Exposing the subunit diversity within protein complexes: A mass spectrometry approach

Methods 59 (2013) 270–277

Contents lists available at SciVerse ScienceDirect

Methods journal homepage: www.elsevier.com/locate/ymeth

Exposing the subunit diversity within protein complexes: A mass spectrometry approach Shelly Rozen a, Alessandra Tieri a, Gabriela Ridner a, Ann-Kathrin Stark a, Tilo Schmaler b, Gili Ben-Nissan a, Wolfgang Dubiel b, Michal Sharon a,⇑ a b

Department of Biological Chemistry, Weizmann Institute of Science, Rehovot 76100, Israel Department of Surgery, Charité – Universitätsmedizin, Berlin 10117, Germany

a r t i c l e

i n f o

Article history: Available online 4 January 2013 Communicated by Peter Schuck Keywords: Mass spectrometry/methods Chromatography/methods Protein complexes Post-translational modifications Protein isoforms COP9 signalosome

a b s t r a c t Identifying the list of subunits that make up protein complexes constitutes an important step towards understanding their biological functions. However, such knowledge alone does not reveal the full complexity of protein assemblies, as each subunit can take on multiple forms. Proteins can be post-translationally modified or cleaved, multiple products of alternative splicing can exist, and a single subunit may be encoded by more than one gene. Thus, for a complete description of a protein complex, it is necessary to expose the diversity of its subunits. Adding this layer of information is an important step towards understanding the mechanisms that regulate the activity of protein assemblies. Here, we describe a mass spectrometry-based approach that exposes the array of protein variants that comprise protein complexes. Our method relies on denaturing the protein complex, and separating its constituent subunits on a monolithic column prepared in-house. Following the subunit elution from the column, the flow is split into two fractions, using a Triversa NanoMate robot. One fraction is directed straight into an on-line ESI-QToF mass spectrometer for intact protein mass measurements, while the rest of the flow is fractionated into a 96-well plate for subsequent proteomic analysis. The heterogeneity of subunit composition is then exposed by correlating the subunit sequence identity with the accurate mass. Below, we describe in detail the methodological setting of this approach, its application on the endogenous human COP9 signalosome complex, and the significance of the method for structural mass spectrometry analysis of intact protein complexes. Ó 2013 Elsevier Inc. All rights reserved.

1. Introduction Production of ATP, DNA replication, transcription, protein synthesis and degradation, are only a small list of the biological processes that are carried out by protein complexes. In fact, the entire cell can be viewed as a factory in which its many diverse functions are carried out by the orchestrated activities of such complexes, each of which requires the coordinated action of multiple subunits that assemble into a functional unit of distinct composition and structure. In recent years, experimental advances on several fronts have expanded our ability to study the structural and functional properties of such large protein complexes [1]. Techniques to isolate and purify multi-subunit complexes have been developed, and technological innovations have enabled the complete sequencing of several eukaryotic genomes [2]. This information, coupled with worldwide proteomics initiatives, have yielded detailed lists of

⇑ Corresponding author. Fax: +972 8 934 6010. E-mail address: [email protected] (M. Sharon). 1046-2023/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ymeth.2012.12.013

the subunits that comprise protein complexes [3]. Determining the accurate composition of protein assemblies is an important step towards understanding their function. However, such knowledge might not be sufficient for unraveling the full complexity of protein complexes’ mode of action, and the regulatory mechanisms that underlie their activities. Protein complexes are likely not uniform in their structure and function. This is due to the fact that a single subunit may be encoded by more than one gene [4], a single gene might produce several alternative splice forms [5–7], protein subunits may be post-translationally modified (PTM) [8,9] or cleaved [10], and single nucleotide polymorphisms (SNPs) may be present [11]. These multiple forms of a single subunit may be integrated within a protein complex. As a consequence, the presence of a specific protein complex within a cell may actually be represented by a diverse group of functionally distinct entities. Moreover, protein complexes are most likely dynamic assemblies whose composition is altered according to the tissue type, the nature of the cell itself (e.g., normal vs. diseased), and the intracellular localization. Such variability in subunit composition is expected to create unique functional properties, enabling the complex to respond and adapt

S. Rozen et al. / Methods 59 (2013) 270–277

to varying cellular conditions [12]. These abilities may be highly developed, and subject to sophisticated modes of regulation [13,14]. Clarifying the heterogeneity of subunit composition is essential for uncovering the multiple functional forms of protein complexes, and their regulatory pathways. However, relatively little attention has been given to this topic. In the past, only a few studies have sought to determine the variability of protein subunits (for examples, see [15–18]). This is mainly due to the fact that large-scale proteomic analysis, which involves the digestion of proteins into peptides, is limited in its ability to describe subunit diversity. This approach, while providing a high number of identifications, often yields low sequence coverage, hindering the detection of endogenous protein cleavages, PTMs and isoforms, all of which are likely to have high sequence homology. Notably, even when the full array of PTMs is probed and their exact localization within the protein sequence is defined, information regarding their composition on a single subunit, and the number of co-existing combinations, is lost. Thus, of necessity, specialized methods had to be developed, in order to identify the full repertoire of subunit compositions. Two-dimensional gel electrophoresis followed by liquid chromatography mass spectrometry (LC–MS) analysis constitutes a known strategy for characterizing the heterogeneity of protein species [17,19]. However, the method relies on the extraction of protein variants from polyacrylamide gels, which often yields low recovery. The 2D gel electrophoresis approach was recently advanced by combining it with intact protein separation and mass measurements [15]. ‘‘Top down’’ approaches based on mass spectrometric analyses of intact proteins [20] have also been applied for mapping protein variants [16–18]. However, these ‘‘top down’’ approaches present a challenge, due to the difficulty in predicting fragmentation pathways, compared to tandem MS of peptides [21]. Nevertheless, an elegant study recently used this approach for large-scale intact protein analysis [16], and 3000 protein species were identified. However, since cell extracts were used, information regarding the ability of each subunit variant to be integrated within a functional complex was lost. Here, we describe our approach for determining the diversity of protein subunit composition. The underlying principle of the method entails correlation between the accurate mass of the protein subunit, and its sequence identity. Our experimental set-up couples capillary-LC separation, a NanoMate robotic system, and a QToF mass spectrometer (Fig. 1). Initially, the protein complex is loaded onto a monolithic column prepared in-house, where it is decomposed into its ‘‘building blocks,’’ by a gradual increase in organic solvent concentration. The individual subunits are then separated from one another, based on their size and chemical properties. Subsequently, the eluted flow is split into two, using the NanoMate robot. One fraction is sprayed directly into the mass spectrometer to accurately determine the mass of the individual subunits, whereas the second fraction is collected into a 96-well plate for sequence identification by tryptic digestion, LC–MS/MS proteomic analysis, and a database search. Our approach is automated and, as a result, easily reproducible, with low sample consumption. Furthermore, this strategy improves upon a method we previously applied, which required two LC separations of the protein subunits [22]: one for exact mass determination, and one for sequence identification. With the introduction of the NanoMate robot, the two applications are integrated in a single LC run. Not only is our new method faster, but it also reduces sample consumption, a critical factor when only minute amounts of the protein complex are available.

271

2. Rationale In the course of investigating the structural properties of protein complexes by MS, we realized that there is little agreement between the actual measured mass of a protein subunit, and that predicted by protein sequence databases. Consequently, it was often impossible to determine which protein subunits are present within a particular complex, especially when investigating less established multi-protein assemblies of unknown composition and stoichiometry. Thus, a method enabling correlation of accurate mass with the sequence identity of protein subunits is required. Such an approach would enable us to identify all proteins present in a multi-subunit complex, together with their corresponding intact masses. Only then could we successfully utilize structural MS to determine the non-covalent interactions between subunits, and the overall architecture of the complex [23–25]. 3. Materials and methods 3.1. Chemicals and reagents For column synthesis, fused-silica capillary tubings, polyimidecoated with 200 lm I.D. and 375 lm O.D, were purchased from Polymicro Technologies. The silanization agent: 3-(trimethoxysilyl)propyl methacrylate 98% (3-TMPM); inhibitor: 2,20 -diphenyl1-picrylhydrazyl (DPPH); monomer: hexyl methacrylate 99% (HEMA); crosslinker: ethylene glycol dimethacrylate 98% (EDMA); porogenic solvents: 1-propanol 99.8%, 1,4-butanediol 99%; and radical initiator: azobisisobutyronitrile, 98.0% (AIBN); were all purchased from Sigma. For LC separation, methanol and acetonitrile (ULC/MS grade) and trifluoroacetic acid (TFA) (HPLC grade) were purchased from Bio-Lab. Formic acid (FA) >98% was purchased from Fisher Scientific. Ammonium acetate buffer (7.5 M), dithiothreitol (DTT), iodoacetamide (IAA), trypsin and ammonium bicarbonate were purchased from Sigma. Ultrapure water (Milli Q water) was obtained from a Direct-Q 3 Ultrapure Water System (Millipore). Commercially available proteins were used as standards (see Table 1) following purification by size exclusion chromatography, using a Superdex 200 column (GE Healthcare), with 200 mM ammonium acetate. 3.2. Monolithic column synthesis The fused-silica capillary was pretreated by etching with 1 M NaOH at 100 °C for 3 h, followed by 1 M HCl at 70 °C for 3 h. The capillary was then washed with Milli Q water and methanol until a neutral pH was achieved. Silanization was carried out by treating the inner wall of the capillary with a solution of 3-(trimethoxysilyl)propyl methacrylate (3-TMPM) in neat toluene [30/70 (v/v)] containing 0.005% of the inhibitor 2,20 -diphenyl-1-picrylhydrazyl (DPPH). The silanized capillary was washed with toluene and methanol, and dried by a nitrogen stream. For preparation of the HEMA monolithic column, a polymerization mixture was prepared by mixing 1-propanol (48.5% (v/v)), 1,4butanediol [41.6% (v/v)], Milli Q water [9.9% (v/v)] with HEMA [40% (w/w), EDMA (60% (w/w)] and AIBN (1% w/w to the monomers)giving a ratio of monomers to porogenic solvents of 60/40. The mixture was degassed by a stream of nitrogen before it was loaded into an 18-cm pretreated capillary using a syringe pump (Harvard Apparatus) at a flow rate of 10 ll/min. The capillary was sealed at both ends with two silicon rubber caps, and incubated at 50 °C for 24 h. The monolithic capillary column thus obtained was then connected to a nanoAcquity UPLC system (nUPLC) (Waters Corp.), and

272

S. Rozen et al. / Methods 59 (2013) 270–277

B

D

H

C

E A

F G Fig. 1. An overview of the experimental setup. The protein complex (A), is separated into its composing subunits using a monolithic column on a nUPLC system (B). The nUPLC is coupled to a NanoMate robot, which splits the column effluent into two fractions (C). A small portion of the flow is directed on-line into the mass spectrometer, for mass measurements of intact proteins (D). The remaining flow is fractionated by the NanoMate into a 96-well plate (E), for trypsin digestion (F) and LC–MS/MS proteomic analysis (G). Based on the elution profile, the identity of the subunits is correlated with their accurate mass (H).

Table 1 Protein standards used in this study. Protein

Acronym

Cat. No.

MW (kDa)

Cytochrome C, from bovine heart Lysozyme, from chicken egg whites Carbonic anhydrase, from bovine erythrocytes Alcohol dehydrogenase, from S. cerevisiae Albumin, from bovine serum

CytC Lys CAD

Sigma, C3131 Sigma, L6876 Sigma, C7025

12.2 14.3 29.0

ADH

Sigma, A3263– 15 kU Sigma, A7906

36.8

BSA

brated over 48 h with solvent B (acetonitrile, 0.05% FA, 0.035% TFA), followed by a 10 min wash with solvent A (Milli Q water, 0.05% FA, 0.035% TFA) at 10 ll/min. After equilibration, separation was achieved using a linear gradient of 20% to 60% solvent B over 40 min, at a flow rate of 15 ll/min at room temperature. For separation of the mixture of standard proteins, we loaded 5 pmol of each protein, in a total volume of 5 ll. These same LC conditions were then used to separate the components of the CSN complex, following injection of 5 ll at 1.9 lM into the column.

66.4

3.5. NanoMate parameters washed with 100 column volumes of acetone and 20 column volumes of methanol at a flow rate of 10–50 ll/min, to remove unreacted monomers and other residuals. 3.3. Purification of the CSN complex The endogenous COP9 signalosome complex (CSN) was extracted from human erythrocytes, according to a previously published protocol [26]. The activity of the complex was validated by a deneddylation assay, in which the ability of the complex to deconjugate Nedd8 from S. pombe HA-Cullin1, in protein extracts from a CSN1-null strain, was examined. The final sample concentration of the complex was 3.75 lM at 20 mM Tris–HCl, 50 mM KCl, and 10% glycerol. 3.4. Liquid chromatography conditions Microcapillary reverse phase liquid chromatography (LC) was performed with a nUPLC system, using the monolithic column prepared in-house. About 3 lg of protein can be loaded onto this type of column, a quantity sufficient for running both the ESI-MS and proteomic analyses. Prior to initial use, the column was equili-

The Triversa NanoMate (Advion) was coupled to the nUPLC, as well as to the mass spectrometer ESI-ionization source. The splitting settings of the NanoMate were programmed to divide the eluted flow, such that 5 ll/min were directed on-line to the mass spectrometer (Qstar XL, MDS Sciex) and 10 ll/min were fractionated into a 96-well plate (Cat. No. AB-1000, Thermo Fisher Scientific). Fused silica capillaries of 50–120 lm inner diameter were fitted to maintain the chromatographic separation. In addition, the NanoMate was set to collection mode, enabling time-based fractionation of the eluted flow at 60 s intervals. The fractions were collected into a 96-well plate, pre-filled with 10 ll of Milli Q water, and stored at 4 °C in the NanoMate device, to prevent sample adsorption and evaporation. 3.6. Proteomic analysis To reduce acetonitrile levels and enable tryptic digestion, protein-containing fractions (20 ll) were transferred into 0.5 ml microcentrifuge tubes, diluted with 25 ll of 100 mM ammonium bicarbonate, and partially dried in a SpeedVac. This step was repeated and the sample evaporated, until a final volume of 10 ll was reached. Prior to enzymatic digestion by trypsin, the samples were reduced with DTT, and alkylated with iodoacetamide. Briefly,

S. Rozen et al. / Methods 59 (2013) 270–277

reduction was carried out at 56 °C for 1 h, to a final concentration of 10 mM DTT. Iodoacetamide was then added to a final concentration of 55 mM, and the mixture was incubated at room temperature for 30 min. The reaction was quenched with 10 mM DTT. Tryptic digestion was carried out overnight at 37 °C, at an enzyme-to-protein ratio of 1:20 (wt/wt.) All samples were partially dried in a SpeedVac. The resulting peptide mixtures were diluted in 80% formic acid and immediately diluted 1:10 with water, prior to analysis by on-line C18 nanoLC–MS/MS. Reverse phase LC was then performed, using a 15 cm fused-silica capillary column (inner diameter, 75 lm) prepared in-house, and packed with 3 lm ReproSil-Pur C18AQ media (Dr. Maisch, GmbH) using an UltiMate 3000 Capillary/Nano LC System (LC Packings, Dionex). The LC system was used in conjunction with an LTQ Orbitrap XL (Thermo Fisher Scientific), operated in the positive ion mode and equipped with a nano-ESI ion source. Peptides were separated with a 50-min gradient from 5% to 65% acetonitrile (buffer A, 5% acetonitrile, 0.1% formic acid and 0.005% TFA; buffer B, 90% acetonitrile, 0.2% formic acid and 0.005%TFA). A voltage of 1.2 kV was applied to the union to produce an electrospray. The mass spectrometer was operated in the data-dependent mode. Survey mass spectrometry scans were acquired in the Orbitrap, with the resolution set to a value of 60,000. Up to six of the most intense ions per scan were fragmented and analyzed in the linear ion trap. For peptide analysis, survey scans were recorded in the Fourier transform mode, followed by data-dependent, collision-induced dissociation of the six most intense ions in the linear ion trap. Raw data files were searched with the MASCOT (Matrix Science) software against an NCBInr database.

3.7. Mass spectrometric analysis of intact subunits Eluted subunits were analyzed on-line by a Qstar XL mass spectrometer, using the following experimental parameters: capillary 5.8 kV, declustering potential of 40 V, focusing potential of 200 V, and second declustering potential of 20 V. The mass range was defined as 500–5000 m/z. Spectra were calibrated using a solution of Reserpine (1 lM). Minimal smoothing and centering parameters were used. The measured MW was compared to the reported mass in the neXtProt databases.

3.8. MS and MS/MS analysis of the intact CSN complex For analysis of the human endogenous CSN complex, mass spectra were acquired on a QToF Q-STAR Elite instrument (MDS Sciex), modified for improved transmission of large non-covalent complexes [27]. Prior to mass spectrometry, 1 ml of the CSN sample was concentrated and buffer-exchanged into 0.5 M ammonium acetate (pH 7.5), using Amicon centrifugal filter units (0.5 ml, 3 kDa CO; Millipore), to 0.6 lM. Aliquots (2 ll) were introduced via nanoflow capillaries prepared in-house [28]. The conditions within the mass spectrometer were adjusted to preserve non-covalent interactions. To improve desolvation, mass spectra were recorded in the presence of 10% acetonitrile and 10% methanol. The mass spectrometer was operated at a capillary voltage of 1180 V, a declustering potential of 180 V, a focusing potential of 200 V, and a second declustering potential of 25 V. For MS/MS analysis, the 38+ charge state at 9215 m/z value was selected in the quadrupole, and argon was used as a collision gas at maximum pressure, with a collision energy ranging from 100 to 240 V. The mass spectrometer was calibrated using a solution of cesium iodide (50 mg/ ml). Spectra are shown here with minimal smoothing, and without background subtraction.

273

4. Results and discussion 4.1. Characterization of the HEMA monolithic column The experimental system described in this manuscript was initially established using a commercial PepSwift 500 lm  5 cm monolithic column (Cat. No. 164087, Thermo Fisher Scientific, Inc.). While this column enabled efficient and reproducible separation of protein subunits, its lifetime was short, and its costs were high. We therefore decided to polymerize our own capillary monolithic column from a continuous methacrylic stationary phase. The quality of the column was assessed by examining its morphology, by means of scanning electron microscopy. The column displayed a uniform monolithic matrix possessing an interconnected structure, with continuous porous channels through the monolithic bed (Fig. 2). The HEMA matrix was well attached to the inner wall of the capillary, demonstrating that the silanization process succeeded in anchoring the polymer to the column wall. In terms of LC performance, the column proved to be as efficient as the commercially available column, but far more cost-effective. The preparation of the column proved to be fast and straightforward, with the added benefit that its length and diameter can be readily customized to suit the specific protein complex being analyzed. 4.2. Efficiency of subunit separation The efficiency of the chromatographic separation was optimized using a mixture of five commercially available proteins with a mass range of 12–66.5 kDa, that mimic the size of subunits present in large protein complexes. Validation and optimization was performed over a range of parameters, including loading amount, flow rate, and elution gradient. We then coupled the LC system to the NanoMate robot, to enable the splitting and fractionation of the flow. At this stage, particular attention was paid to the length and width of the tubing succeeding the column, with the aim of preventing sample diffusion and low hydraulic resistance. The NanoMate settings were carefully adjusted to maintain system pressure, and avoid sample flow into the waste. In particular, the tubing size and geometry of the NanoMate fraction collector (entrance, splitting, reference, collector’s input and exit to MS and waste) were optimized. Likewise, the delay time between spectra acquisition and fraction collection was adjusted, to sustain the separation reliability, and enable correct subunit identification. Notably, in our set-up, we used the ESI source of the mass spectrometer rather than the NanoMate’s integrated chip-based electrospray ionization technology, as in our hands, the latter highly compromised signal intensity. Overall, under these experimental conditions, we could accurately record spectra and measure the average masses of each of the standard proteins (Fig. 3), as well as undertake proteomic identification of the collected fractions. 4.3. Characterizing the subunit composition of the human CSN complex We then tested our approach on the endogenous COP9 signalosome (CSN) complex, isolated at its natural expression level directly from human red blood cells. This evolutionarily conserved, multifunctional complex is involved in regulating the ubiquitin– proteasome pathway (for reviews, see [29–32]). The CSN contains eight subunits, denoted CSN1–8, in order of decreasing molecular weight. PTMs of CSN subunits, especially phosphorylations, have been identified [33–36], and for several CSN subunits, multiple isoforms have been reported (documented in neXtProt, http:// www.nextprot.org). Given these observations, we sought to

274

S. Rozen et al. / Methods 59 (2013) 270–277

Fig. 2. Scanning electron microscopy images of the polymerizied HEMA monolithic column. Different magnifications of the column body: (A) 2000 and (B) 10000.

Fig. 3. LC separation and mass measurement of intact proteins. A protein mix containing cytochrome C, lysozyme, two isoforms of alcohol dehydrogenase, carbonic anhydrase, and bovine serum albumin, was separated on a monolithic column, followed by on-line MS analysis and simultaneous fractionation. Peaks representing the eluted proteins are highlighted in color, and correlated with the recorded ESI-MS spectra.

determine which CSN subunit variants could be integrated within the active complex. Following the purification of the complex and confirmation of its activity, we correlated between the identity and mass of the CSN subunits. The denaturing conditions of the LC run induced the disassembly of the complex into its constituent subunits, which were then separated chromatographically, using the prepared monolithic column. The on-line direct spray of the sample into the ESI enabled the mass measurement of the eluted subunits, and the time-based fractionation facilitated subsequent peptide sequencing and subunit identification. All eight subunits comprising the CSN complex could be identified, including the two different isoforms of CSN7, CSN7a and CSN7b [37] (Fig. 4A). Although CSN1, CSN4 and CSN6 co-eluted, it was possible to differentiate between the three subunits by analyzing different regions of the peak. Our results indicate that multiple forms co-exist for almost every CSN subunit, revealing the heterogeneity of the assembled complex.

To illustrate the applicability of the technique, results obtained for CSN3, CSN4 and CSN8 are described, as follows: To minimize experimental variations and increase the reliability of the results, we took into account both technical and biological variations. Three different CSN preparations were used, and at least two technical replicates. Specified masses are an average of both technical and biological measurements taking into consideration both instrument mass accuracy errors and biological variation between human samples. Altogether, we could conclude that three variants are obtained for both CSN3 and CSN4; however, biological variation in the abundance of the three forms exists (Fig. 4B). One variant, which did not appear in all runs for CSN4 (exemplified in Fig. 4B), corresponds to the expected mass caused by the removal of the first methionine residue (131 Da), followed by acetylation (+42 Da). Such an N-terminal modification has previously been reported for both CSN3 and CSN4 [33]. The second and third isoforms are consistent in mass to a similar N-terminal modification, in addition to single and double sites of phosphorylation (+80 and +160 Da,

S. Rozen et al. / Methods 59 (2013) 270–277

275

Fig. 4. Correlating between accurate mass and sequence identity of the human CSN complex. (A) The complex was separated into its composing subunits, using a monolithic column under denaturing conditions. Peaks representing the eluted CSN subunits are highlighted in color. (B) The resulting ESI-QToF (right panel) and an example of one C18 LC–MS/MS Orbitrap mass spectra (left panel) of CSN3, CSN4 and CSN8 are shown. The ESI-QToF spectra enable the intact mass of the subunit to be determined, while the LC– MS/MS, performed subsequent to tryptic digestion and in combination with a database search, enables protein identification. Indicated masses are an average of biological and technical measurements. Statistical mass errors were calculated by taking the square root of the sum of the square of individual errors. Subunit variants are differentiated by labeling with circles and triangles. LC-MS/MS peaks matching the theoretical fragments ions are labeled as y ions (blue), b ions (red) and water loss for b ions (green).

respectively), in agreement with the multiple phosphorylation sites that have been documented for these two subunits [33,35,38–41]. For CSN8, we could identify two different series of peaks in the ESI-MS spectrum, corresponding in mass to 23,094 ± 4 and 22,638 ± 2 Da, respectively (Fig. 4B). While the mass of the heavier CSN8 variant can be easily explained by the known removal of the first methionine [33], the mass of the lighter CSN8 form was inconsistent with any of the reported CSN8 isoforms. Our examination of the CSN8 sequence raised the possibility of an alternative translation initiation site at Met6, a possibility that was previously suggested for both mouse and human CSN8 [42]. Calculating the mass of CSN8, while taking into account

the second initiation site, fits nicely with the experimental mass yielded by removal of the first methionine and N-terminal acetylation. Moreover, according to the relative intensity of the two forms, it could be determined that only about 25% of the CSN8 integrated within the CSN complex corresponds to the shorter form of the subunit, while the majority of CSN8 corresponds to the full-length protein. Overall, the intact molecular weight measurements of the individual CSN subunits, coupled with their identification, enabled us to reveal the array of subunit variants incorporated into the CSN complex. How such subunit variability affects CSN functionality is an intriguing question, yet to be answered.

276

S. Rozen et al. / Methods 59 (2013) 270–277

B

A

100

%

+38

0

D

8000

11000

m/z

+13

100

C

%

+13 +13

+38

100

F +23

0

m/z 1950

1600 1600

100

E

+22 +14

+21

%

0

0

m/z 2400

2000

2000

4000

Csn8 Csn8’

+14

%

%

100

6000

8000

0

10000

12000

12000

Csn3 Csn4

Csn8

24000

14000

16000

18000

Csn3/ Csn4 +

20000

22000

m/z

m/z

Csn8

Csn3/ Csn4

Fig. 5. Structural MS analysis of the human CSN complex. (A) The subunit interaction map of the CSN complex, generated by analysis of the recombinant complex (see [22]). (B) The MS spectrum of the endogenous, intact CSN complex. The 38+ charge state, highlighted in yellow, was subjected to MS/MS analysis in the presence of 10% methanol and 10% acetonitrile (C). The recorded data shows dissociation of the peripherally located CSN8 isoforms (D), and CSN3 and CSN4 subunits (E). The high m/z series of peaks are assigned to the corresponding stripped complexes (F).

4.4. Subunit composition characterization as the basis for structural MS analysis Correlating between subunit sequence identity and mass not only exposes the compositional variability of the protein assembly, but also provides the basis for structural analysis of the intact complex [25]. At the stage where the ‘‘building blocks’’ of the complex are identified, and the unique mass (or masses) of each subunit is (are) defined, it is possible to study the non-covalent interactions between subunits. Various approaches are available for such analysis, including cross-linking [43], peptide array [44] and point mutations analysis [45]. Here we demonstrate the application of structural MS for the study of the intact CSN. This method enables determination of both subunit stoichiometry, and the architecture of the protein complex [23–25] (Fig. 5A). An electrospray mass spectrum of the human endogenous CSN complex, recorded under conditions that maintain the non-covalent interactions, is shown in Fig. 5B. The series of peaks are assigned to the intact complex with a mass of 337,406 ± 26 Da. As the masses of the individual subunits have been identified (see Section 4.2), it is possible to unambiguously conclude that all eight subunits are present at unit stoichiometry. This finding is in accord with our previous studies of the recombinant complex [22] (Fig. 5A). Applying the tandem mass spectrometry approach to the intact CSN complex induces the dissociation of subunits

CSN3, CSN4 and CSN8, ensuring their expected peripheral position (Fig. 5). Interestingly, only the unphosphorylated forms of CSN3 and CSN4 dissociated from CSN, suggesting that phosphorylation enhances the interactions of these subunits within the complex. Both forms of CSN8, which result from the Met1 and Met6 initiation sites, were expelled from the complex and their relative abundance was approximately 3:1, Met1:Met6, as revealed by the monolithic separation. This finding implies that the extreme N-terminal region of CSN8 does not contribute to the strength of the interaction with the other CSN subunits. Overall, by combining the two MS approaches, monolithic column-based analysis, and structural MS, insights into the contributions of the different CSN isoforms to the stability and integrity of the complex can be determined.

5. Conclusions The method we describe here enables exposure of the diversity of subunit variants, by correlating between the identity of protein subunits, and their accurate mass. The advantage of this approach is that subunit separation prior to proteomic analysis increases the odds of full sequence coverage and PTM identification. In parallel, accurate mass measurements enable clarification of the co-existing

S. Rozen et al. / Methods 59 (2013) 270–277

PTMs, enzymatic cleavages, truncations, and the presence of multiple isoforms or gene products. An additional strength of the method is that it can depict the combinations of PTMs that can occur on a single subunit, exposing the possible interplay between neighboring modifications. It is now becoming clear that the functions of the various PTMs are related to one another, and that the crosstalk between different modifications determines the final biological read-out [46–48]. Within this context, by repeating the analysis described above on protein complexes isolated from different tissues, cellular compartments, cells exposed to various stimuli, or at different time points during the cell cycle, the dynamic cross-regulation caused by the various modifications can be revealed. Acknowledgments A.T. was supported by a Sergio Lombroso Fellowship. M.S. is grateful for the financial support of a Starting Grant from the European Research Council (ERC) under the European Community’s Seventh Framework Programme (FP7/2007-2013)/ERC Grant Agreement no. 239679, and a Grant from the Israel Science Foundation (Grant no. 220/10). M.S. is the incumbent of the Elaine Blond Career Development Chair. References [1] A. Sali, R. Glaeser, T. Earnest, W. Baumeister, Nature 422 (2003) 216–225. [2] M. Yandell, D. Ence, Nat. Rev. Genet. 13 (2012) 329–342. [3] A.C. Gingras, M. Gstaiger, B. Raught, R. Aebersold, Nat. Rev. Mol. Cell Biol. 8 (2007) 645–654. [4] J.S. Taylor, J. Raes, Annu. Rev. Genet. 38 (2004) 615–643. [5] M. Fagnani, Y. Barash, J.Y. Ip, C. Misquitta, Q. Pan, A.L. Saltzman, O. Shai, L. Lee, A. Rozenhek, N. Mohammad, S. Willaime-Morawek, T. Babak, W. Zhang, T.R. Hughes, D. van der Kooy, B.J. Frey, B.J. Blencowe, Genome Biol. 8 (2007) R108. [6] Q. Pan, O. Shai, L.J. Lee, B.J. Frey, B.J. Blencowe, Nat. Genet. 40 (2008) 1413– 1415. [7] E.T. Wang, R. Sandberg, S. Luo, I. Khrebtukova, L. Zhang, C. Mayr, S.F. Kingsmore, G.P. Schroth, C.B. Burge, Nature 456 (2008) 470–476. [8] T.J. Griffin, D.R. Goodlett, R. Aebersold, Curr. Opin. Biotechnol. 12 (2001) 607– 612. [9] M. Mann, O.N. Jensen, Nat. Biotechnol. 21 (2003) 255–261. [10] J.M. Baugh, E.V. Pilipenko, Mol. Cell 16 (2004) 575–586. [11] R. Sachidanandam, D. Weissman, S.C. Schmidt, J.M. Kakol, L.D. Stein, G. Marth, S. Sherry, J.C. Mullikin, B.J. Mortimore, D.L. Willey, S.E. Hunt, C.G. Cole, P.C. Coggill, C.M. Rice, Z. Ning, J. Rogers, D.R. Bentley, P.Y. Kwok, E.R. Mardis, R.T. Yeh, B. Schultz, L. Cook, R. Davenport, M. Dante, L. Fulton, L. Hillier, R.H. Waterston, J.D. McPherson, B. Gilman, S. Schaffner, W.J. Van Etten, D. Reich, J. Higgins, M.J. Daly, B. Blumenstiel, J. Baldwin, N. Stange-Thomann, M.C. Zody, L. Linton, E.S. Lander, D. Altshuler, Nature 409 (2001) 928–933. [12] L.H. Hartwell, J.J. Hopfield, S. Leibler, A.W. Murray, Nature 402 (1999) C47– C52. [13] J. Hanna, D. Finley, FEBS Lett. 581 (2007) 2854–2861. [14] C. Naujokat, D. Fuchs, C. Berges, Biochim. Biophys. Acta 1773 (2007) 1389– 1397. [15] L. Huang, A.L. Burlingame, Methods Enzymol. 405 (2005) 187–236.

277

[16] J.C. Tran, L. Zamdborg, D.R. Ahlf, J.E. Lee, A.D. Catherman, K.R. Durbin, J.D. Tipton, A. Vellaichamy, J.F. Kellie, M. Li, C. Wu, S.M. Sweet, B.P. Early, N. Siuti, R.D. LeDuc, P.D. Compton, P.M. Thomas, N.L. Kelleher, Nature 480 (2011) 254– 258. [17] S. Uttenweiler-Joseph, S. Claverol, L. Sylvius, M.P. Bousquet-Dubouch, O. Burlet-Schiltz, B. Monsarrat, Methods Mol. Biol. 484 (2008) 111–130. [18] S. Wu, N. Tolic, Z. Tian, E.W. Robinson, L. Pasa-Tolic, Methods Mol. Biol. 694 (2011) 291–304. [19] S. Claverol, O. Burlet-Schiltz, J.E. Gairin, B. Monsarrat, Mol. Cell. Proteomics 2 (2003) 483–493. [20] G.E. Reid, S.A. McLuckey, J. Mass. Spectrom. 37 (2002) 663–675. [21] F. Lanucara, C.E. Eyers, Top-down mass spectrometry for the analysis of combinatorial post-translational modifications, Mass Spectrom. Rev. 32 (2013) 27–42. [22] M. Sharon, H. Mao, E. Boeri Erba, E. Stephens, N. Zheng, C.V. Robinson, Structure 17 (2009) 31–40. [23] J.L. Benesch, B.T. Ruotolo, D.A. Simmons, C.V. Robinson, Chem. Rev. 107 (2007) 3544–3567. [24] A.J. Heck, Nat. Methods 5 (2008) 927–933. [25] M. Sharon, J. Am. Soc. Mass Spectrom. 21 (2010) 487–500. [26] B.K. Hetfeld, D. Bech-Otschir, W. Dubiel, Methods Enzymol. 398 (2005) 481– 491. [27] I.V. Chernushevich, B.A. Thomson, Anal. Chem. 76 (2004) 1754–1760. [28] N. Kirshenbaum, I. Michaelevski, M. Sharon, Analyzing large protein complexes by structural mass spectrometry, J Vis Exp, (41), (2012) e1985, http://dx.doi.org/10.3791/1985. [29] G.A. Cope, R.J. Deshaies, Cell 114 (2003) 663–671. [30] C. Schwechheimer, Biochim. Biophys. Acta 1695 (2004) 45–54. [31] N. Wei, G. Serino, X.W. Deng, Trends Biochem. Sci. 33 (2008) 592–600. [32] D.A. Wolf, C. Zhou, S. Wee, Nat. Cell Biol. 5 (2003) 1029–1033. [33] L. Fang, X. Wang, K. Yamoah, P.L. Chen, Z.Q. Pan, L. Huang, J. Proteome Res. 7 (2008) 4914–4925. [34] P. Beli, N. Lukashchuk, S.A. Wagner, B.T. Weinert, J.V. Olsen, L. Baskcomb, M. Mann, S.P. Jackson, C. Choudhary, Mol. Cell 46 (2012) 212–225. [35] S. Matsuoka, B.A. Ballif, A. Smogorzewska, E.R. McDonald 3rd, K.E. Hurov, J. Luo, C.E. Bakalarski, Z. Zhao, N. Solimini, Y. Lerenthal, Y. Shiloh, S.P. Gygi, S.J. Elledge, Science 316 (2007) 1160–1166. [36] S. Uhle, O. Medalia, R. Waldron, R. Dumdey, P. Henklein, D. Bech-Otschir, X. Huang, M. Berse, J. Sperling, R. Schade, W. Dubiel, EMBO J. 22 (2003) 1302– 1312. [37] N. Wei, X.W. Deng, Photochem. Photobiol. 68 (1998) 237–241. [38] N. Dephoure, C. Zhou, J. Villen, S.A. Beausoleil, C.E. Bakalarski, S.J. Elledge, S.P. Gygi, Proc. Natl. Acad. Sci. USA 105 (2008) 10762–10767. [39] S. Gauci, A.O. Helbig, M. Slijper, J. Krijgsveld, A.J. Heck, S. Mohammed, Anal. Chem. 81 (2009) 4493–4501. [40] K. Rikova, A. Guo, Q. Zeng, A. Possemato, J. Yu, H. Haack, J. Nardone, K. Lee, C. Reeves, Y. Li, Y. Hu, Z. Tan, M. Stokes, L. Sullivan, J. Mitchell, R. Wetzel, J. Macneill, J.M. Ren, J. Yuan, C.E. Bakalarski, J. Villen, J.M. Kornhauser, B. Smith, D. Li, X. Zhou, S.P. Gygi, T.L. Gu, R.D. Polakiewicz, J. Rush, M.J. Comb, Cell 131 (2007) 1190–1203. [41] S. Sui, J. Wang, B. Yang, L. Song, J. Zhang, M. Chen, J. Liu, Z. Lu, Y. Cai, S. Chen, W. Bi, Y. Zhu, F. He, X. Qian, Proteomics 8 (2008) 2024–2034. [42] K. Lykke-Andersen, N. Wei, Gene 321 (2003) 65–72. [43] A. Sinz, Mass Spectrom. Rev. 25 (2006) 663–682. [44] C. Katz, L. Levy-Beladev, S. Rotem-Bamberger, T. Rito, S.G. Rudiger, A. Friedler, Chem. Soc. Rev. 40 (2011) 2131–2145. [45] D. Reichmann, O. Rahat, M. Cohen, H. Neuvirth, G. Schreiber, Curr. Opin. Struct. Biol. 17 (2007) 67–76. [46] X.J. Yang, E. Seto, Mol. Cell 31 (2008) 449–461. [47] T. Kouzarides, Cell 128 (2007) 693–705. [48] J.A. Latham, S.Y. Dent, Nat. Struct. Mol. Biol. 14 (2007) 1017–1024.