Conformational Role for the C-Terminal Tail of the Intrinsically

TECHNICAL NOTE pubs.acs.org/jpr

Conformational Role for the C-Terminal Tail of the Intrinsically Disordered High Mobility Group A (HMGA) Chromatin Factors Elisa Maurizio,† Laetitia Cravello,‡ Liam Brady,‡ Barbara Spolaore,§ Laura Arnoldo,† Vincenzo Giancotti,† Guidalberto Manfioletti,*,† and Riccardo Sgarra*,† †

Department of Life Sciences, University of Trieste, Trieste, Italy Waters Corporation, Atlas Park, Manchester, United Kingdom § CRIBI Biotechnology Centre, University of Padua, Padua, Italy ‡

bS Supporting Information ABSTRACT: The architectural factors HMGA are highly connected hubs in the chromatin network and affect key cellular functions. HMGA have a causal involvement in cancer development; in fact, truncated or chimeric HMGA forms, resulting from chromosomal rearrangements, lack the constitutively phosphorylated acidic C-terminal tail and display increased oncogenic potential, suggesting a functional role for this domain. HMGA belong to the intrinsically disordered protein category, and this prevents the use of classical approaches to obtain structural data. Therefore, we combined limited proteolysis, ion mobility separation-mass spectrometry (IMS-MS), and electrospray ionization
mass spectrometry (ESI
MS) to obtain structural information regarding full length and C-terminal truncated HMGA forms. Limited proteolysis indicates that HMGA acidic tail shields the inner portions of the protein. IMS-MS and ESI
MS show that HMGA proteins can assume a compact form and that the degree of compactness is dependent upon the presence of the acidic tail and its constitutive phosphorylations. Moreover, we demonstrate that C-terminal truncated forms and wild type proteins are post-translationally modified in a different manner. Therefore, we propose that the acidic tail and its phosphorylation could affect HMGA post-translational modification status and likely their activity. Finally, the mass spectrometrybased approach adopted here proves to be a valuable new tool to obtain structural data regarding intrinsically disordered proteins. KEYWORDS: intrinsically disordered proteins, Ion Mobility Separation
Mass Spectrometry, limited proteolysis, oncogene, post-translational modifications, phosphorylation

1. INTRODUCTION The High Mobility Group A (HMGA) proteins are chromatin factors, which together with the HMGB and HMGN proteins belong to the HMG superfamily.1 Besides a role in modulating chromatin structure, the three main members of the HMGA family (HMGA1a, HMGA1b, and HMGA2) are functionally involved in a plethora of biological processes ranging from transcriptional regulation to chromatin remodeling.2 They are proven to be essential for proper embryonic development and are implicated, through different mechanisms, in the development of both benign and malignant neoplasias, which is why they are referred to as oncofetal proteins.3 For instance, HMGA2 protein is a key regulator of the epithelial-mesenchymal transition (EMT) through the transcriptional control of SNAIL14 and promotes cellular proliferation regulating cyclin A1 expression,5 while HMGA1 prevents p53 activation in tumor cells by direct protein
protein interaction.6,7 Moreover, HMGA are causally linked to neoplastic transformation and considered a hallmark of almost all human tumors.3 Notably, chromosomal rearrangements of HMGA genes, leading to the production of C-terminal truncated or chimeric proteins, are a feature of most benign r 2011 American Chemical Society

human mesenchymal tumors.3 The relevant impact of HMGA on cell biology resides also in the fact that they are highly connected proteins. Indeed, their molecular network consists of more than 100 different factors.2 Expression of HMGA genes is under strict transcriptional and post-transcriptional control,8 and DNA and protein contacts of HMGA proteins are both finely modulated by post-translational modifications (PTMs).9,10 HMGA are small proteins of about 100 residues, with a modular domain organization characterized by three DNA binding domains (DBDs), a protein
protein interaction domain, which partially overlaps with the second DBD, and a highly acidic C-terminal tail that is constitutively phosphorylated.1 Notably, despite any clear functional assignment, the importance of the C-terminal tail is coming to light. The loss of the C-terminal domain has been demonstrated to have an impact on cellular transformation, suggesting an inhibitory role in the regulation of HMGA proliferative and transforming ability. In fact, HMGA truncation confers an increased growth rate to different cell types,11 and HMGA deleted forms display a higher Received: February 11, 2011 Published: May 06, 2011 3283

dx.doi.org/10.1021/pr200116w | J. Proteome Res. 2011, 10, 3283–3291

Journal of Proteome Research oncogenic activity.12,13 This suggests that the lack of the C-terminal tail could enhance HMGA oncogenic activity through changes leading to alterations of their interactions with DNA, protein partners, and modifying enzymes. In addition, the C-terminal tail of HMGA2 has been demonstrated to have a transactivating function in the context of Snail1 gene regulatory sequences14 and cells expressing C-terminal tail truncated HMGA2 have a different transcriptional profile with respect to those expressing full-length protein.15 For a large number of proteins, conventional techniques cannot provide specific structural assignments. These molecules belong to the peculiar category of the intrinsically disordered (ID) proteins,16 for which the structure/function paradigm does not apply.16 Their ID status confers a peculiar plasticity thus enabling a high connectivity; this is the reason why they are often “hubs” of protein molecular networks.17 HMGA factors are considered to be prototypical ID proteins.2 In fact, the only structural information regarding HMGA concerns the region comprising the second and the third AT-hook that, by means of multidimensional NMR spectroscopy, has been demonstrated to undergo a disordered-to-ordered transition upon DNA binding.18 Electrospray ionization
mass spectrometry (ESI
MS), alone or in combination with limited proteolysis, has been employed for protein conformational investigations.19,20 Moreover, the combination of gas-phase ion mobility spectrometry with mass spectrometry led to the development of ion mobility separation/spectroscopy
mass spectrometry (IMS-MS) that, in the last 10 years, has proved invaluable for probing protein structure.21 By means of these techniques, we aimed to probe the conformational equilibria of HMGA proteins, which are largely unknown but could help to characterize the mechanisms by which they contribute to carcinogenesis. Moreover, HMGA are affected by a multiplicity of PTMs.9 PTMs are known to modulate protein functions by creating or disrupting binding sites, altering local structural conformations or posing barriers to conformational transitions. Therefore, in this work we also explored a conformational role for HMGA PTMs, in particular for constitutive phosphorylation of the HMGA C-terminal tail. Limited proteolysis highlighted a shielding effect of the tail toward the inner portion of the proteins. IMS-MS and ESI
MS showed that HMGA, despite their ID status, can assume a compact conformation and that the C-terminal tail and its constitutive phosphorylations are responsible for a modulation of protein compactness. The observed different susceptibility of truncated forms with respect to full-length proteins toward modifying enzymes could therefore be attributed to these structural differences. This evidence, together with the already known in vivo functional outcome of HMGA C-terminal truncation,11
13 suggests a structure/function link between HMGA C-terminal tails, their PTMs, and the proteins oncogenic properties, paving the way for the development of interfering therapeutic strategies based on targeting HMGA proteins.22

2. MATERIALS AND METHODS 2.1. Recombinant Protein Production and Purification

Recombinant HMGA, GST-PRMT1 and GST-PRMT6 proteins were produced and purified as already described.23
25 In the work they are named as follow: HMGA1a full-length: HMGA1a FL; HMGA2 full-length: HMGA2 FL; HMGA1a 1
89: HMGA1a CT; HMGA2 1
93: HMGA2 CT; HMGA2 tetra- and penta-phosphorylated: HMGA2 4P and 5P, respectively.

TECHNICAL NOTE

2.2. Methylation and Phosphorylation of HMGA Proteins

C-terminal phosphorylated HMGA2 was obtained and purified as previously described.9,23 In vitro methylation of FL and CT HMGA1a by GST-PRMT1 and GST-PRMT6 both with [3H]labeled and nonlabeled S-adenosyl-L-methionine (SAM), electrophoresis, and fluorography of methylation reactions were carried out as previously described.24,25 Briefly, equimolar quantities (216 pmoles) of FL and CT HMGA1a were incubated in the same batch with GST-PRMT1, GST-PRMT6, or GST (5 μg) in 30 μL of PBS in the presence of 4 μL [3H]-SAM (0.55 mCi/mL
Perkin-Elmer). Methylation reactions for LC
MS analyses were carried out incubating FL and CT HMGA1a (432 pmoles) in the same batch with GST-PRMT1, GST-PRMT6, or GST (2 μg) in 30 μL of PBS in the presence of 1 μL SAM 15 mM at 30 °C for 16 h. In vitro phosphorylation of HMGA proteins by CDK1 were performed by incubating equimolar quantities (650 pmoles) of FL and CT recombinant HMGA1a or HMGA2 proteins in the same batch with 20 Units of CDK1 (New England BioLabs) in 30 μL of reaction volume (50 mM Tris/HCl, 10 mM MgCl2, 2 mM DTT, 1 mM EGTA, 0.01% Brij 35, pH 7.5, and ATP 200 μM) at 30 °C. Phosphorylation assays were stopped by trifluoroacetic acid (TFA) addition (1% final) at 0.5, 1, 2.5, 5 and 16 h. HMGA proteins after methylation and phosphorylation assays were made 0.1% TFA and LC-MS analyzed with a 1200 Series Capillary HPLC (Agilent) coupled to an HCTultra ion trap (Bruker Daltonics). Proteins were separated on a Zorbax 300SBC18 3.5 μm, 150  0.5 mm column (Agilent) using a water 0.05% TFA - (A)/acetonitrile 0.05% TFA - (B) gradient. 2.3. Limited Proteolysis Analyses

FL and CT HMGA2 forms were subjected to limited proteolysis using trypsin (Promega) in 100 mM ammonium acetate pH 7.0 at an E:S ratio of 1:1000 (w/w) directly in nano electrospray capillaries. Limited proteolysis using Lys-C (Roche) were performed in 80 mM ammonium bicarbonate pH 8.1 at an E:S ratio of 1:5000 (w/w) at 25 °C and stopped by acetic acid addition (5% final) after 5 and 15 min. ESI
MS measurements of limited proteolysis with trypsin and Lys-C were performed on a Q-TOF Micro mass spectrometer (Micromass, U.K.) equipped with a Z-spray nanoflow electrospray ionization interface. Mass spectra of the peptide digests were acquired using the nano electrospray source operating at capillary, cone, and extractor voltages of 1800, 35, and 1 V, respectively (positive ion mode). Immediately after the addition of trypsin, HMGA digestion solutions were loaded in capillaries to directly follow the ongoing of the proteolytic cleavage. Instrument control and data acquisition and processing were performed with the MassLynx software (Micromass, U.K.). Limited proteolysis with Lys-C endoprotease were also analyzed by LC
MS as described above. 2.4. IMS-MS Analyses

HMGA2 proteins were diluted in 100 mM ammonium acetate. Samples were infused into the ESI source of a SYNAPT HDMS (Waters Corp., Manchester, U.K.) at a flow rate of 10 μL/min. An ESI capillary voltage of 3.2 kV was used, together with a sampling cone voltage of 30 V. The desolvation gas flow and temperature were held at 800 L/h and 300 °C respectively. The source backing pressure was maintained at 5.2 mbar to produce intact gas phase ions from the protein in solution. The pressure in the Trap and Transfer T-Wave regions was about 0.02 mbar of argon and the pressure in the IMS T-Wave was 0.47 mbar of nitrogen. IMS traveling wave velocities of 300 m/s were used, and the wave pulse height was 6.5 V. Instrument 3284

dx.doi.org/10.1021/pr200116w |J. Proteome Res. 2011, 10, 3283–3291

Journal of Proteome Research

TECHNICAL NOTE

HMGB, HMGA and HMGN sequences are predicted to be totally disordered. However, a peculiar characteristic common only to HMGA and HMGB is the presence of a highly negative C-terminal tail (Figure 1A, C) that in both cases is predicted to be ID (Figure 1B). The HMGB1 acidic tail can intramolecularly shield the positively charged HMGB DBDs modulating HMGB1-DNA interactions and impairing accessibility to modifying enzymes.27
31 Similarly, HMGA proteins have highly positively charged DBDs and an acidic C-terminal tail whose negative charge density is comparable or even higher, being enhanced by constitutive phosphorylation (Figure 1C). This evidence suggests a role for the HMGA acidic domain similar to that of HMGB, despite the structural differences between the two proteins. 3.1. C-Terminal Tail of HMGA Proteins Shields the Protein
Protein Interaction Domain

Figure 1. Prediction of intrinsically disordered (ID) status of HMG proteins and comparison between HMGA and HMGB1 acidic tails. (A) Schematic representation of HMG proteins (HMGA1a, HMGA2, HMGB1, and HMGN1) and their functional domains. (B) Prediction of ID regions in HMGA1a, HMGA2, HMGB1, and HMGN1 by PONDR (Predictor Of Natural Disordered Regions
Molecular Kinetics, Inc.).26 Scores greater than 0.5 indicate propensity to disorder. (C) Amino acid sequences of C-terminal tails of HMGB1, HMGA1a, and HMGA2 are shown. Phosphorylable residues are showed in red. For each tail, the sum of negatively charged amino acid is indicated (z) together with charge density (z density: no. negative charge/no. aa residues). Phosphate group charges have been estimated using pKa1 = 1.2 and pKa2 = 6.539 and physiological pH 7.4.

control and data analysis were performed using MassLynx v.4.1 software (Waters Corp., Manchester, U.K.).

3. RESULTS AND DISCUSSION Truncated HMGA proteins are known to have enhanced oncogenic properties with respect to full-length forms.11
13 We asked ourselves whether this difference is linked to variation in structural conformations. Since no specific structural information is available for HMGA1 and HMGA2, we started by comparing the already known structural properties of the different proteins belonging to the HMG superfamily. Therefore, we performed a comparison of HMG’s structural domain organization and a bioinformatic prediction of disordered sequences in HMG proteins with the Predictor Of Naturally Disordered Regions (PONDR) software26 (Figure 1A, B). In contrast to

The intrinsic disorder of HMGA proteins prevents the evaluation of any kind of intramolecular interaction by conventional strategies and no data are currently available as to whether they assume any particular conformation.2 As expected, the farUV circular dichroism (CD) spectra of full-length (FL) and C-terminal truncated (CT) HMGA forms did not show significant differences, both being characterized by an intense negative minimum centered at 198 nm, which is typical of a random coil conformation (Supplementary Figure S1, Supporting Information). Therefore, as a first step in evaluating a conformational role for the C-terminal tail, we investigated the structural properties of HMGA by limited proteolysis, focusing on HMGA2 as representative of all three HMGA proteins. Figure 2 and Supplementary Figure S2 (Supporting Information) show the results of limited proteolysis experiments performed on both FL and CT HMGA2. Both FL and CT HMGA2 forms are very rapidly cleaved by trypsin at R26, R28, R32, R48, and R76 (Supplementary Figure S2). Surprisingly, not all potential trypsin proteolytic sites are randomly cleaved, as would be expected for a fully accessible, flexible, and unstructured protein.32,33 This implies that both HMGA2 forms acquire some bended and hence “protected” conformation. The first sites of cleavage within both FL and CT HMGA2 are located in their DBDs suggesting that these domains have a higher accessibility with respect to the other portions of the protein. On the other hand, the majority of Lys residues is initially resistant to trypsin hydrolysis (Supplementary Figure S2). The most informative data in proteolysis experiments are obtained by the detection of the very first cleavage events,32,33 since those occurring later involve already digested protein portions. For this reason, the fact that both FL and CT HMGA proteins are cleaved very rapidly by trypsin at the level of Arg residues within their DBDs suggests that they are not fully extended proteins but prevents the evaluation of any difference between FL and CT forms. We therefore performed limited proteolysis reactions using endoprotease Lys-C with equimolar amounts of both FL and CT HMGA2 (Figure 2A, B). In Figure 2B, we report the overlaid chromatograms obtained by injecting both FL and CT Lys-C digestions (5 min) and in Figure 2C and D the LC
MS mass spectra and a summary of the identified peptides are shown, respectively. Comparing the UV chromatograms and the mass spectra (see the insert in Figure 2B and C) there is a clear difference between FL and CT for the abundance of peaks corresponding to peptides generated by cleavage at Lys residues embedded in HMGA protein
protein interaction domain 3285

dx.doi.org/10.1021/pr200116w |J. Proteome Res. 2011, 10, 3283–3291

Journal of Proteome Research

TECHNICAL NOTE

Figure 2. Differential accessibility to proteolytic cleavage of CT and FL HMGA2 forms. (A) Overlaid view of UV chromatograms (Abs 220 nm) relative to LC
MS separations of equivalent quantities (2 μg) of FL HMGA2 (red) and CT HMGA2 (blue) to assess the reliability of the protein quantification method adopted. (B) Overlaid view of UV chromatograms (Abs 220 nm) relative to LC
MS separations of limited proteolysis performed with Lys-C and equimolar quantities of FL (red) and CT (blue) HMGA2 (injected 200 pmoles
equivalent to 2 μg and 1.7 μg of FL and CT, respectively). (C) Mass spectra of Lys-C generated peptides eluting in the chromatographic region corresponding to the peaks shown in the insert of panel B. (D) Schematic representation of the LC
MS identified peptides. (Asterisks indicate m/z signals at 639.2 common to both peptides 1
61 and 1
55. The intensity of m/z 639.29þ belonging to peptide 1
55 has been arbitrary attributed considering the whole charge state distribution (CSD) of this peptide.)

(Figure 2D). The abundance of these peaks is higher in the CT form indicating that the presence of the C-terminal tail partially impairs the Lys-C endoprotease activity. This clearly suggests a shielding effect of the C-terminal tail toward the protein
protein interaction domain that could be elicited either directly by intramolecular interactions or indirectly by electrostatic repulsion. The same results have been obtained from the analysis of 15 min digestions (Supplementary Figure S3, Supporting Information), thus confirming the different accessibility of CT and FL proteins. Only N-terminal-containing peptides were evaluated to compare Lys-C cleavage rate between CT and FL forms since C-terminal-containing peptides are obviously different between the two protein forms and cannot be quantitatively compared both in UV and MS detection.

Therefore, by using limited proteolysis we overcame limitations of conventional strategies in obtaining structural data regarding ID proteins. The fact that both trypsin and Lys-C have preferential cleavage sites supports the idea that FL and CT HMGA forms do not assume fully extended and accessible conformations. Moreover, the different rates of Lys-C proteolysis clearly showed a difference in accessibility, suggesting a shielding effect of the C-terminal tail. 3.2. Conformational Role of the HMGA Acidic Tail

To confirm the masking effect of the acidic C-terminal tail we took advantage of the IMS-MS technique, which has recently emerged as a very powerful tool to probe for the presence of multiple conformational isoforms of proteins or macromolecular complexes.21 We analyzed FL, CT, and C-terminal phosphorylated 3286

dx.doi.org/10.1021/pr200116w |J. Proteome Res. 2011, 10, 3283–3291

Journal of Proteome Research

TECHNICAL NOTE

Figure 4. Loss of HMGA1a acidic tail increases PRMTs methylation. Reconstructed mass spectra, obtained by LC
MS analyses, of C-terminal tail truncated (CT) and full-length (FL) HMGA1a forms in vitro methylated by PRMT6 (A) or by PRMT1 (B). M indicates the addition of a methyl group.

Figure 3. IMS-MS analyses reveal a conformational role of HMGA acidic tail. (A, B, and C) DriftScope views of C-terminal truncated (CT), full-length (FL), and C-terminal tail phosphorylated (FL 4P and FL 5P) HMGA2, respectively. The drift time (abscissa) is plotted vs the m/z (ordinate); the abundances of ions are given by a color graduation starting from black (no signal), to blue, red, and yellow (high intensity). The 7þ charge states (signals corresponding to compact HMGA2 proteins chosen for ion mobility measurement shown in panels D
G) are showed in white. 8þ and 11þ/14þ signals correspond to the center of compact and extended HMGA2 m/z distributions, respectively. (D
G) Arrival time distribution (mobilogram) for 7þ ions belonging to CT, FL, FL 4P, and FL 5P HMGA2 forms. F, M, and S indicate ions with a fast, medium and slow mobility, respectively.

(FL 4P and FL 5P) HMGA2 forms. C-terminal tail phosphorylation of HMGA proteins is of particular interest since all HMGA molecules are constitutively phosphorylated in vivo there.9 Figure 3A
C show the DriftScope views and Figure 3D
G show the mobilograms of 7þ ions. DriftScope views show the existence of two charge state distributions (CSDs) for each protein, one centered at high m/z values (8þ), the other at lower ones (11þ or 14þ). In accordance to this, ESI
MS analyses, performed under near-physiologic conditions (Supplementary Figure S4, Supporting Information), in which FL HMGA2 is compared to Myoglobin and

Cytochrome C, also indicate that HMGA2 can assume both a lowand a high-charged CSD. Protein low-charged CSDs are representative of a compact/native conformation, while high-charged CSDs are usually attributed to extended/denatured conformations.34 Only ions carrying a low number of charges were chosen for the IMS-MS comparison (Figure 3D
G) since they belong to proteins in a “native-like” condition.34 In the mobilograms of 7þ ions (Figure 3D
G), all the proteins show two distinct ion mobility peaks, indicating that all of them are present in at least two different conformations. Proteins sharing the same charge and shape should have different drift time depending on their mass. Smaller ions move faster than bigger ones because of the higher acceleration they experience and the lower friction they encounter in the drift tube (Supplementary Figure S5A, Supporting Information). Therefore, assuming that CT and FL HMGA2 forms have the same shape, the lighter 7þ CT ions should have a lower drift time than 7þ FL ions. As shown in Figure 3D and E, this is not the case. CT HMGA2 ions are distributed between two distinct conformations, one of which with a drift time even higher than that of FL ions (S: slow mobility). This peculiar behavior indicates that the effect of shape is predominant with respect to the contribution of the mass of the protein. This clearly points out that the CT form assumes a conformation with a steric hindrance comparable or even higher to that of the longer FL (see Supplementary Figure S5B
D for a schematic representation of this phenomenon, Supporting Information). From these data we assume that the presence of the acidic C-terminal tail is responsible for further compacting the protein with respect to the CT form. As mentioned above, FL form (Figure 3E) is present in two distinct conformations (M: medium mobility and F: fast mobility). Considering FL 4P and FL 5P (Figure 3F, G), it is clear that the abundance of the F form increases with the presence and number of phosphate groups. These observations indicate that the acidic C-terminal tail allows HMGA2 to assume the more compact conformation (F) and that the equilibrium between F and M forms shifts toward the F form when the acidic 3287

dx.doi.org/10.1021/pr200116w |J. Proteome Res. 2011, 10, 3283–3291

Journal of Proteome Research

TECHNICAL NOTE

Figure 5. Loss of HMGA acidic tail increases CDK1 phosphorylation rate. Reconstructed mass spectra, obtained by LC
MS analyses, of time course CDK1 phosphorylation assays (0.5, 1, 2.5, 5, and 16 h) performed with C-terminal tail truncated forms (CT) and full-length (FL) HMGA proteins (HMGA1a and HMGA2, panels A, B, C, D, and E and F, G, H, I, and J, respectively). P indicates the addition of a phosphate group.

C-terminal tail is phosphorylated. Similar conclusions can be drawn looking at HMGA1a IMS-MS analyses (Supplementary Figure S6, Supporting Information). The shielding effect of the acidic tail on the protein
protein interaction domain highlighted by limited proteolysis experiments is thus strongly supported by IMS-MS analyses. 3.3. C-Terminal Truncation of HMGA Proteins Makes Them More Susceptible to PTMs

To provide a functional significance to the structural features showed above and considering that HMGA activities are finely modulated by several PTMs,10 we asked ourselves whether there is a structure/PTM relationship involving the acidic C-terminal tail. For this purpose, equimolar amounts of FL and CT HMGA proteins were incubated with different HMGA modifying enzymes (Protein Arginine Methyltransferase 1 and 6 (PRMT1 and PRMT6) and Cyclin Dependent Kinase 1 (CDK1)) in order to gain insights into the effect of the C-terminal tail of HMGA1a and HMGA2 on the accessibility of serine/threonine and arginine residues with respect to these modifying enzymes. Arginine methylation has not been observed in vivo for HMGA2 and

therefore methylation experiments were performed only with the HMGA1a protein. In fact PRMT1 and PRMT6 are known to methylate HMGA1a. On the other hand, phosphorylation experiments were performed with both HMGA1a and HMGA2 using CDK1, which has been described to in vivo phosphorylate both of them (see Supplementary Figure S7 for the various enzyme specificities toward HMGA proteins, Supporting Information).9,10 The methyl-transferase activity toward HMGA1a proteins has been checked by LC
MS analyses (Figure 4) and radiolabeling experiments (Supplementary Figure S8, Supporting Information). Time course experiments in combination with LC
MS detection have been performed to follow the kinetic of HMGA phosphorylation (Figure 5). Figure 4 shows the reconstructed mass spectra of FL and CT HMGA1a forms modified by PRMT6 and PRMT1. Both PRMTs are able to methylate CT HMGA1a much more efficiently (up to 4 methyl groups by PRMT6 and up to 8 methyl groups by PRMT1) with respect to FL HMGA1a (Figure 4A, B). These data were confirmed by parallel experiments performed with [3H]-SAM and visualized by fluorography (Supplementary Figure S8). Phosphorylation assays performed with CDK1 show that both CT 3288

dx.doi.org/10.1021/pr200116w |J. Proteome Res. 2011, 10, 3283–3291

Journal of Proteome Research

Figure 6. A model for HMGA conformational transitions. ESI-MS and limited proteolysis data suggest that HMGA proteins could be in equilibrium between a fully extended and a compact conformation (1). IMS-MS suggests that phosphorylation at the acid tail could enhance such compactness (2) and that truncated HMGA forms, on the contrary, could assume a more relaxed structure (3). These structural differences are responsible for a different accessibility to modifying enzymes (4) highlighting that there could be a structural/functional relationship at the basis of the enhanced transforming ability of truncated HMGA (5).

HMGA1a and CT HMGA2 forms are more rapidly phosphorylated than FL proteins. In fact, at each time point of the time course CT HMGA forms have a higher phosphorylation degree with respect to FL proteins (Figure 5). Given the different behavior of FL and CT HMGA proteins with respect to enzymatic activities it is possible to conclude that the HMGA C-terminal tail is involved in hampering the accessibility of the modifying enzymes to their specific consensus sites.

4. CONCLUSIONS Genomic mutations and rearrangements leading to altered protein products that are linked to evident phenotypic alterations are often essential clues to highlight protein functions.35 In fact, the cue for this work derives from the finding that cells expressing C-terminal tail truncated (CT) HMGA forms display a more aggressive neoplastic phenotype with respect to those expressing full-length (FL) HMGA.11
13 This clearly links the C-terminal tail to modulation of HMGA activities. Moreover, very recently different biological effects of CT HMGA1 with respect to FL form have been observed in transgenic mice.36 In addition, a gene-specific transactivating function has been assigned to the acidic tail of HMGA2,14 and gene expression analyses showed that overexpression of CT HMGA2 in mesenchymal stem-like cells affects the expression of a higher number of genes than FL HMGA2.15 All these observations suggest a functional and modulatory role of the C-terminal domain toward HMGA activities. Structural data are normally required to unravel the function of specific domains within a given protein, unveiling their threedimensional localization and relationship within the whole protein and suggesting possible molecular mechanisms. These approaches were not however suitable for HMGA given their ID status. These proteins are described as multifunctional, with no specific three-dimensional organization, and displaying two main biochemical properties: the ability of binding with a “relaxed” specificity to DNA and of interacting with a huge number of other molecular partners.2 These properties are finely modulated by several PTMs, adding a further level of complexity to HMGA activities/functions.10

TECHNICAL NOTE

In this work, both the primary sequence and the contribution of PTMs have been taken into account and we demonstrated that although ID, HMGA proteins are not extended but, on the contrary, acquire a compact conformation, requiring a more sophisticated interpretation of the term “intrinsically disordered”. Indeed, this term suggests the lack a fixed three-dimensional structure but it does not exclude that the polypeptide chain preferentially acquires some bended conformations which are favored by the formation of intramolecular interactions. All the biochemical and MS techniques used to assay HMGA conformations support this view (see Figure 6 for a schematic representation of HMGA conformational transitions). The most relevant result of this work is that we provide a link between a biochemical property of HMGA proteins and their structural features, that is, accessibility with respect to a set of modifying enzymes is linked to the presence of their acidic tail. Our data support a model in which the tail masks the protein
protein interaction domain causing a protein conformation alteration that impairs enzymatic accessibility. Moreover, IMSMS analyses suggest a role for constitutive phosphorylation of the C-terminal tail in enhancing the compactness level of HMGA proteins. These results now allow us to better interpret our previous data demonstrating that truncated HMGA2 has altered DNA- and protein
protein interaction properties23,37 and that the tail phosphorylation modulates DNA-binding affinity.9 In analogy with the HMGB1 protein,27
31 we thus propose a model in which intramolecular interactions occurring between the tail and inner positive portions of HMGA proteins lead to a compaction of the whole molecule. The resulting conformational alterations could be responsible for the different transforming ability of CT forms with respect to FL HMGA proteins. Obviously, the functional difference between CT and FL forms could be also due to other additional mechanisms. One possibility is that the acidic C-terminal tail constitutes a steric hindrance or an electrostatic repeller with respect to HMGA interactions. We cannot furthermore exclude that the C-terminal tail could have a role in modulating HMGA protein stability, as was previously demonstrated for HMGB proteins.30 The reported data constitute a proof of concept for HMGA as compact proteins, implying that not all the amino acid residues are equally accessible. This is important information for the development of strategies in which HMGA activities are targeted by means of interacting molecules. This is in the light of recent findings showing that molecules able to bind to HMGA1a with high affinity and to compete with HMGA1a cellular partners counteract its oncogenic activity. HMGA-interacting L-RNA oligonucleotides (Spiegelmers) have for example been demonstrated to act as efficient therapeutic agents in a pancreatic adenocarcinoma xenograft mouse model.38 Our data suggest that the region comprised between the second and the third DBD could be a potential target to interfere with HMGA activities. Indeed, this region seems to be the one most protected by the C-terminal tail and thus most linked to the negative modulatory role of the tail with respect to HMGA oncogenic activities. It is striking that the target emerging from this work overlaps almost perfectly with the HMGA protein
protein interaction domain.2 To our knowledge, this is the first example of a characterization, obtained by IMS-MS, of changes in the conformational equilibria of an ID protein due to PTMs or sequence deletion. ID proteins constitute a growing protein family playing crucial biological functions.16 Since this technique preserve native 3289

dx.doi.org/10.1021/pr200116w |J. Proteome Res. 2011, 10, 3283–3291

Journal of Proteome Research structures and is also used to study protein complexes, it could be adopted as a screening method to assess the influence of both PTMs and small interacting molecules on ID protein conformations, thus revealing their efficiency in perturbing a “native” condition and, hopefully, protein functions.

’ ASSOCIATED CONTENT

bS

Supporting Information Supplementary figures. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]; manfi[email protected]. Tel.: þ39-040-5583676/-3675. Fax: þ39-040-5583694.

’ ACKNOWLEDGMENT We are grateful to S. Zammitti, S. Pegoraro, I. Pellarin, G. Ros, and L. Spagnolo for a critical reading of the manuscript and we particularly thank A. Rustighi and A. Tossi for helpful revision and comments. This work was supported by grants from: Universit
a degli Studi di Trieste (Italy), Associazione Italiana per la Ricerca sul Cancro (AIRC), Milan (Italy), Regione Friuli-Venezia Giulia [Progetto AITT 2007], Ministero dell’Istruzione, dell’Universit
a e della Ricerca (MIUR) [PRIN 2007 and FIRB 2008, project ‘Rete Nazionale Proteomica’ RBRN07BMCT] to G.M., and Agenzia Spaziale Italiana (ASI) [MOMA project] of to V.G. ’ REFERENCES (1) Sgarra, R.; Rustighi, A.; Tessari, M. A.; Di Bernardo, J.; Altamura, S.; Fusco, A.; Manfioletti, G.; Giancotti, V. Nuclear phosphoproteins HMGA and their relationship with chromatin structure and cancer. FEBS Lett. 2004, 574, 1–8. (2) Sgarra, R.; Zammitti, S.; Lo Sardo, A.; Maurizio, E.; Arnoldo, L.; Pegoraro, S.; Giancotti, V.; Manfioletti, G. HMGA molecular network: From transcriptional regulation to chromatin remodeling. Biochim. Biophys. Acta 2010, 1799, 37–47. (3) Fusco, A.; Fedele, M. Roles of HMGA proteins in cancer. Nat. Rev. Cancer 2007, 7, 899–910. (4) Thuault, S.; Valcourt, U.; Petersen, M.; Manfioletti, G.; Heldin, C. H.; Moustakas, A. Transforming growth factor-beta employs HMGA2 to elicit epithelial-mesenchymal transition. J. Cell Biol. 2006, 174, 175–83. (5) Tessari, M. A.; Gostissa, M.; Altamura, S.; Sgarra, R.; Rustighi, A.; Salvagno, C.; Caretti, G.; Imbriano, C.; Mantovani, R.; Del Sal, G.; Giancotti, V.; Manfioletti, G. Transcriptional activation of the cyclin A gene by the architectural transcription factor HMGA2. Mol. Cell. Biol. 2003, 23, 9104–16. (6) Pierantoni, G. M.; Rinaldo, C.; Esposito, F.; Mottolese, M.; Soddu, S.; Fusco, A. High Mobility Group A1 (HMGA1) proteins interact with p53 and inhibit its apoptotic activity. Cell Death Differ. 2006, 13, 1554–63. (7) Frasca, F.; Rustighi, A.; Malaguarnera, R.; Altamura, S.; Vigneri, P.; Del Sal, G.; Giancotti, V.; Pezzino, V.; Vigneri, R.; Manfioletti, G. HMGA1 inhibits the function of p53 family members in thyroid cancer cells. Cancer Res. 2006, 66, 2980–89. (8) Cleynen, I.; Van de Ven, W. J. The HMGA proteins: a myriad of functions. Int. J. Oncol. 2008, 32, 289–305. (9) Sgarra, R.; Maurizio, E.; Zammitti, S.; Lo Sardo, A.; Giancotti, V.; Manfioletti, G. Macroscopic differences in HMGA oncoproteins posttranslational modifications: C-terminal phosphorylation of HMGA2 affects its DNA binding properties. J. Proteome Res. 2009, 8, 2978–89.

TECHNICAL NOTE

(10) Zhang, Q.; Wang, Y. High mobility group proteins and their posttranslational modifications. Biochim. Biophys. Acta 2008, 1784, 1159–66. (11) Pierantoni, G. M.; Battista, S.; Pentimalli, F.; Fedele, M.; Visone, R.; Federico, A.; Santoro, M.; Viglietto, G.; Fusco, A. A truncated HMGA1 gene induces proliferation of the 3T3-L1 pre-adipocytic cells: a model of human lipomas. Carcinogenesis 2003, 24, 1861–69. (12) Fedele, M.; Berlingieri, M. T.; Scala, S.; Chiariotti, L.; Viglietto, G.; Rippel, V.; Bullerdiek, J.; Santoro, M.; Fusco, A. Truncated and chimeric HMGI-C genes induce neoplastic transformation of NIH3T3 murine fibroblasts. Oncogene 1998, 17, 413–18. (13) Li, Y.; Lu, J.; Prochownik, E. V. Dual role for SUMO E2 conjugase Ubc9 in modulating the transforming and growth-promoting properties of the HMGA1b architectural transcription factor. J. Biol. Chem. 2007, 282, 13363–71. (14) Thuault, S.; Tan, E. J.; Peinado, H.; Cano, A.; Heldin, C. H.; Moustakas, A. HMGA2 and Smads co-regulate SNAIL1 expression during induction of epithelial-to-mesenchymal transition. J. Biol. Chem. 2008, 283, 33437–46. (15) Henriksen, J.; Stabell, M.; Meza-Zepeda, L. A.; Lauvrak, S. A.; Kassem, M.; Myklebost, O. Identification of target genes for wild type and truncated HMGA2 in mesenchymal stem-like cells. BMC Cancer 2010, 10, 329. (16) Uversky, V. N.; Dunker, A. K. Understanding protein nonfolding. Biochim. Biophys. Acta 2010, 1804, 1231–64. (17) Uversky, V. N.; Oldfield, C. J.; Dunker, A. K. Showing your ID: intrinsic disorder as an ID for recognition, regulation and cell signaling. J. Mol. Recognit. 2005, 18, 343–84. (18) Huth, J. R.; Bewley, C. A.; Nissen, M. S.; Evans, J. N.; Reeves, R.; Gronenborn, A. M.; Clore, G. M. The solution structure of an HMGI(Y)-DNA complex defines a new architectural minor groove binding motif. Nat. Struct. Biol. 1997, 4, 657–65. (19) Konermann, L.; Tong, X.; Pan, Y. Protein structure and dynamics studied by mass spectrometry: H/D exchange, hydroxyl radical labeling, and related approaches. J. Mass Spectrom. 2008, 43, 1021–36. (20) Yan, X.; Watson, J.; Ho, P. S.; Deinzer, M. L. Mass spectrometric approaches using electrospray ionization charge states and hydrogen-deuterium exchange for determining protein structures and their conformational changes. Mol. Cell. Proteomics 2004, 3, 10–23. (21) Kanu, A. B.; Dwivedi, P.; Tam, M.; Matz, L.; Hill, H. H., Jr. Ion mobility-mass spectrometry. J. Mass Spectrom. 2008, 43, 1–22. (22) Reeves, R.; Adair, J. E. Role of high mobility group (HMG) chromatin proteins in DNA repair. DNA Repair (Amst) 2005, 4, 926–38. (23) Noro, B.; Licheri, B.; Sgarra, R.; Rustighi, A.; Tessari, M. A.; Chau, K. Y.; Ono, S. J.; Giancotti, V.; Manfioletti, G. Molecular dissection of the architectural transcription factor HMGA2. Biochemistry 2003, 42, 4569–77. (24) Frankel, A.; Yadav, N.; Lee, J.; Branscombe, T. L.; Clarke, S.; Bedford, M. T. The novel human protein arginine N-methyltransferase PRMT6 is a nuclear enzyme displaying unique substrate specificity. J. Biol. Chem. 2002, 277, 3537–43. (25) Sgarra, R.; Lee, J.; Tessari, M. A.; Altamura, S.; Spolaore, B.; Giancotti, V.; Bedford, M. T.; Manfioletti, G. The AT-hook of the chromatin architectural transcription factor high mobility group A1a is arginine-methylated by protein arginine methyltransferase 6. J. Biol. Chem. 2006, 281, 3764–72. (26) Li, X.; Romero, P.; Rani, M.; Dunker, A. K.; Obradovic, Z. Predicting protein disorder for N-, C-, and internal regions. Genome Inform. 1999, 10, 30–40. (27) Kawase, T.; Sato, K.; Ueda, T.; Yoshida, M. Distinct domains in HMGB1 are involved in specific intramolecular and nucleosomal interactions. Biochemistry 2008, 47, 13991–6. (28) Watson, M.; Stott, K.; Thomas, J. O. Mapping intramolecular interactions between domains in HMGB1 using a tail-truncation approach. J. Mol. Biol. 2007, 374, 1286–97. (29) Wang, Q.; Zeng, M.; Wang, W.; Tang, J. The HMGB1 acidic tail regulates HMGB1 DNA binding specificity by a unique mechanism. Biochem. Biophys. Res. Commun. 2007, 360, 14–9. 3290

dx.doi.org/10.1021/pr200116w |J. Proteome Res. 2011, 10, 3283–3291

Journal of Proteome Research

TECHNICAL NOTE

(30) Knapp, S.; M€uller, S.; Digilio, G.; Bonaldi, T.; Bianchi, M. E.; Musco, G. The long acidic tail of high mobility group box 1 (HMGB1) protein forms an extended and flexible structure that interacts with specific residues within and between the HMG boxes. Biochemistry 2004, 43, 11992–7. (31) Pasheva, E.; Sarov, M.; Bidjekov, K.; Ugrinova, I.; Sarg, B.; Lindner, H.; Pashev, I. G. In vitro acetylation of HMGB-1 and
2 proteins by CBP: the role of the acidic tail. Biochemistry 2004, 43, 2935–40. (32) Fontana, A.; Polverino de Laureto, P.; Spolaore, B.; Frare, E.; Picotti, P.; Zambonin, M. Probing protein structure by limited proteolysis. Acta Biochim. Pol. 2004, 51, 299–321. (33) Fontana, A.; Polverino de Laureto, P.; De Filippis, V.; Scaramella, E.; Zambonin, M. Probing the partly folded states of proteins by limited proteolysis. Fold. Des. 1997, 2, R17–26. (34) Kaltashov, I. A.; Abzalimov, R. R. Do ionic charges in ESI MS provide useful information on macromolecular structure?. J. Am. Soc. Mass. Spectrom. 2008, 19, 1239–46. (35) Stuart, D.; Sellers, W. R. Linking somatic genetic alterations in cancer to therapeutics. Curr. Opin. Cell Biol. 2009, 21, 304–10. (36) Fedele, M.; Visone, R.; De Martino, I.; Palmieri, D.; Valentino, T.; Esposito, F.; Klein-Szanto, A.; Arra, C.; Ciarmiello, A.; Croce, C. M.; Fusco, A. Expression of a truncated Hmga1b gene induces gigantism, lipomatosis and B-cell lymphomas in mice. Eur. J. Cancer 2010, 47, 470–8. (37) Sgarra, R.; Tessari, M. A.; Di Bernardo, J.; Rustighi, A.; Zago, P.; Liberatori, S.; Armini, A.; Bini, L.; Giancotti, V.; Manfioletti, G. Discovering high mobility group A molecular partners in tumour cells. Proteomics 2005, 5, 1494–506. (38) Maasch, C.; Vater, A.; Buchner, K.; Purschke, W. G.; Eulberg, D.; Vonhoff, S.; Klussmann, S. Polyetheylenimine-polyplexes of Spiegelmer NOX-A50 directed against intracellular high mobility group protein A1 (HMGA1) reduce tumor growth in vivo. J. Biol. Chem. 2010, 285, 40012–8. (39) Halligan, B. D.; Ruotti, V.; Jin, W.; Laffoon, S.; Twigger, S. N.; Dratz, E. A. ProMoST (Protein Modification Screening Tool): A webbased tool for mapping protein modifications on two-dimensional gels. Nucleic Acids Res. 2004, 32, W638–44.

3291

dx.doi.org/10.1021/pr200116w |J. Proteome Res. 2011, 10, 3283–3291