 8: Desalting chromatogram. Shown in a whole black line is the absorbance at 280nm and in a dotted line is the conductivity. Fractions to the left of the vertical line are the protein fractions; the fractions to the right are the low molecular weight components. On the contrary, our results clearly show that if the proteins are separated from low molecular weight components of lysate before loading on an IMAC column, the column does not loose its binding capacity as it does in the case of full lysate. Accordingly, for very large lysate volumes of a simulated 6xhis tagged target protein present at low levels in lysate an IMAC column is able to bind an order of magnitude more of the target protein when using the protein fraction of desalted lysate as compared with the full lysate (figure 9 and 10). The loss of binding capacity caused by lysate cannot be completely reproduced by the low molecular weight components. In figure 9 the yield of 6xhis tagged GFP, a simulated 6xhis tagged target protein present at low levels in lysate, after different treatments of the same lysate stock

36 can be seen. The loss in binding capacity corresponds to the nickel ion leakage from the columns during lysate loading (figures 9 and 11). Columns treated with full lysate or the low molecular weight components of lysate loose approximately half of their nickel upon lysate loading while the nickel leakage from a IMAC column loaded with the high molecular weight components of lysate looses only slightly more of its nickel than an IMAC column loaded with buffer (figure 11).

Figure 9: The yield of GFP protein after loading with different lysate variants on IMAC columns measured using the absorbance at 280 nm or fluorescence. Lysate is the GFP lysate diluted with empty lysate loaded directly, low Mwt. is the GFP lysate diluted with the low molecular weight components from desalting, proteins is the protein fraction from desalting of GFP lysate diluted with empty lysate, buffer is GFP lysate diluted with buffer and 5 mL is the GFP lysate diluted with empty lysate loaded on 5mL IMACcolumn. Buffer all and 5 mL all is the amount of protein when a larger elution volume was pooled and measured (10 mL instead of 6 mL)

Figure 10: SDS-PAGE gel on the pooled elution fractions shown in figure 9. See figure 9 for capitation.

Who is to blame? We see a strong concurrence in the pattern of loss in binding capacity and nickel ion leakage of IMAC columns that has led us to believe that the loss in binding capacity is a result of loss of the immobilized nickel of the columns. We also see that the low molecular weight components of E.coli lysate cause both the nickel ion leakage and loss in binding capacity. The low molecular weight components either extract or expel nickel from the columns. We did preliminary studies on both possible extractors and expellers but have not been able to convince ourselves of either mechanism. We find it likely though that chelators in E.coli lysate extract the nickel from the IMAC column upon loading. No matter what the cause, it is clear that IMAC purifications are not, in the form used in most SGP labs today, nearly as generic as previously believed. It is clear that a study where the low molecular weight component that causes loss in binding capacity and nickel leakage is identified and eliminated, is very urgent and has the potential to alter IMAC purifications so that it can be used for purifying proteins present at low levels in lysate routinely.

Figure 11: Analysis of the nickel content of different fractions after loading on IMAC column. Full lysate, the low molecular weight components of lysate (low mwt. comp), lysate without the low molecular weight components (proteins only) and no lysate (buffer). Flow through is the amount of nickel detected in flow through under loading, metal elution is the amount of nickel left on the column after loading and the total Ni calc. is the sum of the first two values.

39 Summary

Two structures of PP2A regulators have been solved; those of B56γ and PTPA. PTPA is an essential protein important for the proper function of the ubiquitous and abundant PP2A system (Fellner et al., 2003; Hombauer H, 2007; Rempola et al., 2000). However, the MO of PTPA is still unclear. Whether PTPA is a PPIase, a phosphotyrosyl phosphatase activator, a metal chaperone, a novel regulatory subunit or something else, it is clear that there are contradictory data and interpretations in previous studies which needs to be resolved before the assignment of a correct biochemical activity for PTPA. The structure of a free PP2Ab, B56γ, has allowed us to propose a mechanism for holoenzyme assembly. We believe that the C-terminal tail of PP2Ac is not a critical initiator of the assembly; rather that it is a last step of the process. A highly flexible loop on the B56γ that is conserved in the B56 family, is a likely candidate for reining in the heterotrimeric complex by interaction of its acidic cluster with an arginine on PP2Ac. The scenario proposed in figure 6 needs be verified. Two studies addressing challenging protein production problems have been described; The CoFi blot is a novel method developed in our group that allows for screening for expression of soluble recombinant proteins on colony level from any DNA library. The CoFi blot is a powerful method to identify clones overexpressing recombinant target proteins. It is suitable for SGPs as the clones identified can be used for scale up directly after verification, that is, no recloning is required. Furthermore, the 6xhis purification tag is utilized for identification of soluble overexpression of recombinant proteins, which is the most widely used purification tag in SGPs (Consortium, 2008). We have examined the well established IMAC purifications and detected and characterized severe limitations of it. We have also found ways to circumvent these limitations. Our findings have the potential to greatly improve the yield of target protein that can not be over expressed in high amounts, but even to increase the yield of well behaving proteins that are easily over expressed and can now be purified in tens of mgs from IMAC column instead of only a few mgs. We believe that this effect cannot be the result of bad effects on the target protein or the polyhistidine tag as total protein binding is decreased accordingly (figure 9). When an IMAC column

40 is exposed to lysate it loses the binding affinity it has for 6xhis tagged protein through an unknown mechanism involving low molecular weight components of E.coli lysate, possibly metal chelators, but not other proteins. We have seen this effect with different types of buffers, experimental setups and researchers. We urge protein producers to consider the consequences of loading E.coli lysate on an IMAC column and to take precautions so as not to waste valuable time and money on purifications yielding no, or very little, target protein.

41 Future perspective

The future is bright for PP2A structural research. After more than a dec- ade of frustrating unsuccessful attempts at determining the structures of the PP2A holoenzyme and its subunits, in one year, the structures of the the PP2A activator PTPA and the PP2A holoenzyme and core dimer were determined. Structures of the core dimer in complex with the simian virus 40 small T antigen (Chen et al., 2007; Cho et al., 2007) and with the PME-1, the methyl esterase (Xing et al., 2008) soon followed as did the structure of the free B56γ PP2Ab. However, a myriad of exiting and hopefully informative research is on the horizon. The assignment of the correct biochemical action of PTPA as well as a structure of the complex of PTPA with the core dimer is acute. The structural determination of a PP2A holoenzyme with a member of one of the other PP2Ab families will shed further light on the molecular mechanisms behind holoenzyme assembly and recognition of the PP2Abs. An especially exiting project is the elucidation of substrate recognition at the molecular level. Very little is know about substrate recognition; about how and when substrate is recognized, and about the effect of substrate binding to the PP2A holoenzyme. Studies on this subject will be of paramount importance for PP2A research.

Things are also looking a little bit brighter for protein production. The Cofi blot has the potential to help identify variants of proteins of interest that can be recombinantly expressed in a soluble form. However, it is not yet widely used. Furthermore, we have identified serious limitations to the popular IMAC purification method that is likely to make purification of proteins present at medium low levels in E.coli lysate, much more efficient. Hope- fully, it will also lead to further optimization of the IMAC method so that it will be generally applicable to purification of recombinant proteins from whole cell lysate, even if the recombinant protein is present at low levels.

42 Ett stort tack till

Pär för din enthusiasm och ovanliga men intuitiva syn på cellen och forsk- ningen, för all din hjälp och stöd och den frihet du ger oss doktorander. He- lena för dit mod, beslutsamhet och dina kunskaper. Said för att ha organise- rat och strukturerat upp labbet så att det fungerar, för din tålamod och varma utstrålning. Sue-Li och Tobias för att förstå och översätta för andra min dåli- ga svenska, eran underbara bröllopsfest, vår resa i Island och allt det andra. Pål: Kúkum á Kerfið! För alla fester, samtal, sällskap och först och främst för att ha lärt mig krystallografi. Anna-Karin: allt blir sjukt roligt med dig, Ulrika: Jagur harur inteur jobbatur påur AC04ur änur, tack för allt kul vi haft. Lola: for your susceptiveness and humor, Karin: för din sällskap och för hjälp och råd med linux och krystallografi, Susanne Flodin: för alla frukt- samma samarbeten och djupa samtal, Mina kompanjoner i tjejmafian: your support has been priceless! 3D: Daniel MM (life is agony! tack för bl.a. trev- liga synchrotron resor och Ekeberg onykterhet), Daniel Gurmu (för tidiga morgonfika) och Damien (för datorhjälp) 3C: Christian, Christine and Ced- ric: for good times in Tällberg 2008. Ruairi: for reading my thesis and being geðveikt funny. Monika för all hjälp i början av mitt project, Marina for help with molecular biology, Pelle för goda råd och goda samtal. Anna, Maria, Hanna, Hennrik, Marie, Elisabeth, Mikaela och Malin and to people who have left the group: Heidi, Albert, Amin, Anders, Andreas, Ken, Beni- ta, Therese, Jessica, Victoria, Florian, Karl-Magnus, Agnes och Martin A, Hä, Hö och M.

Ett superstort tack till alla på SGC för er hjälp och vänlighet, speciellt till Susanne vd B, Susanne G, Ida, Aja, Martin W, Martin H, Derek, Johanna, Tomas, Martina och Andreas

Arna, Karvel, Sigrún, Snævar, Hrönn, Georg og krakkarnir fyrir öll afmælin, jólaglöggin, skíðaferðina og aðrar ánægjulegar stundir.

Elín Esther: Hugrekki er dyggð sem þú átt nóg af. Hlakka til að kynnast þér betur. Mummi, tónlistarsykurpabbi: Það væri löngu búið að leggja mig inn án sam- talanna okkar.

43 Ósk og Magga: Þið haldið mér á jörðinni. Mamma og Pabbi: Ómetanlegar fyrirmyndir fyrir að breyta lífi ykkar til hins betra og þora að takast á við ógnandi verkefni eins og doktorsnám í útlön- dum og virkjanagerð. Heiðrún og Þorvarður: fyrir að taka foreldra mína að ykkur, hrista upp í þeim og gera þau ánægð. Ömmi og Kata bestu litlusystkini sem ég gæti óskað mér, ég er svo stolt af ykkur. Emelía, Anna Eir og Bumbubúinn, fyrir allar broshrukkurnar sem þið hafið valdið, endalausan stuðning og fyrir að þola mig í gegnum vott og þurrt.

And to you who I forgot.

44 References

Andrade, M.A., Petosa, C., O'Donoghue, S.I., Muller, C.W. and Bork, P. (2001) Comparison of ARM and HEAT protein repeats. Journal of Molecular Biology, 309, 1- 18. Arnold, H. and Sears, R. (2008) A tumor suppressor role for PP2A-B56γ through negative regulation of c-Myc and other key oncoproteins. Cancer and Metastasis Reviews, 27, 147-158. Arnold, H.K. and Sears, R.C. (2006) Protein Phosphatase 2A Regulatory Subunit B56{alpha} Associates with c-Myc and Negatively Regulates c-Myc Accumulation. Molecular and Cellular Biology, 26, 2832-2844. Arroyo, J.D. and Hahn, W.C. (2005) Involvement of PP2A in viral and cellular transformation. Oncogene, 24, 7746-7755. Baek, S. and Seeling, J.M. (2007) Identification of a novel conserved mixed-isoform B56 regulatory subunit and spatiotemporal regulation of protein phosphatase 2A during Xenopus laevis development. BMC Developmental Biology, 7, 1-15. Baneyx, F. and Mujacic, M. (2004) Recombinant protein folding and misfolding in Es- cherichia coli. Nature Biotechnology, 22, 1399-1408. Barford, D., Das, A.K. and Egloff, M.P. (1998) The structure and mechanism of protein phosphatases: Insights into catalysis and regulation. Annual Review of Biophys- ics and Biomolecular Structure, 27, 133-164. Beck, T. and Hall, M.N. (1999) The TOR signalling pathway controls nuclear localization of nutrient-regulated transcription factors. Nature, 402, 689-692. Bennin, D.A., Don, A.S.A., Brake, T., McKenzie, J.L., Rosenbaum, H., Ortiz, L., DePaoli- Roach, A.A. and Horne, M.C. (2002) Cyclin G2 associates with protein phosphatase 2A catalytic and regulatory B ' subunits in active complexes and induces nuclear aberrations and a G(1)/S phase cell cycle arrest. Jour- nal of Biological Chemistry, 277, 27449-27467. Benoist, M., Gaillard, S. and Castets, F. (2006) The striatin family: A new signaling platform in dendritic spines. Journal of Physiology, 99, 146-153. Bolanos-Garcia, V.M. and Davies, O.R. (2006) Structural analysis and classification of native proteins from E. coli commonly co-purified by immobilised metal affinity chromatography. Biochimica et Biophysica Acta (BBA) - General Subjects, 1760, 1304-1313. Bornhorst, J.A. and Falke, J.J. (2000) [16] Purification of proteins using polyhistidine affinity tags. In Thorner, J., Emr, S.D. and Abelson, J.N. (eds.), Methods in Enzy- mology. Academic Press, Vol. 326, pp. 245-254. Branden, C. and Tooze, J. (1999) Introduction to Protein Structure. Garland Publishing, Inc., New York. Braun, P. and LaBaer, J. (2003) High throughput protein production for functional proteomics. Trends in Biotechnology, 21, 383-388. Cabantous, S. and Waldo, G.S. (2006) In vivo and in vitro protein solubility assays using split GFP. Nature Methods, 3, 845-854.

45 Calin, G.A., di Iasio, M.G., Caprini, E., Vorechovsky, I., Natali, P.G., Sozzi, G., Croce, C.M., Barbanti-Brodano, G., Russo, G. and Negrini, M. (2000) Low frequency of alterations of the alpha (PPP2R1A) and beta (PPP2R1B) isoforms of the subunit A of the serine-threonine phosphatase 2A in human neoplasms. On- cogene, 19, 1191-1195. Cayla, X., Goris, J., Hermann, J., Hendrix, P., Ozon, R. and Merlevede, W. (1990) Isolation and characterization of a tyrosyl phosphatase activator from rabbit skeletal muscle and Xenopus laevis oocytes. Biochemistry, 29, 658-667. Ceulemans, H. and Bollen, M. (2004) Functional Diversity of Protein Phosphatase-1, a Cellu- lar Economizer and Reset Button. Physiological Reviews, 84, 1-39. Chao, Y., Xing, Y.N., Chen, Y., Xu, Y.H., Lin, Z., Li, Z., Jeffrey, P.D., Stock, J.B. and Shi, Y.G. (2006) Structure and mechanism of the phosphotyrosyl phosphatase activator. Molecular Cell, 23, 535-546. Chen, J., Martin, B.L. and Brautigan, D.L. (1992) Regulation of Protein Serine-Threonine Phosphatase Type-2a by Tyrosine Phosphorylation. Science, 257, 1261- 1264. Chen, J.H., St-Germain, J.R. and Li, Q. (2005) B56 regulatory subunit of protein phosphatase 2A mediates valproic acid-induced p300 degradation. Molecular and Cellu- lar Biology, 25, 525-532. Chen, W., Possemato, R., Campbell, K.T., Plattner, C.A., Pallas, D.C. and Hahn, W.C. (2004) Identification of specific PP2A complexes involved in human cell transformation. Cancer Cell, 5, 127-136. Chen, Y., Xu, Y., Bao, Q., Xing, Y., Li, Z., Lin, Z., Stock, J.B., Jeffrey, P.D. and Shi, Y. (2007) Structural and biochemical insights into the regulation of protein phosphatase 2A by small t antigen of SV40. Nature Structural Biology, 14, 527-534. Cho, U.S., Morrone, S., Sablina, A.A., Arroyo, J.D., Hahn, W.C. and Xu, W. (2007) Struc- tural Basis of PP2A Inhibition by Small t Antigen. PLoS Biology, 5, e202. Cho, U.S. and Xu, W.Q. (2007) Crystal structure of a protein phosphatase 2A heterotrimeric holoenzyme. Nature, 445, 53-57. Chou, D.M., Petersen, P., Walter, J.C. and Walter, G. (2002) Protein phosphatase 2A regulates binding of Cdc45 to the prereplication complex. Journal of Biological Chemistry, 277, 40520-40527. Chung, H.Y., Nairn, A.C., Murata, K. and Brautigan, D.L. (1999) Mutation of Tyr307 and Leu309 in the protein phosphatase 2A catalytic subunit favors association with the alpha 4 subunit which promotes dephosphorylation of elongation factor-2. Biochemistry, 38, 10371-10376. Cohen, P.T.W. (1997) Novel protein serine/threonine phosphatases: Variety is the spice of life. Trends in Biochemical Sciences, 22, 245-251. Cohen, P.T.W., Brewis, N.D., Hughes, V. and Mann, D.J. (1990) Protein Serine Threonine Phosphatases - an Expanding Family. FEBS Letters, 268, 355-359. Cohen, P.T.W., Philp, A. and Vazquez-Martin, C. (2005) Protein phosphatase 4 - from obscu- rity to vital functions. FEBS Letters, 579, 3278-3286. Consortium, S.G. (2008) Protein production and purification. Nature Methods, 5, 135-146. Conti, E. and Izaurralde, E. (2001) Nucleocytoplasmic transport enters the atomic age. Cur- rent Opinion in Cell Biology, 13, 310-319. Cornvik, T., Dahlroth, S.L., Magnusdottir, A., Flodin, S., Engvall, B., Lieu, V., Ekberg, M. and Nordlund, P. (2006) An efficient and generic strategy for producing soluble human proteins and domains in E-coli by screening construct libraries. Proteins-Structure Function and Bioinformatics, 65, 266-273. Cornvik, T., Dahlroth, S.L., Magnusdottir, A., Herman, M.D., Knaust, R., Ekberg, M. and Nordlund, P. (2005) Colony filtration blot: a new screening method for soluble protein expression in Escherichia coli. Nature Methods, 2, 507-509. Dagda, R.K., Zaucha, J.A., Wadzinski, B.E. and Strack, S. (2003) A developmentally regulated, neuron-specific splice variant of the variable subunit B beta targets

46 protein phosphatase 2A to mitochondria and modulates apoptosis. Journal of Biological Chemistry, 278, 24976-24985. Dale, G.E., Oefner, C. and D'Arcy, A. (2003) The protein as a variable in protein crystallization. Journal of Structural Biology, 142, 88-97. Damuni, Z., Xiong, H.S. and Li, M. (1994) Autophosphorylation-Activated Protein-Kinase Inactivates the Protein-Tyrosine-Phosphatase Activity of Protein Phos- phatase 2a. FEBS Letters, 352, 311-314. Das, A., Helps, N., Cohen, P. and Barford, D. (1996) Crystal structure of the protein serine/threonine phosphatase 2C at 2.0 angstrom resolution. EMBO Journal, 15, 6798-6809. Egloff, M.P., Cohen, P.T.W., Reinemer, P. and Barford, D. (1995) Crystal-Structure of the Catalytic Subunit of Human Protein Phosphatase-1 and Its Complex with Tungstate. Journal of Molecular Biology, 254, 942-959. Eichhorn, P.J.A., Creyghton, M.P. and Bernards, R. (2008) Protein phosphatase 2A regulatory subunits and cancer. Biochimica et Biophysica Acta (BBA) - Reviews on Cancer, In Press. Endo, Y. and Sawasaki, T. (2006) Cell-free expression systems for eukaryotic protein production. Current Opinion in Biotechnology, 17, 373-380. Evans, D.R.H. and Hemmings, B.A. (2000) Mutation of the C-terminal leucine residue of PP2Ac inhibits PR55/B subunit binding and confers supersensitivity to microtubule destabilization in Saccharomyces cerevisiae. Molecular and General Genetics, 264, 425-432. Fellner, T., Lackner, D.H., Hombauer, H., Piribauer, P., Mudrak, I., Zaragoza, K., Juno, C. and Ogris, E. (2003) A novel and essential mechanism determining specificity and activity of protein phosphatase 2A (PP2A) in vivo. Genes & De- velopment, 17, 2138-2150. Forstner, M., Leder, L. and Mayr, L.M. (2007) Optimization of protein expression systems for modern drug discovery. Expert Review of Proteomics, 4, 67-78. Gaberc-Porekar, V. and Menart, V. (2001) Perspectives of immobilized-metal affinity chromatography. Journal of Biochemical and Biophysical Methods, 49, 335- 360. Gallego, M. and Virshup, D.M. (2005) Protein serine/threonine phosphatases: life, death, and sleeping. Current Opinion in Cell Biology, 17, 197-202. Gasser, B., Saloheimo, M., Rinas, U., Dragosits, M., Rodríguez-Carmona, E., Baumann, K., Giuliani, M., Parrilli, E., Branduardi, P., C, L., Porro, D., Ferrer, P., Tutino, M., Mattanovich, D. and Villaverde, A. (2008) Protein folding and conformational stress in microbial cells producing recombinant proteins: a host comparative overview. Microbial Cell Factories, 7, 11. Gentry, M.S., Li, Y.K., Wei, H.J., Syed, F.F., Patel, S.H., Hallberg, R.L. and Pallas, D.C. (2005) A novel assay for protein phosphatase 2A (PP2A) complexes in vivo reveals differential effects of covalent modifications on different Sac- charomyces cerevisiae PP2A heterotrimers. Eukaryotic Cell, 4, 1029-1040. Goldberg, J., Huang, H.-b., Kwon, Y.-g., Greengard, P., Nairn, A.C. and Kuriyan, J. (1995) Three-dimensional structure of the catalytic subunit of protein serine/threonine phosphatase-1. Nature, 376, 745-753. Goldberg, Y. (1999) Protein phosphatase 2A: Who shall regulate the regulator? Biochemical Pharmacology, 57, 321-328. Groves, M.R., Hanlon, N., Turowski, P., Hemmings, B.A. and Barford, D. (1999) The structure of the protein phosphatase 2A PR65/A subunit reveals the conformation of its 15 tandemly repeated HEAT motifs. Cell, 96, 99-110. Götz, J., Probst, A., Ehler, E., Hemmings, B. and Kues, W. (1998) Delayed embryonic lethality in mice lacking protein phosphatase 2A catalytic subunit. Proceedings of the National Academy of Sciences of the United States of America, 95, 12370-12375.

47 Hemmings, B.A., Adams-Pearson, C., Maurer, F., Mueller, P., Goris, J., Merlevede, W., Hofsteenge, J. and Stone, S.R. (1990) α- and β-forms of the 65-kDa subunit of protein phosphatase 2A have a similar 39 amino acid repeating structure. Biochemistry, 29, 3166-3173. Hengen, P.N. (1995) Purification of His-Tag fusion proteins from Escherichia coli. Trends in Biochemical Sciences, 20, 285-286. Hombauer H, W.D., Mudrak I, Stanzel C, Fellner T, Lackner DH, Ogris E. (2007) Generation of Active Protein Phosphatase 2A Is Coupled to Holoenzyme Assembly. PLoS Biol., 5, 1355-1365. Hu, P.R., Yu, L., Zhang, M., Zheng, L.H., Zhao, Y., Fu, Q. and Zhao, S.Y. (2000) Molecular cloning and mapping of the brain-abundant B1 gamma subunit of protein phosphatase 2A, PPP2R2C, to human chromosome 4p16. Genomics, 67, 83-86. Inoki, K., Ouyang, H., Li, Y. and Guan, K.L. (2005) Signaling by target of rapamycin proteins in cell growth control. Microbiology and Molecular Biology Reviews, 69, 79-100. Ito, A., Koma, Y., Watabe, K., Nagano, T., Endo, Y., Nojima, H. and Kitamura, Y. (2003) A truncated isoform of the protein phosphatase 2A b56 gamma regulatory subunit may promote genetic instability and cause tumor progression. American Journal of Pathology, 162, 81-91. Janssens, V. and Goris, J. (2001) Protein phosphatase 2A: a highly regulated family of serine/threonine phosphatases implicated in cell growth and signalling. Bio- chemical Journal, 353, 417-439. Janssens, V., Goris, J. and Van Hoof, C. (2005) PP2A: the expected tumor suppressor. Cur- rent Opinion in Genetics & Development, 15, 34-41. Johnson, L.N. and Barford, D. (1993) The Effects of Phosphorylation on the Structure and Function of Proteins. Annual Review of Biophysics and Biomolecular Struc- ture, 22, 199-232. Jordens, J., Janssens, V., Longin, S., Stevens, I., Martens, E., Bultynck, G., Engelborghs, Y., Lescrinier, E., Waelkens, E., Goris, J. and Van Hoof, C. (2006) The Protein Phosphatase 2A Phosphatase Activator Is a Novel Peptidyl-Prolyl cis/trans- Isomerase. Journal of Biological Chemistry, 281, 6349-6357. Junttila, M.R., Puustinen, P., Niemelä, M., Ahola, R., Arnold, H., Böttzauw, T., Ala-aho, R., Nielsen, C., Ivaska, J., Taya, Y., Lu, S.-L., Lin, S., Chan, E.K.L., Wang, X.- J., Grènman, R., Kast, J., Kallunki, T., Sears, R., Kähäri, V.-M. and Westermarck, J. (2007) CIP2A Inhibits PP2A in Human Malignancies. Cell, 130, 51-62. Kamibayashi, C., Lickteig, R.L., Estes, R., Walter, G. and Mumby, M.C. (1992) Expression of the a-Subunit of Protein Phosphatase-2a and Characterization of Its In- teractions with the Catalytic and Regulatory Subunits. Journal of Biological Chemistry, 267, 21864-21872. Kennelly, P.J. (2001) Protein phosphatases - A phylogenetic perspective. Chemical Reviews, 101, 2291-2312. Kissinger, C.R., Parge, H.E., Knighton, D.R., Lewis, C.T., Pelletier, L.A., Tempczyk, A., Kalish, V.J., Tucker, K.D., Showalter, R.E., Moomaw, E.W., Gastinel, L.N., Habuka, N., Chen, X., Maldonado, F., Barker, J.E., Bacquet, R. and Villafranca, J.E. (1995) Crystal structures of human calcineurin and the human FKBP12-FK506-calcineurin complex. Nature, 378, 641-644. Klumpp, S. and Krieglstein, J. (2002) Serine/threonine protein phosphatases in apoptosis. Current Opinion in Pharmacology, 2, 458-462. Kremmer, E., Ohst, K., Kiefer, J., Brewis, N. and Walter, G. (1997) Separation of PP2A core enzyme and holoenzyme with monoclonal antibodies against the regulatory A subunit: Abundant expression of both forms in cells. Molecular and Cel- lular Biology, 17, 1692-1701.

48 Kuo, Y.-C., Huang, K.-Y., Yang, C.-H., Yang, Y.-S., Lee, W.-Y. and Chiang, C.-W. (2008) Regulation of Phosphorylation of Thr-308 of Akt, Cell Proliferation, and Survival by the B55{alpha} Regulatory Subunit Targeting of the Protein Phosphatase 2A Holoenzyme to Akt. Journal of Biological Chemistry, 283, 1882-1892. Letourneux, C., Rocher, G. and Porteu, F. (2006) B56-containing PP2A dephosphorylate ERK and their activity is controlled by the early gene IEX-1 and ERK. EMBO Journal, 25, 727-738. Leulliot, N., Vicentini, G., Jordens, J., Quevillon-Cheruel, S., Schiltz, M., Barford, D., Van Tilbeurgh, H. and Goris, J. (2006) Crystal structure of the PP2A phosphatase activator: Implications for its PP2A-specific PPlase activity. Mo- lecular Cell, 23, 413-424. Levitt, M. (2007) Growth of novel protein structural data. Proceedings of the National Acad- emy of Sciences, 104, 3183-3188. Li, H.H., Cai, X., Shouse, G.P., Piluso, L.G. and Liu, X.A. (2007) A specific PP2A regulatory subunit, B56 gamma, mediates DNA damage-induced dephosphorylation of p53 at Thr55. EMBO Journal, 26, 402-411. Li, X.H., Scuderi, A., Letsou, A. and Virshup, D.M. (2002) B56-associated protein phosphatase 2A is required for survival and protects from apoptosis in Droso- phila melanogaster. Molecular and Cellular Biology, 22, 3674-3684. Li, X.H. and Virshup, D.M. (2002) Two conserved domains in regulatory B subunits mediate binding to the A subunit of protein phosphatase 2A. European Journal of Biochemistry, 269, 546-552. Lizotte, D.L., Blakeslee, J.J., Siryaporn, A., Heath, J.T. and Alison DeLong. (2008) A PP2A active site mutant impedes growth and causes misregulation of native catalytic subunit expression. Journal of Cellular Biochemistry, 103, 1309-1325. Longin, S., Jordens, J., Martens, E., Stevens, I., Janssens, V., Rondelez, E., De Baere, I., Derua, R., Waelkens, E., Goris, J. and Van Hoof, C. (2004) An inactive protein phosphatase 2A population is associated with methylesterase and can be re-activated by the phosphotyrosyl phosphatase activator. Biochemi- cal Journal, 380, 111-119. Magenta, A., Fasanaro, P., Romani, S., Di Stefano, V., Capogrossi, M.C. and Martelli, F. (2008) Protein Phosphatase 2A Subunit PR70 Interacts with pRb and Medi- ates Its Dephosphorylation. Molecular and Cellular Biology, 28, 873-882. Magnusdottir, A., Stenmark, P., Flodin, S., Nyman, T., Hammarstrom, M., Ehn, M., Bakali- H, M.A., Berglund, H. and Nordlund, P. (2006) The crystal structure of a human PP2A phosphatase activator reveals a novel fold and highly conserved cleft implicated in protein-protein interactions. Journal of Biological Chemistry, 281, 22434-22438. Main, E.R.G., Lowe, A.R., Mochrie, S.G.J., Jackson, S.E. and Regan, L. (2005) A recurring theme in protein engineering: the design, stability and folding of repeat proteins. Current Opinion in Structural Biology, 15, 464-471. Manjasetty, B., Turnbull, A., Panjikar, S., Büssow, K. and Chance, M.R. (2008) Automated technologies and novel techniques to accelerate protein crystallography for structural genomics. Proteomics, 8, 612-625. Margolis, S.S., Perry, J.A., Forester, C.M., Nutt, L.K., Guo, Y., Jardim, M.J., Thomenius, M.J., Freel, C.D., Darbandi, R. and Ahn, J.-H. (2006) Role for the PP2A/B56[delta] Phosphatase in Regulating 14-3-3 Release from Cdc25 to Control Mitosis. Cell, 127, 759-773. McCright, B., Brothman, A.R. and Virshup, D.M. (1996a) Assignment of human protein phosphatase 2A regulatory subunit genes B56 alpha, B56 beta, B56 gamma, B56 delta, and B56 epsilon (PPP2R5A-PPP2R5E), highly expressed in muscle and brain, to chromosome regions 1q41, 11q12, 3p21, 6p21.1, and 7p11.2->p12. Genomics, 36, 168-170.

49 McCright, B., Rivers, A.M., Audlin, S. and Virshup, D.M. (1996b) The B56 family of protein phosphatase 2A (PP2A) regulatory subunits encodes differentiation-induced phosphoproteins that target PP2A to both nucleus and cytoplasm. Journal of Biological Chemistry, 271, 22081-22089. McCright, B. and Virshup, D.M. (1995) Identification of a New Family of Protein Phos- phatase 2a Regulatory Subunits. Journal of Biological Chemistry, 270, 26123-26128. McPherson, A. (2003) Introduction to Macromolecular Crystallography. Whiley-Liss, Hobo- ken, NJ. Mitic, N., Smith, S.J., Neves, A., Guddat, L.W., Gahan, L.R. and Schenk, G. (2006) The Catalytic Mechanisms of Binuclear Metallohydrolases. Chem. Rev., 106, 3338-3363. Moreno, C.S., Park, S., Nelson, K., Ashby, D., Hubalek, F., Lane, W.S. and Pallas, D.C. (2000) WD40 repeat proteins striatin and S/G(2) nuclear autoantigen are members of a novel family of calmodulin-binding proteins that associate with protein phosphatase 2A. Journal of Biological Chemistry, 275, 5257- 5263. Moreno, C.S., Ramachandran, S., Ashby, D.G., Laycock, N., Plattner, C.A., Chen, W., Hahn, W.C. and Pallas, D.C. (2004) Signaling and Transcriptional Changes Criti- cal for Transformation of Human Cells by Simian Virus 40 Small Tumor Antigen or Protein Phosphatase 2A B56{gamma} Knockdown. Cancer Re- search, 64, 6978-6988. Nilsson, J., Ståhl, S., Lundeberg, J., Uhlén, M. and Nygren, P.-å. (1997) Affinity Fusion Strategies for Detection, Purification, and Immobilization of Recombinant Proteins. Protein Expression and Purification, 11, 1-16. Ogris, E., Gibson, D.M. and Pallas, D.C. (1997) Protein phosphatase 2A subunit assembly: the catalytic subunit carboxy terminus is important for binding cellular B subunit but not polyomavirus middle tumor antigen. Oncogene, 15, 911- 917. O'Hearn, E., Holmes, S.E., Calvert, P.C., Ross, C.A. and Margolis, R.L. (2001) SCA-12: Tremor with cerebellar and cortical atrophy is associated with a CAG repeat expansion. Neurology, 56, 299-303. Ortega-Gutiérrez, S., Leung, D., Ficarro, S., Peters, E.C. and Cravatt, B.F. (2008) Targeted Disruption of the PME-1 Gene Causes Loss of Demethylated PP2A and Perinatal Lethality in Mice. PLoS ONE, 3, e2486. Oswald, T., Hornbostel, G., Rinas, U. and Anspach, F.B. (1997) Purification of (His)6EcoRV [recombinant restriction endonuclease EcoRV fused to a (His)6affinity domain] by metal-chelate affinity chromatography. Biotechnology and Ap- plied Biochemistry, 25, 109-115. Perrotti, D. and Neviani, P. (2008) Protein phosphatase 2A (PP2A), a drugable tumor suppressor in Ph1(+) leukemias. Cancer and Metastasis Reviews, 27, 159-168. Petersen, P., Chou, D.M., You, Z., Hunter, T., Walter, J.C. and Walter, G. (2006) Protein Phosphatase 2A Antagonizes ATM and ATR in a Cdk2- and Cdc7- Independent DNA Damage Checkpoint. Molecular and Cellular Biology, 26, 1997-2011. Prickett, T.D. and Brautigan, D.L. (2006) The alpha 4 regulatory subunit exerts opposing allosteric effects on protein phosphatases PP6 and PP2A. Journal of Bio- logical Chemistry, 281, 30503-30511. Rempola, B., Kaniak, A., Migdalski, A., Rytka, J., Slonimski, P.P. and di Rago, J.P. (2000) Functional analysis of RRD1 (YIL153w) and RRD2 (YPL152w), which en- code two putative activators of the phosphotyrosyl phosphatase activity of PP2A in Saccharomyces cerevisiae. Molecular and General Genetics, 262, 1081-1092.

50 Ruediger, R., Pham, H.T. and Walter, G. (2001) Disruption of protein phosphatase 2A subunit interaction in human cancers with mutations in the A alpha subunit gene. Oncogene, 20, 10-15. Sablina, A. and Hahn, W. (2008) SV40 small T antigen and PP2A phosphatase in cell transformation. Cancer and Metastasis Reviews, 27, 137-146. Savva, R., Prodromou, C. and Driscoll, P.C. (2007) DNA fragmentation based combinatorial approaches to soluble protein expression: Part II: Library expression, screening and scale-up. Drug Discovery Today, 12, 939-947. Schafer, F., Blumer, J., Romer, U. and Steinert, K. (2000) Ni-NTA for large-scale IMAC processes-systematic investigations of separeation characteristics, storage and CIP conditions, and leaching. Qiagen News, 4, 11-15. Schonthal, A.H. (2001) Role of serine/threonine protein phosphatase 2A in cancer. Cancer Letters, 170, 1-13. Seeling, J.M., Miller, J.R., Gil, R., Moon, R.T., White, R. and Virshup, D.M. (1999) Regula- tion of beta-catenin signaling by the B56 subunit of protein phosphatase 2A. Science, 283, 2089-2091. Silverstein, A.M., Barrow, C.A., Davis, A.J. and Mumby, M.C. (2002) Actions of PP2A on the MAP kinase pathway and apoptosis are mediated by distinct regulatory subunits. Proceedings of the National Academy of Sciences of the United States of America, 99, 4221-4226. Sontag, E. (2001) Protein phosphatase 2A: the Trojan Horse of cellular signaling. Cellular Signalling, 13, 7-16. Sontag, E., Hladik, C., Montgomery, L., Luangpirom, A., Mudrak, I., Ogris, E. and White, C.L. (2004) Downregulation of protein phosphatase 2A carboxyl methylation and methyltransferase may contribute to Alzheimer disease pathogene- sis. Journal of Neuropathology and Experimental Neurology, 63, 1080- 1091. Sorensen, H. and Mortensen, K. (2005) Soluble expression of recombinant proteins in the cytoplasm of Escherichia coli. Microbial cell factories, 4, 1. Stefansson, B., Ohama, T., Daugherty, A.E. and Brautigan, D.L. (2008) Protein Phosphatase 6 Regulatory Subunits Composed of Ankyrin Repeat Domains. Biochemistry, 47, 1442-1451. Stevens, I., Janssens, V., Martens, E., Dilworth, S., Goris, J. and Van Hoof, C. (2003) Identi- fication and characterization of B ''-subunits of protein phosphatase 2 A in Xenopus laevis oocytes and adult tissues - Evidence for an independent N- terminal splice variant of PR130 and an extended human PR48 protein. European Journal of Biochemistry, 270, 376-387. Strack, S., Cribbs, J.T. and Gomez, L. (2004) Critical role for protein phosphatase 2A heterotrimers in mammalian cell survival. Journal of Biological Chemistry, 279, 47732-47739. Strack, S., Ruediger, R., Walter, G., Dagda, R.K., Barwacz, C.A. and Cribbs, J.T. (2002) Protein phosphatase 2A holoenzyme assembly - Identification of contacts between B-family regulatory and scaffolding A subunits. Journal of Bio- logical Chemistry, 277, 20750-20755. Swingle, M.R., Honkanen, R.E. and Ciszak, E.M. (2004) Structural Basis for the Catalytic Activity of Human Serine/Threonine Protein Phosphatase-5. Journal of Bio- logical Chemistry, 279, 33992-33999. Tamaki, M., Goi, T., Hirono, Y., Katayama, K. and Yamaguchi, A. (2004) PPP2R1B gene alterations inhibit interaction of PP2A-A beta and PP2A-C proteins in colo- rectal cancers. Oncology Reports, 11, 655-659. Tolstykh, T., Lee, J., Vafai, S. and Stock, J.B. (2000) Carboxyl methylation regulates phosphoprotein phosphatase 2A by controlling the association of regulatory B subunits. EMBO Journal, 19, 5682-5691. Trockenbacher, A., Suckow, V., Foerster, J., Winter, J., Krausz, S., Ropers, H.-H., Schneider, R. and Schweiger, S. (2001) MID1, mutated in Opitz syndrome, encodes an

51 ubiquitin ligase that targets phosphatase 2A for degradation. Nature Genet- ics, 29, 287-294. Turowski, P., Favre, B., Campbell, K.S., Lamb, N.J.C. and Hemmings, B.A. (1997) Modula- tion of the enzymatic properties of protein phosphatase 2A catalytic subunit by the recombinant 65-kDa regulatory subunit PR65 alpha. European Jour- nal of Biochemistry, 248, 200-208. Van Kanegan, M.J., Adams, D.G., Wadzinski, B.E. and Strack, S. (2005) Distinct protein phosphatase 2A heterotrimers modulate growth factor signaling to extracel- lular signal-regulated kinases and Akt. Journal of Biological Chemistry, 280, 36029-36036. Vanhoof, C., Cayla, X., Bosch, M., Merlevede, W. and Goris, J. (1994) The Phosphotyrosyl Phosphatase Activator of Protein Phosphatase 2a - a Novel Purification Method, Immunological and Enzymatic Characterization. European Jour- nal of Biochemistry, 226, 899-907. Voorhoeve, P.M., Hijmans, E.M. and Bernards, R. (1999) Functional interaction between a novel protein phosphatase 2A regulatory subunit, PR59, and the retinoblastoma-related p107 protein. Oncogene, 18, 515-524. Wall, M.E., Ealick, S.E. and Gruner, S.M. (1997) Three-dimensional diffuse x-ray scattering from crystals of Staphylococcal nuclease. Proceedings of the National Academy of Sciences of the United States of America, 94, 6180-6184. Wang, S.S., Esplin, E.D., Li, J.L., Huang, L.Y., Gazdar, A., Minna, J. and Evans, G.A. (1998) Alterations of the PPP2R1B gene in human lung and colon cancer. Science, 282, 284-287. Wang, Y. and Burke, D. (1997) Cdc55p, the B-type regulatory subunit of protein phosphatase 2A, has multiple functions in mitosis and is required for the kineto- chore/spindle checkpoint in Saccharomyces cerevisiae. Molecular and Cel- lular Biology, 17, 620-626. Waugh, D.S. (2005) Making the most of affinity tags. Trends in Biotechnology, 23, 316-320. Wu, J., Tolstykh, T., Lee, J., Boyd, K., Stock, J.B. and Broach, J.R. (2000) Carboxyl methylation of the phosphoprotein phosphatase 2A catalytic subunit promotes its functional association with regulatory subunits in vivo. EMBO Journal, 19, 5672-5681. Xing, Y., Li, Z., Chen, Y., Stock, J.B., Jeffrey, P.D. and Shi, Y. (2008) Structural Mechanism of Demethylation and Inactivation of Protein Phosphatase 2A. Cell, 133, 154-163. Xing, Y., Xu, Y., Chen, Y., Jeffrey, P.D., Chao, Y., Lin, Z., Li, Z., Strack, S., Stock, J.B. and Shi, Y. (2006) Structure of Protein Phosphatase 2A Core Enzyme Bound to Tumor-Inducing Toxins. Cell, 127, 341-353. Xu, Y., Xing, Y., Chen, Y., Chao, Y., Lin, Z., Fan, E., Yu, J.W., Strack, S., Jeffrey, P.D. and Shi, Y. (2006) Structure of the Protein Phosphatase 2A Holoenzyme. Cell, 127, 1239-1251. Xu, Z. and Williams, B.R.G. (2000) The B56 alpha regulatory subunit of protein phosphatase 2A is a target for regulation by double-stranded RNA-dependent protein kinase PKR. Molecular and Cellular Biology, 20, 5285-5299. Yan, L., Lavin, V.A., Moser, L.R., Cui, Q., Kanies, C. and Yang, E. (2008) PP2A Regulates the Pro-apoptotic Activity of FOXO1. Journal of Biological Chemistry, 283, 7411-7420. Yan, Z., Fedorov, S.A., Mumby, M.C. and Williams, R.S. (2000) PR48, a novel regulatory subunit of protein phosphatase 2A, interacts with Cdc6 and modulates DNA replication in human cells. Molecular and Cellular Biology, 20, 1021-1029. Yang, J., Wu, J.L., Tan, C. and Klein, P.S. (2003) PP2A : B56 epsilon is required for Wnt/beta-catenin signaling during embryonic development. Development, 130, 5569-5578. Yu, X.X., Du, X.X., Moreno, C.S., Green, R.E., Ogris, E., Feng, Q., Chou, L., McQuoid, M.J. and Pallas, D.C. (2001) Methylation of the protein phosphatase 2A catalytic

52 subunit is essential for association of B alpha regulatory subunit but not SG2NA, striatin, or polyomavirus middle tumor antigen. Molecular Biology of the Cell, 12, 185-199. Zheng, Y. and Jiang, Y. (2005) The yeast phosphotyrosyl phosphatase activator is part of the tap42-phosphatase complexes. Molecular Biology of the Cell, 16, 2119- 2127. Zhou, J., Pham, H.T. and Waltert, G. (2003) The formation and activity of PP2A holoenzymes do not depend on the isoform of the catalytic subunit. Journal of Bio- logical Chemistry, 278, 8617-8622. Zhou, X.-W., Li, X., Bjorkdahl, C., Sjogren, M.J., Alafuzoff, I., Soininen, H., Grundke-Iqbal, I., Iqbal, K., Winblad, B. and Pei, J.-J. (2006) Assessments of the accumulation severities of amyloid [beta]-protein and hyperphosphorylated tau in the medial temporal cortex of control and Alzheimer's brains. Neurobiology of Disease, 22, 657-668.