57 Has Characteristics of an

Sep 13, 2008 - We observed this protein as an elongated particle of low compactness, leading us to conclude that it belongs to the increasing family o...
11 downloads 14 Views 5MB Size
Download PDF
Human Regulatory Protein Ki-1/57 Has Characteristics of an Intrinsically Unstructured Protein Gustavo C. Bressan,†,‡ Ju ´ lio C. Silva,†,| Ju ´ lio C. Borges,§ Dario O. dos Passos,† ⊥ Carlos H. I. Ramos, Iris L. Torriani,†,| and Jo ¨ rg Kobarg*,†,‡ National Synchrotron Light Laboratory - LNLS, 13083-970, Campinas, SP, Brazil, Institute of Biology, State University of Campinas - UNICAMP, 13083-970, Campinas, SP, Brazil, Institute of Chemistry of Sa˜o Carlos, University of Sa˜o Paulo, CEP 13560-970, Sa˜o Carlos, SP, Brazil, Institute of Physics “Gleb Wataghin”, State University of Campinas - UNICAMP, 13083-970, Campinas, SP, Brazil, and Institute of Chemistry, University of Campinas - UNICAMP, 13083-970, Campinas, SP, Brazil Received July 15, 2008

The human protein Ki-1/57 was first identified through the cross reactivity of the anti-CD30 monoclonal antibody Ki-1, in Hodgkin lymphoma cells. The expression of Ki-1/57 in diverse cancer cells and its phosphorylation in peripheral blood leukocytes after mitogenic activation suggested its possible role in cell signaling. Ki-1/57 interacts with several other regulatory proteins involved in cellular signaling, transcriptional regulation and RNA metabolism, suggesting it may have pleiotropic functions. In a previous spectroscopic analysis, we observed a low content of secondary structure for Ki-1/57 constructs. Here, Circular dichroism experiments, in vitro RNA binding analysis, and limited proteolysis assays of recombinant Ki-1/57(122-413) and proteolysis assays of endogenous full length protein from human HEK293 cells suggested that Ki-1/57 has characteristics of an intrinsically unstructured protein. Smallangle X-ray scattering (SAXS) experiments were performed with the C-terminal fragment Ki-1/ 57(122-413). These results indicated an elongated shape and a partially unstructured conformation of the molecule in solution, confirming the characteristics of an intrinsically unstructured protein. Experimental curves together with ab initio modeling approaches revealed an extended and flexible molecule in solution. An elongated shape was also observed by analytical gel filtration. Furthermore, sedimentation velocity analysis suggested that Ki-1/57 is a highly asymmetric protein. These findings may explain the functional plasticity of Ki-1/57, as suggested by the wide array of proteins with which it is capable of interacting in yeast two-hybrid interaction assays. Keywords: intrinsically unstructured protein (IUP) • SAXS • protein-protein interaction • circular dichroism • limited proteolysis • hydrodynamic characterization

Introduction Analysis of malignant Hodgkin and Sternberg-Reed cells in the Hodgkin lymphoma revealed that the monoclonal antibody Ki-1 not only reacted with the trans-membrane CD30 protein of 120 kDa, but also with an intracellular antigen of 57 kDa, called Ki-1/57.1-4 Further studies showed that Ki-1/57 is a phosphor-protein associated with serine-threonine protein kinase activity3 and is localized in the cytoplasm and nucleus, where it is frequently found in association with several nuclear substructures.5 Because Ki-1/57 was also found to bind in vitro to hyaluronic acid and other negatively charged glycosami* To whom correspondence should be addressed. Jo¨rg Kobarg, Laborato´rio Nacional de Luz Sı´ncrotron, Centro de Biologia Molecular Estrutural, Rua Giuseppe Ma´ximo Scolfaro 10.000, C.P. 6192, 13084-971 Campinas - SP, Brazil. Tel: (+55)19-3512-1125. Fax: (+55)19-3512-1006. E-mail: jkobarg@ lnls.br. † National Synchrotron Light Laboratory. ‡ Institute of Biology, UNICAMP. | Institute of Physics “Gleb Wataghin”, UNICAMP. § University of Sa˜o Paulo. ⊥ Institute of Chemistry, UNICAMP. 10.1021/pr8005342 CCC: $40.75

 2008 American Chemical Society

noglicans, it was also named intracellular hyaluronic acid binding protein 4 (IHABP4).6 The human protein CGI-55 has 40% of amino acid sequence identity and 67% of similarity with Ki-1/57, suggesting these two proteins could be paralogues. Recent studies have revealed the interaction of both Ki-1/57 and CGI-55 with the chromatin remodeling protein CHD-37 and several others transcriptional regulators, such as Topors, Daxx and PIAS-3.8,9 Furthermore, Ki-1/57 was found to interact with proteins involved in RNA metabolism, such as PRMT-1,10 cellular signaling, such as the PKC adaptor protein RACK-1,11 and several p53 interacting proteins.9 Moreover, Ki-1/57 associates itself with p53 and has an inhibitory effect on its transcriptional activity.9 Although the exact cellular role of Ki-1/57 is still unknown, its protein-protein interaction profile indicates that it may be involved in gene expression regulation either on the transcriptional level or via its influence on RNA metabolism. The analysis of the structural properties of Ki-1/57 is an important step to clarify its functional activity. We previously performed spectroscopic analyses of the interaction between Journal of Proteome Research 2008, 7, 4465–4474 4465 Published on Web 09/13/2008

research articles the signaling adaptor protein RACK-1 and constructs of the C-terminal region of Ki-1/57.12 The observed low content of regular secondary structure of these constructs has suggested that Ki-1/57 may belong to the class of the intrinsically disordered proteins. This fastly growing class of proteins has gained much attention lately, since many of its members are involved in regulatory mechanisms, signal transduction, noncatalytic interaction with DNA/RNA and other proteins.13-15 It is believed that the inherent flexibility of these proteins may be crucial to their functions, which commonly demand for plasticity and for the combination of high specificity and low affinity during interaction with its molecular partners.16,17,14 Moreover, the observation that many proteins or domains that are functionally associated with cancer and others human diseases have long disordered regions18,19 further highlights the importance of this new class of proteins. It has been even suggested that this group of proteins necessitates new and alternative efforts in drug design.19 In this study, we used low resolution approaches to characterize the structural properties of a C-terminal construct of Ki-1/57 spanning the amino acids 122-413. Together with theoretical predictions, all experiments we performed support the hypothesis that the protein Ki-1/57 belongs to the class of the intrinsically unstructured proteins. Our SAXS study reveals an extended and flexible conformation for the Ki-1/57(122-413) suggesting the presence of largely unfolded regions. In agreement with this, an elongated shape was also observed in the size-exclusion chromatography and analytical ultracentrifugation experiments. Furthermore, Ki-1/57(122-413) as well as the endogenous Ki-1/57 from human HEK293 cell lysates was very sensible to limited proteolysis. Finally, circular dichroism experiments demonstrated that Ki-1/57 shows no significant change in its secondary structure upon binding to poly-U RNA, however gained R-helical content upon addition of 5-10% of TFE. In summary, our data pointed that Ki-1/57 has features of an intrinsically unstructured protein, supplying new directions to explain and study the plasticity of its protein interaction profile.

Results and Discussion Primary Sequence Analysis. Particular sequence signatures have been observed for intrinsically unstructured proteins, which permit the prediction of their structural disorder from amino acids sequence.20 The comparison analysis of a set of globular proteins against a set of disordered proteins has allowed the characterization of specific features of natively unfolded protein sequences.16 It has been shown that disordered proteins sequences are enriched in disorder-promoting (A, R, G, Q, S, P, E, K) and depleted in order-promoting amino acids (W, C, F, I, Y, V, L, N).16 In this way, in comparison to globular proteins, the whole amino acid sequence of Ki-1/57 as well as its C-terminal region studied here (residues 122-413) are enriched in disorder-promoting residues (63.9 and 59.9%, respectively) and depleted in order-promoting amino acids residues (19.8 and 21.4%, respectively) (Figure 1A), similar to the majority of intrinsically unstructured proteins.21 We also observed a high net charge and low mean hydropathy for Ki1/57 sequence (Figure 1B). This represents another characteristic observed for disordered proteins.22 In general, the low content of hydrophobic amino acids residues flanked by numerous acidic residues inhibits the formation of a well defined hydrophobic core and disfavors the formation of secondary structure elements. This in turn results in an 4466

Journal of Proteome Research • Vol. 7, No. 10, 2008

Bressan et al. extended conformation due to electrostatic repulsion.23,24,20 These observations can also be attributed to Ki-1/57 and its construct Ki-1/57(122-413), which contain a high proportion of negatively charged amino acids in its sequence. Trying to discriminate unfolded from potentially ordered regions in the sequence of Ki-1/57, we used the predictors PONDR VL-XT25,26 and DISORPRED.27 Not surprisingly, both predictors showed long disordered regions throughout the entire Ki-1/57 sequence (Figure 1C, E). In the same direction, we basically saw similar results with the software PSIPRED,28 which predicts a low content of R-helices (∼16%) and β-sheets (∼2%) for Ki-1/57 sequence (Figure 1D). In combination, all of these different bioinformatics tools demonstrate that at the level of its primary structure, Ki-1/57 possesses the typical amino acid composition of intrinsically unstructured proteins.16,23 Due to the low bacterial expression, degradation and instability of the recombinant full-length Ki-1/57 protein, we focused our analysis on a construct spanning the C-terminal region of Ki-1/57: Ki-1/57(122-413). Small Angle X-Ray Scattering (SAXS) Studies. SAXS is a very useful technique for the study of flexible and low compactness proteins, providing important parameters like the overall size and shape of the macromolecules in solution.29-31 Figure 2A displays the corrected and normalized experimental SAXS curves for Ki-1/57(122-413), together with the curve fitting calculated using GNOM (solid lines). The p(r) function resulting from this calculation is shown in the inset of Figure 2A. The maximum dimension obtained for the particle was 150 Å. The Rg values derived from Guinier’s law, Debye’s law, and the p(r) function were respectively (47 ( 2) Å, (48.1 ( 0.4) Å and (47.0 ( 0.6) Å, all values being considered in close agreement. In comparison with the BSA scattering data, the molecular mass for Ki-1/57(122-413) was estimated from the SAXS data as being ∼40 kDa. This value is in agreement with the 37 kDa predicted from ProtParam32 and indicates that the particle is monomeric in solution. These overall parameters suggest an elongated shape for Ki-1/57(122-413) which is also evidenced by the corresponding asymmetric shape of the p(r) function (inset of Figure 2A). The Kratky plot [q2I(q) versus q] is not a bell shaped curve, with a well defined maximum usually obtained for compact and globular particles. Instead, a plateau for q > 0.07 Å-1 is observed (Figure 2B). The absence of a maximum in the plot suggests that Ki-1/57(122-413) possesses flexible chains and may not have a well-packed core. These results support the fact that Ki-1/57 represents an unfolded or at least partially unfolded molecule in solution. In the inset of Figure 2B we can observe that the intensity curve behavior in the region 0.07 < q < 0.2 Å-1 has a q-2 behavior, as expected for a typical Gaussian chain model.33-35 Ab Initio Approaches. By using two ab initio approaches provided by the DAMMIN and GASBOR programs described in the section Experimental Procedures, overall molecular envelopes were obtained for the Ki-1/57(122-413) protein using the scattering data. Keeping in mind that the modeling procedure does not produce a unique solution,36-38 ten independent runs were performed for each approach. Different conformations were observed for each run. However, all of the models had the typical, recurring extended shape. Figure 2C (light blue model) displays the filtered average DAMMIN model obtained for Ki-1/57(122-413). Five representative GASBOR dummy-residue models are displayed alongside the DAMMIN average model (Figure 2C, green models). The normalized spatial discrepancies (NSDs)39 for the set of DAMMIN models

Intrinsically Unstructured Ki-1/57

research articles

Figure 1. Bioinformatical analysis of the amino acid sequence of Ki-1/57. (A) Frequencies of order-promoting (W, C, F, I, Y, V, L and N) or disorder-promoting residues (A, R, G, Q, S, P, E and K) given for Ki-1/57 (full-length), Ki-1/57(122-413) and for globular proteins.16 (B) Net charge versus hydropathy plot for Ki-1/57 (0). Values for a number of known natively unfolded proteins (gray circles) and folded proteins (black squares) analyzed by Uversky et al.22 are also shown. (C, D, E) Schematic view of the analysis of Ki-1/57 primary amino acid sequence with DISORPRED, PSIPRED, and PONDR VL-XT, respectively. Above the score threshold of 0.05 and 0.5, the amino acid residues are considered disordered for the DISORPRED and PONDR predictors, respectively. (D) Cylinders and arrows indicate the prediction of R-helices and β-strands respectively (PSIPRED).

attained values from 0.88 to 1.01 (considered acceptable for this modeling approach). However, since the protein seems to be unfolded and highly hydrated, the excluded volume calculated for this model is overestimated (1.38 × 105 Å3). The low resolution molecular reconstruction of this representation only shows that the conformation of this protein is extended in solution, giving few structural details. When using the dummy residues approach (GASBOR program), the set of models obtained had higher NSD values, varying from 2.32 to 2.43, but this type of reconstruction gives a more realistic image of the conformational differences existing in partially unfolded proteins in solution. An average conformation cannot be calculated for these higher resolution models, but the GASBOR approach gives an idea of the conformational space occupied by the protein more adequately. The resolution of the models does not permit an unambiguous determination of the spatial positions of secondary structure elements, but they portray the overall shape of the most frequent conformations adopted by the molecules in solution. The models obtained for the Ki-1/ 57(122-413) protein appear to explore a large conformational space in solution. These results follow the lines of the general features obtained for other intrinsically unstructured proteins including ab initio SAXS models available in the literature.40,30 Taken as a whole, Kratky plots and low resolution models calculated from SAXS data for Ki-1/57(122-413) reveal that it has the typical characteristics of proteins that lack a globular structure.

Hydrodynamic Behavior of Ki-1/57(122-413). Aiming to study the overall shape of Ki-1/57(122-413) through its hydrodynamic properties, we used the analytical size-exclusion chromatography and analytical ultracentrifugation techniques. When size exclusion chromatography is used for globular proteins, an apparent molecular mass can be obtained from the linear relationship between the molecular mass of standard proteins, which are approximately globular in shape. However, due to their elongated shape, intrinsically unstructured proteins normally display a discrepant larger molecular mass when eluted from a gel filtration column.41 Based to the elution volume observed for Ki-1/57(122-413), we calculated an apparent molecular mass of around 101 kDa, larger than the expected 37 kDa, theoretically predicted by ProtParam (Figure 3A). The calculated hydrodynamic radius (Stokes radius - Rs), 41 Å, was also larger than expected for a “globular” protein of 37 kDa and indicates that Ki-1/57(122-413) is a typical native premolten globule under the studied conditions. This conclusion could be achieved through the described log(Rs) versus log(MW) dependencies obtained for different conformations of globular proteins.23 According to a set of published equations,23 the expected Rs for a 37 kDa protein in its ordered, molten globular, premolten globular and coil-like states is around 26.7, 29.7, 40.0, and 50.3 Å, respectively. Therefore, the Rs experimentally determined for Ki-1/57(122-413) (41 Å) is in excellent agreement to that of a premolten globule of 37 kDa. Journal of Proteome Research • Vol. 7, No. 10, 2008 4467

research articles

Bressan et al.

Figure 2. Experimental SAXS curves and ab initio modeling for 6×His-Ki-1/57(122-413) in solution. (A) Experimental scattering curve (O) and theoretical fitting (solid line) of the data using the program GNOM. Inset: Pair distance distribution functions p(r). (B) Kratky Plot. (Inset) log[I(q)] vs log[q] and the curve behavior as 1/q2. (C) Low resolution average DAMMIN model (light blue) followed by five representative GASBOR models (green) alongside. The five GASBOR models displayed were chosen from several independent calculations based on the lowest NSD values. The models were displayed using the program PyMOL.64

As the large dimension observed by analytical gel filtration could be also due to an oligomerization of the protein under experimental conditions, we subsequently used analytical ultracentrifugation, to investigate the exact oligomeric state of Ki-1/57(122-413). The sedimentation distribution shown on Figure 3B is characteristic of a single species protein at the concentrations tested. To avoid interferences caused by solution viscosity and molecular crowding,42 we determined the standard sedimentation coefficient at 0 mg/mL s020,w extrapolated from the determined standard sedimentation coefficients (s20,w) at each analyzed protein concentrations (Figure 3B, inset). The value of s020,w was 2.20 ( 0.04 S and D was estimated to be 4.7 ( 0.1 × 10-7 cm2/seg through dynamic light scattering experiments (at standard conditions). The estimated frictional ratio (f/f0) was of around 1.9 ( 0.2 evidencing Ki-1/57(122-413) as an asymmetric particle. The f/f0 value indicates how a particle differs, in asymmetry, when compared to a sphere representing a globular protein of same M.42,43 This estimated frictional ratio f/f0 is too large to be explained only by the hydration contribution of the protein related to the model sphere used to calculate f0. Through the ratio of sedimentation to diffusion coefficient (eq 3) or by the peak of the c(M) analysis (see Methods) we showed that the protein is monomeric in solution with a molecular mass M of 39 ( 1 kDa or 40 ( 2 kDa, respectively. 4468

Journal of Proteome Research • Vol. 7, No. 10, 2008

The M determined for Ki-1/57(122-413) is in agreement with the SAXS results and with the predicted 37 kDa for the monomer. It is interesting to notice that studying the full-length protein; we observed previously no homodimerization in a yeast two-hybrid assay, where we tested Ki-1/57 fused separately to both the lexA and the GAL4 activation domains (data not shown). This finding suggests that the full-length Ki-1/57 is monomeric in an in vivo system, therefore, in agreement to the monomeric state observed for the C-terminal construct Ki1/57(122-413), here, in vitro. Proteinase K sensitivity assays. Due to their intrinsic flexibility and to the absence of well defined hydrophobic core, intrinsically unstructured proteins are more easily hydrolyzed than globular proteins when incubated with low concentrations of proteases.14,44 To investigate this behavior, we treated the recombinant construct Ki-1/57(122-413) with limiting concentration of proteinase K (Figure 4A). Degradations products appear as early as 2 min after incubation with the protease and none of them accumulate in time. This indicates that the degradation products are being further processed and degraded by proteinase K and suggests that the entire fragment Ki-1/ 57(122-413) is flexible and may not represent a stable hydrophobic core. As a control, the globular protein BSA was also submitted to identical conditions but no sign of proteolysis was

research articles

Intrinsically Unstructured Ki-1/57

Figure 3. Hydrodynamic characterization of 6×His-Ki-1/ 57(122-413). (A) Analytical size exclusion chromatography. In set: Calibration curve using the molecular size standards proteins: aldolase (158 kDa), BSA (67 kDa), ovalbumin (43 kDa) and ribonuclease A (13,7 kDa). The arrows indicate the elution volumes of the standards proteins. (B) Sedimentation velocity experiments. The figure displays c(s) distributions obtained with 0.6, 0.80, and 1.0 mg/mL of the Ki-1/57(122-413) performed at 20 °C with the rotor AN-60Ti at 40.000 rpm. The c(s) distributions were fitted using the SedFit software and the maximum of Gaussian curves resulted in the apparent s. (Inset) Plots of s20,w versus protein concentrations was fitted by linear regression to calculate the s020,w.

observed (Figure 4A). This indicates that the observed degradation of the unfolded Ki-1/57(122-413) was a protein specific effect. It is interesting to notice that the initial perception of this high proteolytical sensitivity was possible in the protein purification procedures, which required the use of additional strategies during all the steps of production of the recombinant Ki-1/57(122-413) due to the frequent signs of degradation. Considering that disordered proteins could adopt a more ordered structure upon interaction with their molecular partners,23,14 we investigated the proteinase K degradation of the endogenous Ki-1/57 in its cellular context. To detect the degradation state of the proteins in the cellular lysates we used antibodies against Ki-1/57 and GAPDH, used here as an endogenous control for globular proteins (Figure 4B). In comparison with the control GAPDH, the endogenous Ki-1/57 was also sensitive to protease digestion, indicating that it has flexible and disordered characteristics even in the cell, where it is probably involved in larger complexes or associated with interacting proteins such as RACK1.11 Although this is an indirect approach, it is a useful way to discriminate folded/

unfolded states of proteins in the cellular context and was successfully applied for globular and unfolded proteins from the intact nuclear pore complex in purified yeast nuclei.44 Taken together, the observed susceptibility to limited proteolysis demonstrated that both recombinant Ki-1/57(122-413) and endogenous full length Ki-1/57 from its native context are flexible, thereby supporting the hypothesis that Ki-1/57 is an intrinsically unstructured protein. Secondary Structure Analysis. The CD spectrum of Ki-1/ 57(122-413) exhibited predominantly a negative minimum at ∼200 nm, indicating the absence of regular secondary structure (Figure 5A). This is a typical signature of random coil or denatured proteins and confirms our previous observations.12 It has however been reported, that most of the intrinsically unstructured proteins can acquire folded structures upon binding to their biological partners, including proteins or nucleic acids.24 The protein CGI-55, which is also called PAIRB1 (Plasminogen activator inhibitor-RNA binding protein 1), shares 40% amino acid sequence identity and 67% similarity with Ki-1/57, and has been reported as its candidate paralogue.7 PAI-RB1 has been described to bind to RNA in vitro.45 In gelshift experiments we therefore tested if Ki-1/57 can also bind to RNA in vitro (data not shown). We found that Ki-1/57 binds selectively to a 25-mer poly-U RNA but not to A, C, or G homopolymers, in vitro. Based on these studies, we therefore tested if Ki-1/57(122-413) upon RNA binding undergoes any unfolded-to-folded transition. However, we were not able to detect any change in the secondary structure of Ki-1/57(122-413) in circular dichroism spectroscopy analyses (Figure 5A). On the other hand, the functional binding of Ki-1/57(122-413) to the poly-U RNA in vitro, as detected by the mobility shift assay technique (Figure 5B), demonstrated the functionality of the protein construct in vitro. This suggests that the RNA interaction by itself is not sufficient to change the overall secondary structure of Ki-1/ 57(122-413). This may point to the necessity of a possible third element in the interaction complex to promote such a structural transition. Another widely used way to investigate the propensity of an amino acids sequence to undergo induced folding is the use of TFE as a molecular probe. It has been reported that it increases the propensity of amino acid sequences to form regular secondary structure by reinforcing hydrogen bonds within the polypeptide backbone in TFE/H2O mixtures and through favorable interaction of hydrophobic amino acid side chains with TFE relative to H2O.46 Upon addition of increasing amounts of TFE, we observed modifications of the CD pattern of Ki-1/57(122-413), suggesting an increasing in the content of regular structures (Figure 5C). This is a typical behavior of intrinsically unstructured proteins and thereby points that Ki1/57(122-413) has propensity to form secondary structure. This transition toward regular structures is particularly common in binding site regions that undergo disorder-to-order transition upon protein-protein complex formation.47

Conclusions In this work we provide new information concerning the conformational properties of the C-terminal region of Ki-1/57 in solution. The extended conformational space determined for Ki-1/57(122-413), together with its secondary structure properties and proteolytical degradation susceptibility, lead us to conclude that it is intrinsically disordered. This therefore supports the hypothesis that human Ki-1/57 is an intrinsically Journal of Proteome Research • Vol. 7, No. 10, 2008 4469

research articles

Bressan et al.

Figure 4. Limited proteolysis analysis of the recombinant protein 6×His-Ki-1/57(122-413) and the endogenous cellular Ki-1/57. (A) SDS-PAGE of aliquots from samples of 6×His-Ki-1/57(122-413) or the control protein BSA after incubation with proteinase K at 37 °C during the indicated times. (B) Proteinase K susceptibility analysis of the endogenous Ki-1/57 or the control globular protein GAPDH from lysates of human embryonic kidney cells (HEK 293).

unstructured protein. Several reviews have been written to discuss the functions and properties of natively unfolded proteins.13,14,21-24 Intrinsic disorder is mostly correlated with membrane transport, molecular recognition, post-translational modifications, cellular signaling and regulatory mechanisms such as transcription, translation and the cell-cycle checkpoints. It is believed that these functions cannot be carried out by a rigid structure. Instead, it is associated with the freedom of the polypeptide chain to fluctuate among alternatives states, conferring the ability to bind to multiple different targets without lost of specificity.18,20 Protein-protein interaction networks are characterized by the presence of nodes, called hubs, which connect multiple biological responses throughout the interaction with multiple partners.48 These interactions may occur dynamically at different times and locations, connecting various processes in the network.49 Interestingly, these hub proteins possess the structural tendency in having long disordered regions, loops and high surface charge as compared with nonhub proteins.50 To date, the functional data available suggest a complex role for Ki-1/57 in gene expression regulation. Similarly to known hub proteins, Ki-1/57 has the ability to bind to several partners involved in cellular signaling, transcriptional regulation and RNA metabolism.9-11 Moreover, we discovered the specific binding of Ki-1/57 to U-rich RNA sequences (Figure 5A). Therefore, these observations fit very well with the characteristics of the intrinsically unstructured proteins and suggest the necessity of a very flexible structure for Ki-1/57 to allow its functional plasticity. Long disordered regions have been found in several human disease associated proteins. These include diabetes, autoimmune disease, neurodegenerative disease, cardiovascular disease and cancer18,19 and reflect the functional importance of the class of intrinsically unstructured proteins. Recently, we 4470

Journal of Proteome Research • Vol. 7, No. 10, 2008

demonstrated the functional association and binding of Ki-1/ 57 to the suppressor tumor protein p53,9 suggesting the involvement of this interaction in the cellular growth control. This points to the importance of a better understanding of the structural characteristics of Ki-1/57, as we presented here, as well as the details of its interaction and functional relationship with this tumor suppressor.

Experimental Procedures Amino Acid Sequence Analysis. The percentages of order/ disorder promoting residues for the amino acids sequence of full-length Ki-1/57 (1-413) (accession number NP _ 055097) were determined based on Tompa and co-workers.21 The program PONDR, Prediction of Natural Disordered Regions, http:// www.pondr.com (accessed July 11, 2008) was used to obtain the hydrophobicity/mean net charge plot as well as to determine the disordered regions in the Ki-1/57 sequence using the default predictor VL-XT.25,26 Access to PONDR is provided by Molecular Kinetics Indianapolis, http://www.molecularkinetics. com (accessed July 11, 2008). Further disorder predictions, were performed using DISORPRED, http://bioinf.cs.ucl.ac.uk/disopred (accessed July 11, 2008).27 Analysis of the secondary structural elements was done with PSIPRED, http://bioinf.cs. ucl.ac.uk/psipred/ (accessed July 11, 2008).28 Plasmid Constructions, Protein Expression, and Purification. Cloning of the cDNA coding for Ki-1/57(122-413) into the pPROEx plasmid (Invitrogen) was performed as described previously.12 The pPROEX-Ki-1/57(122-413) recombinant vector expresses a construct containing 29 additional amino acids at the N-terminal of Ki-1/57(122-413). These extra residues encopass the 6×His-tag and other poly-linker encoded amino acids. Induction of protein expression by IPTG resulted in the

Intrinsically Unstructured Ki-1/57

research articles

Figure 5. Circular dichroism spectroscopy analysis and RNA binding activity of 6×His-Ki-1/57(122-413). (A) Spectra obtained after incubation with the ligand RNA poly-U for 1 h on ice (RNA/protein molar proportion 2:1). (B) RNA binding activity analysis of Ki-1/ 57(122-413) with 32P-labeled 25-mer poly-U (electrophoretic mobility shift assay, EMSA) after 20 min of incubation on ice. The poly-U RNA probe was incubated with crescent amounts of Ki-1/57(122-413) [0.5-1.0-2.0 µM] or GST [4.0 µM], used as negative control. The RNA-protein complexes were run out on native polyacrylamide gels (10%) and visualized with a phosphoimager system (Fuji). (C) Spectra obtained in the presence of increasing concentration of TFE. All far-UV Circular dichroism spectra were acquired at 20 °C and in phosphate buffer 20 mM, pH 7.5 with 20 mM NaCl.

expression of a 6×His N-terminally tagged protein Ki-1/ 57(122-413) for brevity also simply called Ki-1/57(122-413) here. To obtain protein preparations free of degradation and up to 95% of purity, Ki-1/57(122-413) was purified in two steps as follows. After harvesting (6000× g, 10 min), the bacterial cells were resuspended and incubated for 40 min at 4 °C in lysis buffer: 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 10% glycerol, 1 mg/ml lysozyme, 1 mM phenylmethylsulfonyl fluoride, and 0.05 mg/ml DNase. Additional pellet disruption was performed by 10 cycles of sonication in an ice bath, followed by centrifugation at 18 000× g, 4 °C for 40 min. The obtained supernatant was loaded onto a HiTrap Chelating HP column (Ge Healthcare) and eluted by a gradient of 0-500 mM imidazole. The obtained Ni2+-affinity purified fractions were pooled, dialyzed against the buffer: 50 mM Tris-HCl (pH 7.5), 100 mM NaCl, 10% glycerol, 5 mM EDTA. Next, samples were loaded onto a Q-Sepharose ion-exchange column (GE Healthcare). Proteins were eluted by a gradient of 0-1 M NaCl. All the procedures

were performed in the presence of protease inhibitors (Calbiochem) at 4 °C to avoid protein degradation. SAXS Experiments. The small-angle X-ray scattering experiments were performed at the D02A-SAXS2 beam line of the Laborato´rio Nacional de Luz Sı´ncrotron (LNLS, Campinas-SP, Brazil). Before the analysis, the samples of 6×His-tagged Ki1/57(122-413) in buffer 50 mM Tris-HCl (pH 7.5), 400 mM NaCl, 15% glycerol, were centrifuged at 356 000× g for 30 min at 4 °C to remove any possible aggregates or particles. The temperature-controlled (T ) 20 °C via water circulation) measurements were performed in a 1-mm path length cell with mica windows51 with a monochromatic X-ray beam (wavelength of λ ) 1.488 Å). The X-ray patterns were recorded using a two-dimensional position-sensitive MARCCD detector. The sample-to-detector distance was 1374.4 mm, corresponding to the scattering vector range of 0.01< q