Mass Spectrometric Approach for Characterizing the Disordered Tail

Jan 16, 2015 - Kazumi Saikusa†, Aritaka Nagadoi†, Kana Hara†, Sotaro .... were applied: 0.8–1.0 kV of capillary voltage, 20 V of cone voltage,...
0 downloads 0 Views 1MB Size
Article pubs.acs.org/ac

Mass Spectrometric Approach for Characterizing the Disordered Tail Regions of the Histone H2A/H2B Dimer Kazumi Saikusa,†,§ Aritaka Nagadoi,† Kana Hara,†,⊥ Sotaro Fuchigami,† Hitoshi Kurumizaka,‡ Yoshifumi Nishimura,† and Satoko Akashi*,† †

Graduate School of Medical Life Science, Yokohama City University, 1-7-29 Suehiro-cho, Tsurumi-ku, Yokohama, Kanagawa 230-0045, Japan ‡ Graduate School of Advanced Science and Engineering, Waseda University, 2-2 Wakamatsu-cho, Shinjuku-ku, Tokyo 162-8480, Japan S Supporting Information *

ABSTRACT: The histone H2A/H2B dimer is a component of nucleosome core particles (NCPs). The structure of the dimer at the atomic level has not yet been revealed. A possible reason for this is that the dimer has three intrinsically disordered tail regions: the N- and C-termini of H2A and the N-terminus of H2B. To investigate the role of the tail regions of the H2A/ H2B dimer structure, we characterized behaviors of the H2A/ H2B mutant dimers, in which these functionally important disordered regions were depleted, using mass spectrometry (MS). After verifying that the acetylation of Lys residues in the tail regions had little effect on the gas-phase conformations of the wild-type dimer, we prepared two histone H2A/H2B dimer mutants: an H2A/H2B dimer depleted of both N-termini (dN-H2A/dN-H2B) and a dimer with the N- and C-termini of H2A and the N-terminus of H2B depleted (dNC-H2A/dN-H2B). We analyzed these mutants using ion mobility-mass spectrometry (IM-MS) and hydrogen/deuterium exchange mass spectrometry (HDX-MS). With IM-MS, reduced structural diversity was observed for each of the tail-truncated H2A/H2B mutants. In addition, global HDX-MS proved that the dimer mutant dNCH2A/dN-H2B was susceptible to deuteration, suggesting that its structure in solution was somewhat loosened. A partial relaxation of the mutant’s structure was demonstrated also by IM-MS. In this study, we characterized the relationship between the tail lengths and the conformations of the H2A/H2B dimer in solution and gas phases, and demonstrated, using mass spectrometry, that disordered tail regions play an important role in stabilizing the conformation of the core region of the dimer in both phases.

N

the histone H2A/H2B dimer and (H3/H4)2 tetramer, the components of the histone octamer. In our previous paper, gas-phase structures of the histone H2A/H2B dimer and (H3/H4)2 tetramer were investigated by ion mobility-mass spectrometry (IM-MS) and molecular dynamics (MD) simulation.9 Because IM-MS provides not only mass information but also collision cross sections (CCS) of the analyte ions, even for large protein complexes, it is a promising technique for characterizing protein complexes with intrinsically disordered regions (IDRs).9−14 However, IM-MS cannot suggest the atomic-level structure on its own. Thus, the combined method of IM-MS with MD simulation is effective for analyzing protein complexes, including proteins with IDRs. Our previous report on IM-MS and MD simulation of the histone H2A/H2B dimer and (H3/H4)2 tetramer demonstrated that in the gas phase, tail regions were partly collapsed

ucleosome core particles (NCPs) are the minimum structural units for DNA storage in the nucleus of eukaryotic cells, and they are composed of a histone octamer wrapped by ∼146 bp of DNA. The histone octamer contains two molecules of each histone protein, H2A, H2B, H3, and H4, which possess highly flexible tail regions. It is known that modifications of the tail regions manage dynamic structural changes in NCPs and control regulatory activity of DNA transcription, replication, and repair, with the aid of chromatin remodeling factors such as histone chaperones.1−6 For example, acetylation of Lys side chains in the flexible tail regions generally allows the NCP structure to loosen, and transcription is subsequently activated. Therefore, structural characterization of histone proteins including their tail regions is significant for understanding the mechanisms of NCP assembly and disassembly. However, X-ray crystallography demonstrates only an organized structure for the core part, and little information on the flexible tail regions is available.7,8 Furthermore, no atomic-level structure has been revealed for © 2015 American Chemical Society

Received: October 2, 2014 Accepted: January 16, 2015 Published: January 16, 2015 2220

DOI: 10.1021/ac503689w Anal. Chem. 2015, 87, 2220−2227

Article

Analytical Chemistry or unfolded at different levels but the core regions were almost completely maintained. This resulted in two conformational population groups with compact or loose structures. In the present study, we focused on characterization of the role of the disordered tail regions of the histone H2A/H2B dimer using mass spectrometry. First, highly acetylated wildtype histone H2A/H2B dimer was prepared and analyzed by electrospray ionization (ESI) IM-MS to assess the effect of charges in the tail regions on the structural variety of the H2A/ H2B dimer. Next, two H2A/H2B dimer mutants, an H2A/H2B dimer with both N-terminal tails depleted (dN-H2A/dN-H2B) and a dimer with the N- and C-terminal tails of H2A depleted and the N-terminal tail of H2B depleted (dNC-H2A/dN-H2B), were prepared to investigate the role of the histone tails directly. We then analyzed them and the wild-type dimer using ESI-IM-MS. Stability of their folded conformation in solution was also examined by hydrogen/deuterium exchange mass spectrometry (HDX-MS).15,16 MD simulation was used in parallel to investigate the solution and gas-phase structures of the mutant H2A/H2B dimers. With these data, for the first time, we analyzed the relationship between tail lengths and conformations in the solution and gas phases, enabling discussion on the significance of the histone dimer disordered tail regions in both phases.

Figure 1. Amino acid sequences and expected secondary structures of histone H2A and H2B. The secondary structure was drawn with reference to the X-ray structure of human NCP.8 Dotted lines show the tail regions defined by the MD simulation data of the H2A/H2B dimer in solution based on the X-ray structure of human NCP.9 Blue and red arrows indicate the dN- and dNC-histone proteins, respectively.



EXPERIMENTAL SECTION Preparation of Wild-Type Histone Proteins. Wild-type histone H2A and H2B monomers were prepared according to previous reports.17,18 An extra three residues, GSM-, derived from the expression vector, remained at the N-terminus of the protein after thrombin treatment. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and matrixassisted laser desorption/ionization time-of-flight (MALDITOF) MS (Autoflex, Bruker Daltonics, Billerica, MA) were used to confirm the purity and molecular masses of the obtained proteins. To prepare the H2A/H2B dimer, each monomer protein was dissolved in a denaturing buffer (7 M Guanidine-HCl, 500 mM ammonium acetate, or 7 M Guanidine-HCl, 10 mM Tris−HCl (pH 7.5)) and mixed at a 1:1 molar ratio, according to the previous protocol.18 The solution was then subjected to dialysis with a microdialysis tool (Microdialyzer TOR-3K, cutoff molecular weight of 3500, Nippon Genetics Co., Ltd., Japan) against 2 M ammonium acetate to refold the dimer. The sample solution of the H2A/H2B dimer was diluted with 2 M ammonium acetate (pH 6.8), resulting in 10 μM histone dimers. Acetylation of the H2A/H2B Dimer. Acetylation of the wild-type H2A/H2B dimer was performed using histone acetyltransferase (HAT) p300 (Enzo Life Sciences, Farmingdale, NY) at a 10:1 substrate:enzyme ratio (w/w). The mixture of the histone H2A/H2B dimer and HAT p300 was incubated in 50 mM Tris−HCl (pH 8.0), 10% glycerol, 0.1 mM EDTA, and 1 mM DTT in the presence of 1 mM Acetyl-CoA at 37 °C for 26 h. Then, the reaction mixture was dialyzed against 500 mM ammonium acetate with the Microdialyzer TOR-3K. The extent of acetylation of each histone was determined by MALDI-TOF MS. The protein was digested with endoproteinase Glu-C (Roche Diagnostics, Mannheim, 20:1 (substrate:enzyme)) at 37 °C for 18 h after dialysis against 50 mM ammonium acetate (pH 4.0) to analyze the acetylation sites in each histone protein. Digestion products were analyzed with MALDI-TOF MS.

Preparation of Histone Mutant Proteins. The amino acid sequences of human histone dN-H2A, dNC-H2A, and dNH2B are shown in Figure 1. Histone dN-H2A and dNC-H2A are deletion mutants of H2A; dN-H2A corresponds to an Nterminal truncation mutant of H2A (Lys15−Lys129) while dNCH2A corresponds to an N- and C-terminal truncation mutant of H2A (Lys15−Gln104). Histone dN-H2B is an N-terminal deletion mutant of H2B (Arg31−Lys125). Histone dN-H2A, dNC-H2A, and dN-H2B were overexpressed in Escherichia coli with N-terminal His6-tags. The His6-tag was removed with HRV3C (Accelagen, San Diego, CA) treatment during the purification procedure. An extra four residues, GPGM-, derived from the expression vector, remained at the N-terminus of the protein after treatment with HRV3C. The purity and molecular weights of the obtained proteins and the mutant dimers were acquired in the same manner as those of the wild-type dimer. ESI-IM-MS of Histone Dimer. Mass spectra and arrival times for the histone dimers were acquired using a Tri-wave SYNAPT G2 HDMS (Waters, Milford, MA) with a nanoESI source.14,19−21 The samples in ammonium acetate solutions (pH 6.8) were deposited in gold-coated borosilicate capillaries (Humanix, Japan) and placed in the nanoESI source. Because arrival times obtained by IM-MS contained mass-dependent flight times between the ion mobility cell and the TOF analyzer, they were converted into CCS values using the equation given by Ruotolo et al.22 (details are indicated in the Supporting Information). To perform IM-MS experiments, the following parameters were applied: 0.8−1.0 kV of capillary voltage, 20 V of cone voltage, 4 V of trap collision energy, and 40 V of trap bias voltage. Hydrogen/Deuterium Exchange Mass Spectrometry (HDX-MS). Global hydrogen/deuterium exchange (HDX) of the histone H2A/H2B dimer was basically followed the HDXMS protocol for the H2A/H2B dimer by D’Arcy et al.16 Prior to HDX, each histone dimer was dialyzed against 2.75 mM Tris−HCl (pH 7.5), 4.125 M NaCl, 1.1 mM EDTA, and 5.5 2221

DOI: 10.1021/ac503689w Anal. Chem. 2015, 87, 2220−2227

Article

Analytical Chemistry

H2B dimer was acetylated using histone acetyltransferase p300, which enables specific Lys acetylation of histone proteins.23 The acetylated H2A/H2B dimer was then analyzed by MALDITOF MS, and the number of acetylated Lys residues in each histone protein was distributed between 5 and 8 with a maximum at 6 for H2A and between 6 and 9 with a maximum 7 for H2B, as shown in Figure 3. These acetylation sites in each

mM 2-mercaptoethanol. The sample solutions were diluted 10fold with D2O and kept at 4 °C for HDX. After 10 s, the reaction was quenched by mixing with a quench solution at 1:1.5 (v:v) ratio (2.5 M Guanidine HCl, 10% glycerol, and 0.8% (v/v) formic acid). The sample solution was then desalted using an online LC system with a short C18 column (1.0 × 10 mm, GL science, Japan) eluted with 0.1% formic acid at a flow rate of 50 μL/min. The column and injector sample loop (100 μL) were kept at 6 °C in a Shimadzu column oven (CTO20AC, Japan). Desalted proteins were eluted from the column with 80% acetonitrile/0.1% formic acid and introduced to SYNAPT G2 HDMS. For the measurement of mass, the source temperature was set at 120 °C, and electrospray was carried out at 3.0 kV capillary voltage (positive ion mode). Mass spectra were acquired in the range of m/z 500−4000. The number of incorporated deuterium atoms was determined using the observed average mass of the deuterated protein by subtracting of the theoretical average mass of the nondeuterated protein, while considering the actual deuterium percentage (91%) during deuteration, as described in the Results and Discussion section. The HDX experiments were performed in duplicate, and the average mass was used in the analysis.



RESULTS AND DISCUSSION First, the CCS distributions for the multiply charged ions of the wild-type H2A/H2B dimer were analyzed by ESI-IM-MS. Figure 2 shows the ESI mass spectrum and CCS distribution

Figure 3. MALDI-TOF mass spectra of the nonacetylated (A) and acetylated (B) wild-type H2A/H2B.

histone protein were identified by peptide mass mapping of the Glu-C digest, and it was demonstrated that the Lys residues susceptible to acetylation were mostly located in the tail regions, as summarized in Figure S-1 and Table S-1 in the Supporting Information. The IM-MS measurement was then performed for the acetylated H2A/H2B dimer in 2 M ammonium acetate. Figure 4 shows ESI mass spectra and arrival time distributions of the H2A/H2B dimer before and after acetylation. Observed charge states of the H2A/H2B dimer were unchanged by acetylation; mainly 10+, 11+, and 12+ charged ions were seen in both spectra (Figure 4A,B). To avoid artificial structural changes caused by the harsh experimental conditions, relatively low cone, collision, and trap bias voltages were applied. Thus, the many resultant adducts rather broadened the observed peaks. When analyzing the arrival time distributions of the observed multiply charged ions before and after acetylation, no significant difference was recognized. Ions of the acetylated dimer showed slightly larger arrival times than those of nonacetylated dimer (Figure 4C−E). The distribution profiles of arrival times were similar for the corresponding multiply charged ions. Experimental CCS values of the 11+ charged ions of the nonacetylated H2A/H2B dimers were calculated as 2250 Å2 (Group I) and 2440 Å2 (Group II), whereas those of the acetylated dimers were 2280 Å2 (Group I) and 2460 Å2 (Group II), using the method described in the Supporting Information. To evaluate these values, they were compared with the theoretical CCS values of the structural models of the dimers for Group I and II, as generated by MD simulation.9 Structures of the acetylated dimer were generated using the PyMol software.24 The theoretical CCS values of the nonacetylated dimer were 2251 Å2 (Group I) and 2441 Å2 (Group II), whereas those of the acetylated dimer were 2297 Å2 (Group I) and 2456 Å2 (Group II). The differences in the corresponding theoretical CCS values between the nonacetylated and acetylated dimer models are similar to those in the

Figure 2. ESI mass spectrum and CCS distribution (inset) of the wildtype H2A/H2B dimer. Blue, red, and orange lines represent dimers with 10+, 11+, and 12+ charge states, respectively. The CCS distribution denotes two conformational populations: small (Group I) and large (Group II).

plots for the wild-type dimer ions. Ions with 10+, 11+, and 12+ charges were observed, and their CCS distribution plots demonstrated the existence of two populations, compact (Group I) and loose (Group II) structures. Although the 10+, 11+, and 12+ charged ions were all associated with ammonium acetate and ammonia adducts, the intact dimers and their adducts displayed identical mobility profiles. These populations were consistent with the results of our previous study.9 Furthermore, no highly charged dimer ions were observed in the low m/z region. Charge Effect of H2A/H2B Dimer Tails on Conformational Variety. Because histone tail regions are rich in basic amino acid residues, the repulsion of positive charges in the tail region might occur specifically when the dimer ions are introduced into the gas phase. First, the positive charges of the Lys side chains were negated by acetylation to examine if the positively charged cluster in the tail regions affects the gasphase structure of the H2A/H2B dimer. The wild-type H2A/ 2222

DOI: 10.1021/ac503689w Anal. Chem. 2015, 87, 2220−2227

Article

Analytical Chemistry

Figure 4. ESI mass spectra of the nonacetylated (A) and acetylated (B) wild-type H2A/H2B dimer and arrival time distributions of 10+ (C), 11+ (D), and 12+ (E) charged ions of the wild-type H2A/H2B dimer before and after acetylation. Red and orange arrows show the multiply charged ions of the dimer: 10+, 11+, and 12+ for the nonacetylated (A) and 10+, 11+, 12+, and 13+ for the acetylated (B) dimer. Blue and black plots in (C), (D), and (E) show the distribution profiles of the nonacetylated and acetylated dimer, respectively.

Figure 5. ESI mass spectra of the H2A/H2B dimer in 2 M ammonium acetate obtained in the TOF mode. Red, blue, and green arrows indicate the wild-type H2A/H2B, dN-H2A/dN-H2B, and dNC-H2A/dN-H2B dimers, respectively. Purple and orange triangles show the H2A and H2B monomer peaks, respectively.

interactions between the protein and the polyphosphate backbone of DNA, resulting in a loosening of the NCP structure in preparation for transcription initiation. In the present study, we performed acetylation of the H2A/H2B dimer, not of the NCP, and subjected the samples prepared in 2 M ammonium acetate to native MS. To avoid dissociation of the H2A/H2B dimer, which has an interface stabilized mainly by hydrophobic interactions, during the native MS experiments, a high concentration of ammonium acetate was required for the sample preparation. Thus, the fact that no significant size difference was observable between the acetylated and nonacetylated H2A/H2B dimers prepared in 2 M ammonium acetate is in some ways reasonable, as the high ionic strength might have reduced the electrostatic interactions of the tails. No

experimental CCS values. Consequently, it is reasonable to attribute the small increase in the CCS of the acetylated dimer to the mass increase from the acetylation of amino groups. Accordingly, it can be argued that acetylation of the tail regions had nothing to do with the structural change of the H2A/H2B dimer, and that the structural variation derived from the tail regions in the wild-type H2A/H2B dimer was not due to charge repulsion. This will also be discussed later. Acetylation of Lys residues is effective in neutralizing the positive charge on histone tails, and reduces the electrostatic interactions between the phosphate groups of DNA and the amino groups of Lys side chains under physiological conditions. In eukaryotes, acetylation of Lys residues in histone tails of NCP generally activates transcription initiation. This suggests that acetylation triggers a decrease in the electrostatic 2223

DOI: 10.1021/ac503689w Anal. Chem. 2015, 87, 2220−2227

Article

Analytical Chemistry significant difference in the CCS values might have thus resulted for the acetylated and nonacetylated dimers. ESI-IM-MS of Histone Dimer Mutants. We prepared three mutants of the human histone proteins, dN-H2A, dNCH2A, and dN-H2B, in which N- or C-terminal disordered regions were depleted, to examine the effect of the histone multimer tail regions on their conformational variety in more detail (Figure 1). Tail regions were defined by referring to the X-ray crystal structure of the human NCP8 and the MD simulation data of the H2A/H2B dimer in solution9 to design the constructs of the tail-truncated mutants. Consequently, the wild-type H2A/H2B dimer, the dN-H2A/dN-H2B dimer, and the dNC-H2A/dN-H2B dimer were prepared by refolding the H2A and H2B monomers with or without tail regions. ESI mass spectra were then obtained for these three histone dimers in 2 M ammonium acetate, as shown in Figure 5. For the wild-type H2A/H2B dimer, ions with 10+ through 12+ charges were observed in the ESI mass spectra (Figure 5A), whereas 9+ to 11+ charges were observed for the dN-H2A/dNH2B dimer and 8+ to 11+ charges were observed for the dNCH2A/dN-H2B dimer (Figure 5B,C). The most intense peaks were 11+, 10+, and 9+ for the wild-type H2A/H2B, dN-H2A/ dN-H2B, and dNC-H2A/dN-H2B dimers, respectively. The smaller the molecular mass of the dimer, the smaller the observed charge states. In our previous study, MD simulation was carried out for the wild-type H2A/H2B dimer, first in solution then in vacuo, to reproduce the electrospray ionization process.9 An ensemble of 150 simulated structures at 300 K exhibited 11 439 Å2 of the average accessible surface area (ASA) value. This corresponds to the 11+ charge state according to the proposed equation for the relationship between protein ASA values and average charge states of the multiply charged ions observed in the ESI mass spectra (eq 2 in the Supporting Information).25 In a similar manner, the MDsimulated structures of the dN-H2A/dN-H2B and dNC-H2A/ dN-H2B dimers were obtained as described in the Supporting Information. The calculated ASA values were 10 003 and 8995 Å2, corresponding to 10+ and 9+, respectively. These calculated charge states for the three dimers agreed well with the experimentally observed principal charge states, suggesting that tail truncation causes a decrease in the ASA value of the dimer, which results in a reduced charge state in the gas phase. Next, these dimers were subjected to ESI-IM-MS, and the CCS values of the observed ions were analyzed. Figure 6A−C presents the CCS distribution plots of each multiply charged ion observed in the ESI mass spectra. In the case of the wildtype H2A/H2B dimer, the 11+ charged ion, i.e., the most abundant ion, clearly defined two populations: Group I with a centroid CCS value of 2250 Å2 and Group II with one of 2440 Å2, as shown in Figure 6A. Similarly, the CCS distributions of the most abundant ions for the dN-H2A/dN-H2B and dNCH2A/dN-H2B dimers, 10+ and 9+, respectively, can be roughly divided into two groups: small (Group I: ∼2040 Å2 for dNH2A/dN-H2B and ∼1850 Å2 for dNC-H2A/dN-H2B) and large (Group II: ∼2170 Å2 for dN-H2A/dN-H2B and ∼1960 Å2 for dNC-H2A/dN-H2B) (Figure 6B,C). The relative overall ratio of Group II, the conformation population with larger CCS values, got smaller as the tail length was reduced. The 9+ charged ion of the dNC-H2A/dN-H2B dimer, which had extremely short tail regions (97−104 for H2A and 31−37 for H2B) had a tiny shoulder present in the CCS distribution plot for the population corresponding to Group II, and the distribution width was quite narrow (Figure 6C). This strongly

Figure 6. CCS distributions of the wild-type H2A/H2B (A), dNH2A/dN-H2B (B), and dNC-H2A/dN-H2B dimers (C) obtained by ESI-IM-MS. Distribution profiles of each multiply charged ion observed in the mass spectra are indicated.

suggests that the gas-phase conformational variety was derived from the diverse behaviors of the flexible tail regions. Furthermore, non-negligible gas-phase behavior was found for the smallest mutant dimer, dNC-H2A/dN-H2B. The CCS distribution envelopes for the ions with the highest charges, 12+ for the wild-type and 11+ for dN-H2A/dN-H2B, overlapped well with the back side of the Group II CCS distribution envelopes of the most dominantly observed ions (11+ for the wild-type and 10+ for the dN-H2A/dN-H2B dimers), as shown in Figure 6A,B. Additionally, CCS distribution plots for the ions with the lowest charges of these two dimers overlapped with the front side of the distribution envelope of each dominantly observed ion. These results imply that the gas-phase structures of all the multiply charged ions for these two dimers fall into two groups, small (Group I) and large (Group II). In contrast, the CCS distribution peak-top value for the 10+ charged ion of the dNC-H2A/dN-H2B dimer was 2120 Å2, implying that some fraction of the 10+ ions are larger in size than the Group II population of the most abundant 9+ ion, as shown in Figure 6C. Considering that ions of tightly folded globular proteins are present as a single population in the CCS distribution,26 this might suggest that some population of the structure of the dNC-H2A/dN-H2B dimer, the dimer mutant with extremely short tail regions, was loose and could relax easily to a certain degree in the gas phase. This will be discussed in more detail below. HDX-MS of the Wild-Type and Mutant Histone Dimers. Global HDX was performed for each histone dimer (wild-type H2A/H2B, dN-H2A/dN-H2B, and dNC-H2A/dN2224

DOI: 10.1021/ac503689w Anal. Chem. 2015, 87, 2220−2227

Article

Analytical Chemistry Table 1. Global HDX Data of the Histone Dimers

H2A/H2B dimer dN-H2A/dN-H2B dimer dNC-H2A/dN-H2B dimer a

# of amide hydrogen (A)

observed mass increase ± SDa

# of exchanged amide hydrogen (B)

# of unexchanged amide hydrogen = (A) − (B)

H2A H2B dN-H2A

126 121 112

50 ± 3.5 38 ± 1.4 38 ± 2.0

55 42 42

71 79 70

dN-H2B dNCH2A dN-H2B

95 89

18 ± 0.5 25 ± 1.1

20 27

75 62

95

20 ± 2.0

21

74

The average and standard deviation values obtained with two HDX experiments are indicated.

dimers, the number of unexchanged amide hydrogens was ∼5 less than that of wild-type H2B in the wild-type H2A/H2B dimer. In contrast, the number of unexchanged amide hydrogen in dN-H2A in dN-H2A/dN-H2B was similar to that of wildtype H2A in the H2A/H2B dimer. These results suggest that some part of the secondary structure of the core regions might have been disrupted in dNC-H2A and dN-H2B of the mutant dimers due to deletion of the long tail regions that stabilize the dimer core. Perhaps this disruption is related to the other, largeCCS-value population seen in the CCS distribution plot of the 10+ charged ion of the dNC-H2A/dN-H2B dimer. MD simulation was carried out for these three dimers, first in solution then in vacuo, to investigate their atomic-level structures. Details are described in the Supporting Information. As shown in Figure S-2 (Supporting Information), RMSF plots of the MD-simulated structures of these dimers in solution clearly demonstrate that the core regions, at amino acid residues 15−98 for H2A and 35−125 for H2B, were quite stable throughout the calculations for all three dimers. This is inconsistent with our experimental results: HDX-MS suggested that the core region of the smallest mutant dimer was somewhat loosened in the solution phase. MD simulation in the gas phase was performed following MD simulation in solution. The folded core region structure of the wild-type and mutant dimers remained almost unchanged throughout successive MD simulations in vacuo. Furthermore, the MD simulation at 300 K did not suggest the multiple conformational populations observed in the IM-MS experiments. Such a discrepancy might be attributed to the MD simulation time scale, which is much shorter than the time scales for HDX-MS or IM-MS. The MD simulation in solution was performed for 10 ns, whereas HDX-MS was carried out for 10 s. Similarly, the gas-phase MD simulation was performed for 5 ns, whereas ions travel in the mobility region for several tens of ms. Alternatively, extensive examination of the force field for simulation might be required, or the initial structures of the dimers used in the MD simulation, which were constructed based on the H2A/H2B dimer extracted from the human NCP structure, might not be appropriate. If more-appropriate MD simulation conditions are defined to reproduce the IM-MS experiments, the structures accordingly predicated might then be more consistent with the results of HDX-MS and IM-MS. As mentioned above, IM-MS suggested that some population of the dNC-H2A/dN-H2B dimer structure was loosened slightly in the gas phase. This indicates that IM-MS provides structural information in the gas phase, reflecting structural properties in solution. Consequently, the HDX-MS and IM-MS results imply that tail regions are necessary to keep the folded structure of the histone core, and that the histone core alone cannot preserve its folded structure in solution and gas phases.

H2B) in a similar manner to the previous protocol16 to examine the effect of the tail regions on structural rigidness in solution. When amide hydrogens are hydrogen-bonded and involved in α-helices or β-sheets, they do not readily exchange with the deuterium in D2O. In contrast, amide hydrogens in the flexible regions on the protein surface are rapidly exchanged with deuterium upon contact with D2O at a neutral pH. The rate of exchange of a hydrogen atom with deuterium at the backbone amide also depends on the solution temperature. Using HDX in combination with MS, the stability of the α-helices and βsheets in a protein can be characterized as indicated below. The protein is first dissolved in a deuterated solution at pH ∼7, and HDX then commences. After incubation for an appropriate time period, the HDX is quenched by lowering the solution pH and temperature.27,28 When the solution pH is decreased from pH 7 to pH 2.5, the HDX rate is reduced by 4 orders of magnitude. Lowering the solution temperature, e.g., from 20 to 0 °C, also causes a reduction in the exchange rate by approximately 1 order of magnitude. In such a manner, the analysis of deuterium incorporation into amide hydrogens enables structural stability estimation of the α-helices and βsheets within a protein in solution. The difference in the intact protein’s mass before and after HDX represents the flexibility of the entire molecule. Table 1 summarizes the results of HDX of the histone dimers. The mass increases observed for each monomer are 50 (wild-type H2A in wild-type H2A/H2B), 38 (dN-H2A in dNH2A/dN-H2B), 25 (dNC-H2A in dNC-H2A/dN-H2B), 38 (wild-type H2B in wild-type H2A/H2B), 18 (dN-H2B in dNH2A/dN-H2B), and 20 (dN-H2B in dNC-H2A/dN-H2B). Because deuterium exchange was carried out in 91% deuterated aqueous solution, the actual quantity of exchanged amide hydrogens is 1.1 times the value of the observed mass increase, as indicated in Table 1. The number of unexchanged amide hydrogens was calculated by subtracting the number of exchanged amide hydrogens from the number of total exchangeable amide hydrogens. The values are 71 (=126 − (50 × 1.1), wild-type H2A/H2B), 70 (dN-H2A/dN-H2B), and 62 (dNC-H2A/dN-H2B) for H2A and 79 (wild-type H2A/ H2B), 75 (dN-H2A/dN-H2B), and 74 (dNC-H2A/dN-H2B) for H2B. For dN-H2A, dNC-H2A, and dN-H2B, most parts of flexible N- and/or C-terminal regions are truncated, but the same core region of the H2A/H2B dimer remains (15−104 for H2A and 31−125 for H2B). If the secondary structures of the core region within the H2A/H2B dimers are identical, the number of unexchanged amide hydrogens should be the same. However, the number of unexchanged amide hydrogens in dNC-H2A in the dNC-H2A/dN-H2B dimer was ∼10 less than those in wild-type H2A or dN-H2A in each dimer. In the case of dN-H2B in the dN-H2A/dN-H2B or dNC-H2A/dN-H2B 2225

DOI: 10.1021/ac503689w Anal. Chem. 2015, 87, 2220−2227

Article

Analytical Chemistry It is also suggested that truncation of the flexible regions may partly disrupt the intact folded structure, which is essential for proper function under physiological conditions. The lower stability of the helical regions in the tail-truncated dimer is inconsistent with previously reported results; Placek and Gloss had demonstrated that the stability of the dN-H2A/ dN-H2B dimer decreased compared with the wild-type dimer using far-UV circular dichroism (CD) and intrinsic tyrosine fluorescence with varying concentrations of urea.29 They found that a larger number of amino acids were involved in helix formation in the dN-H2A/dN-H2B dimer than in the wild-type dimer, and that the wild-type dimer was easier to denature in the presence of urea than the tail-truncated dimer. They concluded that the electrostatic repulsion of the basic-residue clusters in the N-terminal tail of each histone protein destabilized the folded structure, and that truncation of the 15 and 31 amino acid residues from the N-termini of H2A and H2B led to stabilization of the dimer. Examining the experimental conditions, all of the equilibrium unfolding experiments conducted by Placek and Gloss were carried out for the dimers in 20 mM potassium phosphate (pH 7.2) with no additional salt. In contrast, in the present study, the HDXMS experiments of the dimers were carried out in tris buffer at pH 7.5 with 412.5 mM NaCl16 and the ESI mass spectra were obtained for the dimers in 2 M ammonium acetate. Meanwhile, Karantza et al. reported that the melting temperature (Tm) of the dimer decreased according to the salt concentration, and that Tm of the N-tail truncated H2A/ H2B dimer was lower than that of the wild-type dimer at >25 mM NaCl but higher at