An Integrated Mass Spectrometry-based Proteomic

Research

An Integrated Mass Spectrometry-based Proteomic Approach QUANTITATIVE ANALYSIS OF TANDEM AFFINITY-PURIFIED IN VIVO CROSS-LINKED PROTEIN COMPLEXES (QTAX) TO DECIPHER THE 26 S PROTEASOME-INTERACTING NETWORK*

Cortnie Guerrero‡§, Christian Tagwerker¶储, Peter Kaiser¶, and Lan Huang‡§** We developed an integrated proteomic approach to decipher in vivo protein-protein interactions and applied this strategy to globally map the 26 S proteasome interaction network in yeast. We termed this approach QTAX for quantitative analysis of tandem affinity purified in vivo cross-linked (X) protein complexes. For this work, in vivo formaldehyde cross-linking was used to freeze both stable and transient interactions occurring in intact cells prior to lysis. To isolate cross-linked protein complexes with high purification efficiency under fully denaturing conditions, a new tandem affinity tag consisting of a hexahistidine sequence and an in vivo biotinylation signal was adopted for affinity-based purification. Tandem affinity purification after in vivo cross-linking was combined with tandem mass spectrometry coupled with a quantitative SILAC (stable isotope labeling of amino acids in cell culture) strategy to carry out unambiguous protein identification and quantification of specific protein interactions. Using this method, we captured and identified the full composition of yeast 26 S proteasome complex as well as the two known ubiquitin receptors, Rad23 and Dsk2. Quantitative mass spectrometry analysis allowed us to distinguish specific proteasome-interacting proteins (PIPs) from background proteins and led to the identification of a total of 64 potential PIPs of which 42 are novel interactions. Among the 64 putative specific PIPs, there are ubiquitin pathway components, ubiquitinated substrates, chaperones, and transcription and translation regulators, demonstrating the efficacy of the developed approach in capturing in vivo protein interactions. The method offers an advanced technical approach to elucidate the dynamic protein interaction networks of the proteasome and can find a wide range of applications in the studies of other macromolecular protein complex interaction networks. Molecular & Cellular Proteomics 5: 366 –378, 2006.

From the Departments of ‡Physiology and Biophysics, §Developmental and Cell Biology, and ¶Biological Chemistry, University of California, Irvine, California 92697 and the 储Institute of Biochemistry, University of Innsbruck, A-6020 Innsbruck, Austria Received, September 14, 2005, and in revised form, October 27, 2005 Published, MCP Papers in Press, November 11, 2005, DOI 10.1074/mcp.M500303-MCP200

366

Molecular & Cellular Proteomics 5.2

Protein ubiquitination and proteasome-mediated protein degradation are essential to the regulation of many important biological processes such as cell cycle progression, apoptosis, DNA repair, chromosome maintenance, transcriptional activation, metabolism, immune response, signal transduction, stress response, cell differentiation, etc. (1–3). Disruption of normal ubiquitin-proteasome degradation pathways has been implicated in a wide range of human disease. Therefore, the proteasome and ubiquitination components represent a class of “druggable” targets for pharmaceutical intervention. The 26 S proteasome, the macromolecular degradation machine of the ubiquitin-proteasome pathway, consists of a self-compartmentalized 20 S protease core that is capped at one or both ends by the 19 S regulatory particle, or CAP (also known as PA700 in animal cells) (3– 6). The 20 S core particle, responsible for various proteolytic activities, is made up of two copies each of seven different ␣ and seven different ␤ subunits arranged into four stacked rings (␣7␤7␤7␣7). The two outer ␣ rings are catalytically inactive, whereas three of the seven inner ␤ subunits are catalytically active. Although the 20 S core can degrade fully unfolded proteins in the absence of ATP and ubiquitin, protein degradation by 26 S proteasomes is strictly ATP-dependent and, in almost all cases, requires the presence of a ubiquitin chain on the substrate protein (7). The 19 S regulatory complex is composed of at least 18 different subunits, which are assembled into two main subcomplexes: a base that contains six ATPases plus two non-ATPase subunits and abuts the proteasome ␣ ring and a lid subcomplex containing 10 non-ATPase subunits that sits on top of the base (4). Additional proteasome subunits continue to be identified because of the development of new protein purification and identification techniques (7–10), yet how they fit into the assembly of 19 S has not been determined. There are two major steps involved in the ubiquitin-proteasome-dependent degradation pathway: 1) enzymatic polyubiquitination of protein substrates and 2) transportation to and recognition by the 26 S proteasome prior to degradation. A cascade of enzymes, including the ubiquitin-activating enzyme (E1),1 ubiquitin-conjugating enzymes (E2s), and ubiq-

1 The abbreviations used are: E1, ubiquitin-activating enzyme; E2, ubiquitin-conjugating enzyme; E3, ubiquitin ligase; Ub, ubiquitin; PIP, proteasome-interacting protein; HB, histidine and biotin; SILAC, sta-

© 2006 by The American Society for Biochemistry and Molecular Biology, Inc. This paper is available on line at http://www.mcponline.org

QTAX for Deciphering the 26 S Proteasome-interacting Network

uitin ligases (E3s), are responsible for ubiquitination of a target protein (11). Once a protein is polyubiquitinated, it is generally targeted to and degraded by the 26 S proteasome. Little is known about how ubiquitinated substrates arrive at the proteasome. Proteins with ubiquitin-like and ubiquitin-associated domains, such as Rad23 and Dsk2, have been shown to function as polyubiquitin receptors, which recognize a group of ubiquitinated substrates and translocate them to the 26 S proteasome for degradation (12–15). In addition, one of the 19 S subunits, Rpn10, can bind to multiubiquitin chains (16) and function as a ubiquitin receptor (13, 14). However, because Rpn10 is dispensable for the growth of yeast cells (17), the existence of other ubiquitin-binding proteins in the 19 S complex is likely. Recently in vitro cross-linking experiments on the purified 26 S proteasome suggested that the non-ATPase subunits Rpn1 and Rpn2 can bind to proteins with ubiquitinlike domains (18), and Rpt5/S6, one of the six ATPase subunits, can specifically contact a proteasome-bound polyubiquitin chain (19). However, whether they function as Ub receptors in vivo is not clear. Because the known ubiquitin receptors are only responsible for a subgroup of substrates, additional ubiquitin receptors or pathways have been suggested for targeting various classes of ubiquitinated substrates (14, 15), but their identities remain elusive. Mass spectrometry-based interactive proteomics has evolved as a powerful tool for mapping proteome-wide protein interaction networks (20 –22). Combined with affinity purification, mass spectrometric analyses have identified a number of proteasome-interacting proteins (PIPs) including new proteasome subunits (7, 8); however, many known interactions were not identified by this method including the known Ub receptors Rad23 and Dsk2. This is most likely due to the fact that many in vivo interactions are transient and/or low affinity and generally dependent on the specific cellular environment in which they occur. Therefore, a key strategy of our new approach is to effectively capture and identify these weak interacting proteins in vivo. Chemical cross-linking stabilizes interactions through covalent bond formation, allowing the detection of protein-protein interactions in native cells or tissues that are weak and/or transient. The most commonly used cross-linking reagent is formaldehyde, a water-soluble and cell membrane-permeable molecule, which provides a fast and reversible cross-linking reaction (23). This method has been widely used for the study of protein-DNA and protein-protein interactions (23–30). Recent reports demonstrated that formaldehyde cross-linking/mass spectrometry strategies are effective in capturing and identifying in vivo ble isotope labeling of amino acids in cell culture; TAP, tandem affinity purification; TEV, tobacco etch virus; TEB, TEV elution buffer; co-IP, co-immunoprecipitation; L, light ([12C6]Arg); H, heavy ([13C6]Arg); L/H, relative abundance ratio of light ([12C6]Arg) to heavy ([13C6]Arg); QTAX, quantitative analysis of tandem affinity-purified in vivo cross-linked (X) protein complexes; 1-D, one-dimensional; Ni-NTA, nickel-nitrilotriacetic acid.

protein-protein interactions including low affinity and transient interactions (25, 26, 28). In this work, we describe an integrated approach, QTAX, for quantitative analysis of tandem affinity-purified in vivo cross-linked (X) protein complexes and report the application of this strategy to decipher the 26 S proteasome interaction network. MATERIALS AND METHODS

Chemicals and Antibodies—All general chemicals for buffers and culture media were purchased from Fisher Scientific and/or VWR International. [13C6]Arginine was purchased from Cambridge Isotope Laboratories, Inc. (Andover, MA), and other amino acids were purchased from Sigma. RGS-His antibody was obtained from Qiagen (Germantown, MD), and ImmunoPure streptavidin, horseradish peroxidase-conjugated antibody, and Super Signal West Pico chemiluminescent substrate were from Pierce. Trypsin was purchased from Promega Corp. (Madison, WI), and endoproteinase Lys-C was from Wako chemicals (Osaka, Japan). Yeast Strains and Culture Conditions—Standard yeast growth media and conditions were used (31). All strains used in this study, except the commercial TAP-tagged strains (Open Biosystems, Huntsville, AL), are isogenic to 15Daub⌬, a bar1⌬ ura3⌬ns, a derivative of BF264-15D (32). Rpn11 and Rpt5 were tagged with the HB tag2 at their C terminus at the chromosomal locus by a single step PCR strategy (33). These strains were used for all experiments performed in YEPD medium (1% yeast extract, 2% peptone, 2% dextrose). For the SILAC experiments, arginine auxotroph strains were constructed by deleting ARG4 with a hygromycin resistance marker following a PCR-based strategy (34). For the initial experiments, the Rpn11-HB strain was grown in 100 –500 ml of YEPD medium at 200 rpm at 30 °C to an A600 of ⬃1.5. The strains used in the SILAC experiments were grown in 500 –1000 ml of synthetic complete media supplemented with 20 mg/liter either heavy arginine (wild type, untagged strain) or light arginine (Rpn11-HB strain) to a final A600 of ⬃0.9 before crosslinking experiments and subsequent purifications. In Vivo Formaldehyde Cross-linking—To identify the optimal crosslinking conditions, different concentrations of formaldehyde (1–3%) were directly added to the yeast cell culture for various incubation times (10 –30 min), and cells were incubated at 30 °C. The crosslinking reaction was quenched for 10 min at 30 °C by addition of 2.5 M glycine with a final concentration of 0.125 M. After cross-linking, cells were collected, washed with ice-cold water, and frozen at ⫺80 °C prior to lysis. It was determined that 1% formaldehyde for 10 min at 30 °C was the optimal cross-linking condition, which was used in all subsequent experiments. Tandem Affinity Purification of Cross-linked PIPs—Frozen cells were lysed by bead beating in lysis buffer (8 M urea, 300 mM NaCl, 50 mM NaH2PO4, 0.5% Nonidet P-40, 1 mM PMSF, pH 8; 1 ml of lysis buffer per 100 ml of cultured cells), and cellular debris were removed by centrifugation. Clarified lysate was initially incubated with NiSepharose 6 Fast Flow beads (Amersham Biosciences). After binding, the beads were washed twice with 20 bed volumes of lysis buffer, pH 8, twice with 20 bed volumes of lysis buffer, pH 6.3, and once with 20 bed volumes of lysis buffer, pH 6.3 ⫹ 10 mM imidazole to remove nonspecifically bound proteins. The proteins were eluted from the Ni-Sepharose beads using 10 bed volumes of lysis buffer, pH 4.3. The eluate was brought to pH 8 with 1 M Tris, pH 8 and applied directly to immobilized streptavidin beads (Pierce, Immunopure immobilized streptavidin). To remove any remaining contaminants, streptavidin

2

C. Tagwerker, K. Flick, M. Cui, C. Guerrero, Y. Dou, B. Auer, P. F. Baldi, and L. Huang, submitted for publication.


367


beads were washed stringently using 10 bed volumes each of buffer 2 (8 M urea, 0.2 M NaCl, 0.2% SDS, 100 mM Tris, pH 7.5) and buffer 3 (buffer 2 containing 2% SDS). Residual urea and SDS were removed using buffer 4 (0.2 M NaCl, 100 mM Tris, pH 7.5). Both binding reactions were allowed to proceed overnight at room temperature. Gel Electrophoresis and Immunoblotting—Cell lysates, wash fractions, and elution fractions were separated by SDS-PAGE (7 or 10% gel). Proteins were transferred to a PVDF membrane and analyzed by immunoblotting. The RGS-His6 antibody was used at a 1:2000 dilution (Qiagen). Biotinylated proteins were detected using a streptavidin-horseradish peroxidase conjugate (1:5000). Quantitative Analysis Using SILAC—Arginine auxotroph strains were grown in synthetic complete media supplemented with heavy arginine (wild type, untagged strain) or light arginine (Rpn11-HB strain). Each strain was initially grown to stationary phase overnight (5–20-ml culture) and then transferred to a larger culture (100 –200 ml) until reaching early stationary phase. The culture was then transferred to a final volume of either 500 ml or 1 liter and incubated until reaching A600 of ⬃0.9. Finally the cells were in vivo cross-linked, collected, lysed, clarified, and purified as described above. Equal amounts of lysate from the heavy and light labeled cultures were combined prior to tandem affinity purification. LC MS/MS—At the last step of the tandem affinity purification, the proteins were bound on streptavidin beads and, due to extremely high affinity binding, could not be eluted. Therefore, direct on-bead digestion was performed. For direct trypsin digestion, the urea buffer was replaced with 50 mM NH4HCO3, and the proteins were digested overnight at 37 °C. For Lys-C/trypsin digestion, Lys-C was directly added to the 8 M urea buffer and streptavidin beads for 4 h at 37 °C, and then the urea concentration was reduced to ⬍2 M for trypsin digestion overnight. After digestion, the tryptic peptides were extracted from the streptavidin beads with 25% acetonitrile, 0.1% formic acid three times. The extracts were pooled, concentrated using a SpeedVac, and acidified by 0.1% formic acid prior to mass spectrometric analysis. For one-dimensional (1-D) LC MS/MS analysis, the tryptic digests were directly injected onto the column, whereas the Lys-C/trypsin digests were desalted first with C18 Ziptips (Millipore) before analysis. 1-D LC MS/MS was carried out by nanoflow reverse phase LC (Ultimate, LC Packings, Dionex) coupled on line to a quadrupole-orthogonal-time-of-flight tandem mass spectrometer (QSTAR XL, Applied Biosystems/MDS Sciex). Reverse phase LC was performed using a PepMap column (75-␮m inner diameter ⫻ 150-mm long, LC Packings, Dionex), and the peptides were eluted using a linear gradient of 0% B to 35% B in 100 min at a flow of 200 nl/min. Solvent A contained 98% H2O, 2% acetonitrile, 0.1% formic acid, whereas solvent B was composed of 98% acetonitrile, 2% H2O, 0.1% formic acid. The QSTAR mass spectrometer was operated in an information-dependent mode in which each full MS scan was followed by three MS/MS scans where the three most abundant peptide molecular ions were dynamically selected for CID, thus generating tandem mass spectra. In general, the ions selected for CID were the most abundant in the MS spectrum except that singly charged ions were excluded and dynamic exclusion was used to prevent repetitive selection of the same ions within a preset time. Collision energies were programmed to be adjusted automatically according to the charge state and mass value of the precursor ions. To increase the number of MS/MS spectra acquired from any given sample and improve the dynamic range of mass spectrometric analysis, multiple LC MS/MS runs were performed on the same sample with the exclusion lists (i.e. the m/z list of the ions being sequenced from the previous runs) generated from the previous LC MS/MS runs using Mascot script within the Analyst program. For the two-dimensional LC MS/MS, the digests were first separated by strong cation exchange chromatography, which was per-

368


formed using an AKTA system (Amersham Biosciences). Solvent A (25% acetonitrile, pH 3 adjusted with formic acid) and solvent B (solvent A with 400 mM NH4Cl) were used to develop a salt gradient. The digests were separated using a 1-mm ⫻ 15-cm polysulfoethyl A column (Poly LC, Columbia, MD) at a flow rate of 90 ␮l/min. Peptide elution was monitored by UV detection at 215 and 280 nm. A typical separation used 0% B from 0 to 10 min to allow for sample loading and removal of non-peptide species followed by a gradient of 0 –100% B from 10 to 30 min. Fractions were manually collected based on UV absorbance. All of the strong cation exchange fractions were desalted off line using C18 Ziptips (Millipore) prior to LC MS/MS. Database Searching and Abundance Ratio Calculation—The monoisotopic masses (m/z) of both parent ions and their corresponding fragment ions, parent ion charge states (z), and ion intensities from the MS/MS acquired were automatically extracted using the script in the Analyst software and directly submitted for automated database searching for protein identification using two different search engines, Protein Prospector (University of California, San Francisco) and Mascot (Matrix Science), to improve the confidence level of the protein identifications in the large data sets. The LC-Batchtag program within the developmental version of Protein Prospector was used for database searching. The mass accuracy for parent ions and fragment ions were set as ⫾100 and 300 ppm, respectively. An in-house Mascot program was also used for database searching, and the mass accuracy for parent ions was set as ⫾100 ppm, and 0.3 Da was used for the fragment ion mass tolerance. Both Swiss-Prot and National Center for Biotechnology Information non-redundant (NCBInr) public databases were queried to identify the purified proteins because each database contains unique protein entries. In addition, the Search Compare program within the developmental version of Protein Prospector (35) was used to make a list of proteins that differed between samples. For the SILAC experiments, the Search Compare program was also used to calculate the relative abundance ratios of Argcontaining peptides based on either ion intensities of monoisotopic peaks or their areas observed in the LC MS spectra at the time when the peptides were sequenced and subsequently identified during database searching. The proteins identified by one or two peptides were confirmed by manual inspection of the MS/MS spectra. The relative abundance ratios were also validated by checking the raw spectra. Validation of the Selected Specific PIPs—Several newly identified putative proteasome-interacting partners were validated by co-immunoprecipitation experiments. Yeast strains expressing endogenous levels of TAP-tagged versions of the proteins were purchased (Open Biosystems). Each TAP-tagged strain was grown to an A600 of 1.5 and lysed in a buffer containing 25 mM Tris, pH 7.5, 200 mM NaCl, 0.2% Nonidet P-40, 2 mM DTT, phosphatase inhibitors (50 mM NaF, 0.1 mM Na3VO4, 10 mM Na4P2O7, 5 mM EDTA, 5 mM EGTA), and protease inhibitor complete (Roche Diagnostics). Approximately 5 mg of protein lysate was added to 15 ␮l of antigen affinity gel rabbit IgG (MP Biomedicals) and incubated at 4 °C for 1.75 h. Beads were washed three times with 1 ml of wash buffer (25 mM Tris, pH 7.5, 150 mM NaCl, and 0.2% Triton) and one time with 1 ml of TEB buffer (50 mM Tris, pH7.5, 1 mM EDTA, 1 mM DTT). After washing, 30 ␮l of TEB buffer and 10 units of TEV protease were added, and cleavage was allowed to proceed overnight at 4 °C. Proteins that eluted from the beads were analyzed by immunoblotting with antibodies directed against the 19 S proteasome subunit Rpt6 (GeneTex). A yeast strain expressing a TAP-tagged version of the 20 S subunit Pre1 was used to purify the 26 S proteasome as a positive control. To prevent dissociation of the 19 S proteosome complex from the 20 S core complex, the purification was carried out in the presence of an ATP-regenerating system (7). The TEV eluate from this strain was diluted 1:25 before separation on SDS-PAGE to reduce the signal obtained by immunoblot analysis. A wild type, untagged strain was used as a negative control.


FIG. 1. The QTAX strategy for the purification of proteasome-interacting proteins. A, schematic representation of a proteasome subunit fused to the HB tag at its C terminus. The HB tag shown here consists of a bacterially derived peptide that induces biotinylation in vivo flanked by two hexahistidine tags. One of the hexahistidine (6xhis) tags contains an RGS-His6 epitope that can be used for immunoblot analysis. B, outline of the QTAX strategy for quantitative analysis of tandem affinity-purified in vivo cross-linked protein complexes. Formaldehyde is first added to the cells to cross-link proteins in vivo. The cells are then lysed, and proteins are first purified by Ni2⫹ chelate chromatography. The eluate is subsequently applied to immobilized streptavidin beads for the second purification step. Cell lysis and both purification steps are carried out under fully denaturing conditions to prevent any non-covalent interactions. Proteins bound to streptavidin beads are directly digested with trypsin to release peptides that are analyzed by LC MS/MS, and proteins are identified using automated database searching. 1-D, one-dimensional.

RESULTS

A New Tandem Affinity Tag for Purification of the 26 S Proteasome Complex and Its Interacting Proteins The QTAX strategy for purifying and identifying the 26 S proteasome-interacting network is illustrated in Fig. 1. To use a tandem affinity purification compatible with in vivo crosslinking under fully denaturing conditions, we adapted a new tandem affinity tag (HB tag) that has been developed to specifically purify covalently linked proteins such as ubiquitinated or sumoylated proteins.2 The HB tag consists of a hexahistidine sequence combined with a signal peptide that serves as a biotinylation signal in vivo (36). HB-tagged proteins are efficiently biotinylated in vivo in yeast as well as in mammalian cells (⬎95%) at a specific lysine residue present in the tag. To increase the affinity of the original HB tag to Ni2⫹ chelate resin, a second hexahistidine sequence was added. HBtagged proteins can be purified by Ni2⫹ chelate chromatography followed by binding to streptavidin resin under fully denaturing conditions. To purify the 26 S proteasome and its interacting proteins, two regulatory proteasome subunits, the non-ATPase subunit Rpn11 or the ATPase subunit Rpt5, were selected for HB tagging because these subunits have been successfully used with other tags for purification purposes (7,

8). Rpn11 and Rpt5 were tagged at their C termini at the chromosomal locus so that endogenous levels of the tagged proteins were expressed. The HB-tagged forms of both of the essential subunits Rpn11 and Rpt5 appeared to be fully functional because the growth rate and cell morphology of each tagged strain was indistinguishable from that of an untagged strain (data not shown). As shown in Fig. 1B, prior to cell lysis, in vivo formaldehyde cross-linking was used to freeze the 26 S proteasome interaction networks in yeast cells expressing an HB-tagged proteasome subunit. The cross-linked proteasome complexes were isolated using sequential purification on Ni-NTA resin and immobilized streptavidin under fully denaturing conditions. The purified complexes were subsequently digested with trypsin, and their composition was identified by peptide sequencing using LC MS/MS analysis and automated database searching. To evaluate the purification efficiency of the cross-linked proteasome complexes, we monitored each purification step by immunoblotting. As shown in Fig. 2A, all of the cross-linked products containing an Rpn11-HB proteasome subunit were bound efficiently on Ni-NTA resin because there was no visible amount left in the flow-through after the binding. Proteins eluted from the Ni-NTA resin were subsequently incubated


369


formaldehyde (data not shown). Mass spectrometric analysis was also used to determine the optimal formaldehyde concentration for in vivo cross-linking. Because the optimal cross-linking condition is expected to provide the highest yield of specific cross-linking products but the lowest nonspecific background, we compared a list of proteins identified from samples cross-linked with different formaldehyde concentrations. The number of proteasome subunits identified was comparable in the samples that were cross-linked with 1 or 2% formaldehyde for 10 min at 30 °C, but the number of peptides matched to each subunit was higher in the 1% formaldehyde sample. Cross-linking with a final concentration of 3% formaldehyde resulted in a significantly lower amount of peptides. This is probably due to the fact that higher concentrations of formaldehyde (⬎2%) tend to cause more proteins to cross-link with DNA, leading to a reduced amount of total soluble protein in the lysate.

Identification of the 26 S Proteasome Complex after in Vivo Formaldehyde Cross-linking and Tandem Affinity Purification FIG. 2. In vivo cross-linking and tandem affinity purification. A, tandem affinity purification of in vivo cross-linked Rpn11-HB-containing proteasomes using Ni-NTA resin and streptavidin beads. The Rpn11-HB cross-linked complexes were efficiently isolated using two-step affinity purification as shown by immunoblotting using antiRGS-His6 antibodies. Lane 1, cell lysate; lane 2, unbound fraction after incubation with Ni-NTA resin; lane 3, wash with a buffer (pH 6.3) containing 8 M urea and 10 mM imidazole; lane 4, elution from Ni-NTA resin (i.e. load to streptavidin); lane 5, unbound fraction after binding to streptavidin beads. The samples were separated by 7% SDSPAGE, and Rpn11-HB was detected by immunoblotting using antiRGS-His6 antibodies. B, test of cross-linking efficiency using different concentrations of formaldehyde (FA) as indicated for 10 min. The samples were separated by 10% SDS-PAGE and detected as in A.

with immobilized streptavidin for the second purification step. As shown in Fig. 2A, lane 5 (flow-through after streptavidin binding), the Rpn11-HB-containing complexes were efficiently captured on streptavidin beads. These results demonstrated that HB-tagged Rpn11 is compatible with formaldehyde cross-linking and that cross-linking did not interfere with protein purification. Similar results were observed from cells expressing Rpt5-HB (data not shown).

Optimization of in Vivo Cross-linking of HB-tagged Proteasome-interacting Proteins Using Formaldehyde To optimize the cross-linking condition, different concentrations of formaldehyde were tested at various incubation times in yeast cells expressing the Rpn11-HB. As shown in Fig. 2B, immunoblot analysis indicated that protein-protein cross-linking efficiency did not improve with formaldehyde concentration above 1%. Similarly cross-linking efficiency appeared to plateau after 10 min of incubation at 30 °C with 1%

370


After optimizing the cross-linking and affinity purification conditions, we isolated the cross-linked 26 S proteasome complexes from cells expressing either Rpn11-HB or Rpt5HB. Using these strains, we were able to purify and identify the complete composition of the 26 S proteasome complex as summarized in Table I. In addition to all essential 26 S proteasome subunits, the newly assigned proteasome subunits including Rpn13 (7), Sem1 (9), Ubp6, and Ecm29 (8) were captured. A wild type, untagged strain was processed in parallel and did not result in the identification of any proteasome subunits. Given that the purification was carried out under fully denaturing conditions, these results indicate that the identified proteasome subunits were specifically present in the in vivo cross-linked protein complexes containing the HB-tagged bait Rpn11 or Rpt5.

Quantitative Analysis of PIPs Using SILAC and Tandem Mass Spectrometry Although the tandem affinity purification technique using Ni-NTA/streptavidin resins has improved purification efficiency significantly compared with one-step Ni-NTA affinity purification, there are still some background proteins present after the final purification step. To distinguish specific PIPs from background proteins, we used a quantitative approach using the SILAC strategy (37, 38), which is illustrated in Fig. 3. Briefly one population of cells is grown in medium containing the natural form of an essential amino acid, [12C6]Arg, whereas another population (e.g. untagged wild type cells) is grown in medium containing a stable, isotope-labeled analogue, [13C6]Arg. Equal amounts of protein from the two different populations of cells are mixed and affinity-purified together. The bound proteins on streptavidin beads are


TABLE I Summary of the 26 S proteasome subunits identified using in vivo cross-linking, tandem affinity purification, and mass spectrometry

Accession no.

19 S regulatory subunits P38764 P32565 P40016 Q12250 Q12377 Q06103 Q08723 Q04062 P38886 P43588 P32496 6323453a O94742 P33299 P40327 P33298 P53549 P33297 Q01939 Others P43943 P50086 P38737 20 S core subunits P21243 P23639 P23638 P40303 P32379 P40302 P21242 P38624 P25043 P25451 P22141 P30656 P23724 P30657

Gene name

No. of peptides identified with tagged Rpn11

No. of peptides identified with tagged Rpt5

RPN1 RPN2 RPN3 RPN5 RPN6 RPN7 RPN8 RPN9 RPN10 RPN11 RPN12 RPN13 SEM1 RPT1 RPT2 RPT3 RPT4 RPT5 RPT6

57 68 31 34 27 25 24 30 14 34 17 5 2 36 37 36 36 43 35

53 55 26 35 20 22 19 27 11 24 11 5 1 35 34 30 38 42 33

UBP6 NAS6 ECM29

8 0 4

7 2 47

SCL1 PRE8 PRE9 PRE6 PUP2 PRE5 PRE10 PRE3 PUP1 PUP3 PRE1 PRE2 PRE3 PRE4

15 18 12 9 10 12 8 5 2 1 2 3 2 6

12 10 8 8 10 14 10 9 0 1 1 3 2 5

a

This protein entry is not present in Swiss-Prot database, only in NCBInr.

digested and analyzed by LC MS/MS. If a protein is present in both samples, the resulting arginine-containing peptides will be observed as pairs ([12C6]Arg versus [13C6]Arg) due to the incorporation of different isotope-labeled arginines during cell culture. The relative abundance of proteins present in two different samples can thus be quantified based on the ratios of the peak intensities of arginine containing peptide pairs determined in LC MS experiments. In the LC MS spectra, peptides observed as singlets indicate that they are either arginine-containing peptides derived from the proteins only present in one sample or peptides without arginine (e.g. lysine-containing tryptic peptides or C-terminal peptides). This

FIG. 3. Schematic diagram of a quantitative strategy using SILAC coupled with in vivo cross-linking and tandem affinity purification to identify specific proteasome-interacting proteins. Cells expressing Rpn11-HB were grown in light medium ([12C6]Arg), whereas wild type cells (untagged) were grown in heavy medium ([13C6]Arg).

quantitative method is a powerful approach to identify specific interacting proteins of interest (38). We first evaluated the conditions for the SILAC experiment in yeast cells. Wild type Saccharomyces cerevisiae can synthesize all amino acids, which might lead to incomplete incorporation of isotopically labeled amino acids during in vivo labeling of cells. To circumvent this potential problem, we generated yeast strains that are auxotrophic for arginine by deletion of the ARG4 gene to prevent endogenous arginine from competing with the labeled [13C6]Arg in the medium. To test the incorporation of labeled arginine, cells were grown either in the presence of [13C6]Arg or of [12C6]Arg for 8 h. Proteins were extracted from both samples and separated by 1-D gel electrophoresis. Gel bands corresponding to the same proteins were excised from the two samples, digested with trypsin, and analyzed by LC MS/MS. The results demonstrated that incorporation of [13C6]Arg was close to completion (⬎95%), which was further improved with overnight incorporation, comparable to what has been reported by Gruhler et al. (39). Interestingly in our yeast strain carrying the arg4⌬ deletion, no [13C6]Arg was converted to [13C5]proline. This is advantageous because conversion of [13C6]Arg to [13C5]proline was observed previously in mammalian cells (40)


371


FIG. 4. TOF MS spectra of six Argcontaining peptide pairs from the selected proteins including two unchanged proteins, Pyc1 (L/H, 1. 12) (A) and 60 S Rpl3 (L/H, 1.16) (B); two proteasome subunits, Rpt6 (C) and Scl1 (D); and two putative PIPs, Ssa1 (L/H, 5.1) (E) and Yef3 (L/H, 3.2) (F). The relative abundance ratio of the Arg-containing peptides was calculated as L/H based on either monoisotopic peak intensity or area. All the peptides were observed as either doubly or triply charged ions and identified as peptides containing one arginine; therefore, the mass differences between each peptide pair is 6 Da. The ratios for proteasome subunits could not be calculated because the corresponding 13C-labeled peptides were not observed.

and requires correction for the quantitation of proline-containing peptides. Analysis and quantitation of several other randomly chosen proteins confirmed the excellent labeling efficiency (data not shown), indicating that an arg4⌬ yeast strain is suitable for SILAC experiments using [13C6]Arg labeling. To identify the specific PIPs after in vivo cross-linking, we cultured an Rpn11-HB-expressing strain in [12C6]Arg- and a wild type strain in [13C6]Arg-containing medium (Fig. 3). After in vivo formaldehyde cross-linking, equal amounts of cell lysate from the two different populations of cells were mixed for tandem affinity purification and subsequent trypsin digestion. The digests were analyzed by both 1-D and two-dimensional LC MS/MS to improve the detection sensitivity and dynamic range of the analysis. The list of proteins identified were summarized, compared, and validated, and their relative peptide abundance ratios of Arg-containing peptide pairs were calculated using the Search Compare program within Protein Prospector (35). Three different groups of proteins were classified based on their relative peptide abundance ratios as follows. The 26 S Proteasome Subunits—The first group of proteins are the subunits of the 26 S proteasome complex. We expected that the Rpn11-HB cross-linked complexes would be enriched significantly during the tandem affinity purification

372


from the cells expressing Rpn11-HB but not from the wild type cells. Because Rpn11-HB-tagged cells were grown in the medium containing light Arg (L), whereas wild type cells were cultured in heavy Arg (H) medium, the relative abundance ratios (L/H) of the proteasome subunits were expected to be much larger than 1. For all of the identified proteasome subunits including 19 and 20 S core subunits, we observed the same pattern of the Arg-containing peptides (light versus heavy forms) in the LC MS experiments. This is exemplified on two Arg-containing peptides from two different proteasome subunits, Rpt6 (an ATPase subunit) and Scl1 (␣1 subunit of 20 S) (Fig. 4, A and B). Their sequences were determined by MS/MS. Because only one arginine is present in each sequence, the mass differences between the assumed peptide pairs should be 6 Da if both light and heavy forms were present. Based on the defined mass differences, no corresponding [13C6]Arg-labeled peptides were observed for any identified [12C6]Arg-labeled proteasome peptides, resulting in the observation of Arg-containing peptides as singlets, not pairs. This suggests that the proteasome subunits were selectively enriched and purified from the mixed lysates (Rpn11-HB and wild type cells), so, therefore, their relative abundance ratios are high.


TABLE II Summary of the putative PIPs identified by QTAX with the relative abundance ratios ⬎1.5 Gene name

ORF name

L/H ratio (average)

S.D.

RPS2a UBI4 DSK2 RAD23 RPL13A/B SSE1/SSE2a RPS24A/Ba SSA3 SSA4 SSA2 SSA1 KAR2a SSB2 SEC53a YBRO25Ca SSB1a YEF3a HEF3a SAM1/2a EFT1/2a RPS11A/Ba FPR1a HSC82 ENO2 ENO1 KAP123a HSP82a LEU1a PMA1/2a LYS21a PAB1 LYS20a PGK1 CYC1/7a TIF2 TIF1a FBA1a ERG20a GND1/2a BMH1/2 ACT1a TSA1a CDC19a TAL1a CDC60a ADH1 PDC1 PDC5/6a ASN1a RPS0A/Ba YHB1a ALD6a SER1a THS1a TEF4a TDH1/2/3 TEF1/2 KRS1

YGL123W YLL039C YMR276W YEL037C YDL082W YPL106C YER074W YBL075C YER103W YLL024C YAL005C YJL034W YNL209W YFL045C YBR025C YDL229W YLR249W YNL014W YLR180W YOR133W YDR025W YNL135C YMR186W YHR174W YGR254W YER110C YPL240C YGL009C YGL008C YDL131W YER165W YDL182W YCR012W YJR048W YJL138C YKR059W YKL060C YJL167W YHR183W YER177W YFL039C YML028W YAL038W YLR354C YPL160W YOL086C YLR044C YLR134W YPR145W YGR214W YGR234W YPL061W YOR184W YIL078W YKL081W YJL052W YPR080W YDR037W

No 13C No 13C No 13C No 13C No 13C No 13C No 13C 5.9 5.7 5.2 5.1 5.1 3.5 3.4 3.3 3.2 3.2 3.1 3.1 2.7 2.7 2.7 2.6 2.5 2.5 2.5 2.4 2.1 2.1 2.1 2.1 2 2 2 1.9 1.9 1.9 1.9 1.9 1.8 1.8 1.8 1.8 1.8 1.8 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.7 1.6 1.6 1.6 1.6

N/Ab N/A N/A N/A N/A N/A N/A 1.0 1.2 1.1 0.9 —c 0.6 0.7 0.5 0.5 0.7 0.5 1.1 0.1 — — 0.5 0.6 0.7 — 0.6 0.3 0.2 0.3 0.0 0.3 0.6 0.2 0.5 0.5 0.3 0.1 0.3 0.3 0.1 0.3 0.4 0.2 — 0.6 0.3 0.3 0.2 0.2 0.3 0.1 0.3 0.2 0.1 0.2 0.3 0.2

TABLE II— continued Gene name

ORF name

L/H ratio (average)

S.D.

DPS1a VAS1a RPS27A/Ba GLN1a RPL5 SAH1a

YLL018C YGR094W YKL156W YPR035W YPL131W YER043C

1.6 1.6 1.6 1.6 1.6 1.6

0.2 0.1 0.1 0.1 0.2 0.2

a

PIPs captured that have not been reported previously. N/A, not applicable. c —, only one pair was used for the ratio calculation. b

Putative Specific PIPs—The second group of proteins identified is classified as the putative specific PIPs with relative abundance ratios ⬎1.5. We identified a total of 64 putative specific PIPs that are summarized in Table II. To our knowledge, 42 of these putative PIPs are newly identified and have not been reported before. Among these 64 proteins are polyubiquitins, ubiquitin receptors, ubiquitinated substrates, chaperons, transcription factors, translational factors, karyopherin, etc. Because polyubiquitin and the ubiquitin receptors (Rad23 and Dsk2) are known to be directly involved in ubiquitinproteasome degradation pathways (13, 14), their interactions with the proteasome are believed to be highly specific, and their ratios are also high. As shown in the TOF MS spectra (Fig. 5, A and B), only the light forms of a ubiquitin peptide were clearly observed as either triply charged (MH33⫹ 487.6) or doubly charged (MH22⫹ 730.9) ions, and no [13C6]Arglabeled forms of the same peptide were detected. The peptide sequence was determined as LIFAG(K48(GG))*QLEDGR by a series of y ions (y1–y9) obtained in the MS/MS spectrum (MH33⫹ 487.6) shown in Fig. 5C resulting from the polyubiquitin chain linked to lysine at position 48. After trypsin digestion, the remnant of one ubiquitin molecule on the Lys48 of the other ubiquitin has two glycine residues (i.e. GG) (41). It is noteworthy that only Lys48-linked polyubiquitin chains were detected to be associated with proteasome in this experiment, and no other polyubiquitin-chain linkage types were observed. This is consistent with the notion that Lys48-linked chains are the predominant degradation signal for ubiquitinated substrates (1, 3, 17). Similarly the [12C6]Arg-containing peptides of the known ubiquitin receptors Rad23 and Dsk2 were also identified as singlets due to the absence of the 13 C-labeled forms of these peptides. As an example, the TOF MS spectrum of a tryptic peptide (MH22⫹ 752.85) is shown in Fig. 6, inset, and its sequence was determined as QLNDMGFFDEDR, which matched unambiguously to Dsk2. As shown in Fig. 6B, the relative abundance ratio of this particular peptide should be ⬎8 even if the intensity of the [13C6]Arg-labeled peptide (MH22⫹ 755.85) was the same as the noise level (⬃7 counts). Similar results were obtained for Rad23, indicating that Ub receptors and the components directly involved in the Ubproteasome degradation pathways can be specifically enriched. In addition to polyubiquitin, Rad23, Dsk2, and several other


373


FIG. 5. TOF MS spectra of an Arg-containing peptide from polyubiquitin observed as triply charged ion (MH33⫹ 487. 6) (A) and doubly charged ion (MH22⫹ 730.9) (B). No detectable 13C6-labeled peptides were observed in these spectra; the pattern of Arg-containing peptides (L versus H) was the same as that in Fig. 4, C and D. C, MS/MS spectrum of the triply charged ion MH33⫹ 487.6. The sequence was determined as LIFAG(K48(GG))*QLEDGR, a ubiquitin-modified peptide.

PIPs, i.e. Sse1/Sse2, Rps24A/B, Rpl13A/B, and Rps2, also have high ratios as listed in Table II. In comparison to the PIPs with high ratios, the Arg-containing peptides from other putative PIPs have been observed as pairs, not as singlets, and their relative abundance ratios are ranging from 1.6 to 6. As an example, the selected Arg peptide pairs from Ssa1 and Yef3 are given in Fig. 4, C and D, and their relative abundance ratios of L/H were determined as 5.1 and 3.2, respectively. These results were reproducible from multiple MS analyses and different preparations. Background Proteins—The last groups of proteins with relative abundance ratios ranging from 0.8 to 1.5, i.e. close to 1, are considered background proteins due to nonspecific interactions. The representative TOF MS spectra of the Arg-containing peptide pairs from two unchanged proteins are shown in Fig. 4, E and F. The first peptide sequence was determined as EGDDVADAFQR and matched to pyruvate carboxylase Pyc1 with a ratio of 1.12 (Fig. 4E), and the second was identified as YAQDGAGIER and matched to the 60 S ribosomal subunit Rpl3 with a ratio of 1.16 (Fig. 4F). Our results

374


show that the ratios of different peptides resulting from the same proteins and the ratios from different proteins in this group are reproducible within experimental error. Therefore, based on the relative abundance ratio (⬃1), it is easy to identify these background proteins. It is worth mentioning that three previously reported putative PIPs, Acc1, Ilv6, and Shm2 (7), had relative abundance ratios close to 1, indicating that under our experimental conditions they are likely nonspecific interacting proteins.

Validation of the Selected Identified PIPs We selected eight proteins with different abundance ratios to validate their interactions with the proteasome by an independent method using a different epitope tag. The proteins chosen in order of highest to lowest ratios were: Ssa1, Ybr025c, Yef3, Eft1, Tef4, Pfk1, Gus1, and Acc1. We used yeast strains expressing endogenous levels of TAP-tagged versions of these proteins and purified protein complexes based on the affinity of the TAP tag to IgG-Sepharose. Protein


FIG. 6. A, MS/MS spectrum of a tryptic peptide (MH22⫹ 752. 86). Its sequence, determined as QLNDMGFFDFDR, matched to the Ub receptor Dsk2. B, TOF MS spectrum of the Dsk2 peptide shown in A. No 13C-labeled peptide was observed.

complexes were proteolytically released from the IgG-Sepharose by TEV protease cleavage at the TEV site present in the TAP tag. Protein complexes were analyzed for the presence of proteasomes by immunoblotting using an antibody directed against the proteasome subunit Rpt6. The immunoblotting results are illustrated in Fig. 7. A yeast strain expressing the TAP-tagged 20 S proteasome subunit Pre1 was used as a positive control (Fig. 7, lanes 1 and 9), and an untagged yeast strain was used as a negative control (Fig. 7, lanes 2 and 7). We found that all proteins tested with a ratio of 1.6 or higher (Ssa1, Ybr025c, Yef3, Eft1, and Tef4) interacted specifically with the proteasome (Fig. 7, lanes 3– 6 and 10), whereas the proteins with ratios less than 1.6 (Pfk1, Gus1, and Acc1) did not show positive interactions under these conditions (Fig. 7, lanes 8, 11, and 12). These results suggest that the putative PIPs with ratios ⬎1.5 are most likely to be specific proteasome-interacting proteins. However, it is important to emphasize that these experiments did not involve cross-linking, and purification was carried out under native conditions. Therefore, it is possible that proteins that tested negative in this assay but had abundance ratios ⬎1 are interacting with the proteasome in vivo, but their interactions are very unstable or transient and only able to be captured in combination with in vivo cross-linking. DISCUSSION

FIG. 7. Validation of selected PIPs identified with various abundance ratios by co-immunoprecipitation using TAP-tagged yeast strains. Lanes 1 and 9 are positive controls using yeast strains expressing the TAP-tagged 20 S subunit Pre1. For a better comparison, 25 times less of this sample was loaded as compared with the other samples. Lanes 2 and 7 are negative controls using a wild type yeast strain. Lane 3, Ssa1-TAP (L/H, 5.1); lane 4, Ybr025c-TAP (L/H, 3.3); lane 5, Yef3-TAP (L/H, 3.2); lane 6, Eft1-TAP (L/H, 2.7); lane 8, Acc1-TAP (L/H, 1.0); lane 10, Tef4-TAP (L/H, 1.6); lane 11, Pfk1-TAP (L/H, 1.5); lane 12, Gus1-TAP (L/H, 1.3). Yeast strains expressing the TAP-tagged proteins indicated were lysed, and protein complexes were purified under native conditions based on the affinity of the TAP tag to IgG-Sepharose and eluted by TEV protease cleavage. Eluted protein complexes were analyzed for the presence of proteasomes by immunoblotting with antibodies directed against the 19 S subunit Rpt6. All tested proteins with abundance ratio ⬎1.5 showed copurification of the proteasome subunit Rpt6. The purification of 26 S proteasome from the TAP-tagged Pre1 strain was carried out in the presence of an ATP-regenerating system. *, not applicable.

We have presented a novel integrated proteomics approach, QTAX, for the analysis of protein complexes including transient and weak interaction partners that has been successfully applied to decipher the 26 S proteasome interaction

networks from budding yeast in vivo. This presents an improvement over other strategies using affinity purification and mass spectrometric analysis either with or without cross-


375


linking. Combined with the SILAC strategy, a total of 64 putative PIPs were identified, and they are involved in a variety of biological processes, including proteolysis, RNA processing, translation, metabolism, cell cycle, protein folding, etc. (Table II). In addition to some known interactions, 42 putative new PIPs were identified, suggesting that many proteins may be involved in the ubiquitin-proteasome pathway. In this work, in vivo cross-linking using formaldehyde was carried out to fix protein interactions prior to cell lysis, which can stabilize weak and transient interactions during purification processes. To eliminate the nonspecific purification background (i.e. proteins bound to affinity resins nonspecifically) and prevent the formation of non-covalent interactions after cell lysis, affinity purification of the cross-linked complexes under fully denaturing conditions was preferred, although immunoaffinity purification under native condition has been applied to the study of M-Ras-interacting proteins (28). Previously the only tag available that could be used for affinity purification under fully denaturing conditions was the His tag, which has been successfully used for the purification of ubiquitin and ubiquitin-like modified proteins using Ni-NTA chromatography (41, 42); however, single step purification using the His tag alone is not very specific (42, 43). To improve the purification specificity, a new tandem affinity tag, the HB tag, was adopted in this work that has proven to be compatible for both in vivo cross-linking and subsequent tandem affinity purification of the cross-linked protein complexes using NiNTA chromatography and affinity binding to streptavidin resins under fully denaturing conditions (e.g. 8 M urea). The extremely high affinity binding of biotin to the streptavidin resin (Kd ⫽ 10⫺15) allows stringent wash steps to effectively remove nonspecific interactions, thus achieving higher purification specificity in comparison to one-step Ni-NTA purification. This was further confirmed by mass spectrometric analysis of the background proteins from one-step and two-step purification. The number of the background proteins (⬃40) obtained with two-step Ni-NTA/streptavidin purification is about 16% of that (⬃240) obtained with only one-step Ni-NTA purification from untagged yeast cells after formaldehyde cross-linking. Furthermore cross-linking did not seem to increase nonspecific purification background because the number of background proteins are comparable from cross-linked and non-cross-linked cells (data not shown). In comparison with previous in vivo cross-linking/affinity purification strategies (25, 26, 28), quantitative purification of cross-linked proteasome complexes using the HB tag was carried out using two-step processes under fully denaturing conditions, and the purified cross-linked proteasome complexes were analyzed directly by mass spectrometry without separation by SDSPAGE, which may be advantageous for preserving physiologically formed protein interaction profiles and improving the sensitivity of protein identification. The identification of the full composition of 26 S proteasome complex for the first time in one single experiment

376


suggests that the proteasome assembly is dynamic and that some of the loosely bound subunits can be lost even during one-step affinity purification. Although many peptides (⬎10) of the 19 S and 20 S ␣ subunits were obtained during mass spectrometric analysis (Table I), most of the 20 S ␤ subunits were identified by much smaller number of peptides (one to three), possibly due to the fact that the ␤ subunits were embedded in the center of the stacked ring structure of the 20 S proteasome complex, and cross-linking may block some potential trypsin digestion sites because lysines are the major sites for cross-linking reaction. Therefore, cross-linking reversal may be useful to expose more lysine residues for trypsin digestion, leading to the detection of additional peptides for improved sequence coverage. Comparing the results from Rpn11-HB- and Rpt5-HB-tagged cells, the majority of the proteins identified are very similar. As shown in Table I, some proteins seem to be found preferentially in one type of the tagged complex. Rad23 and Dsk2 were found only in Rpn11HB-tagged cells, whereas Nas6 was only identified in Rpt5HB-tagged cells. The human homolog of Nas6, Gankyrin, was found to interact directly with Rpt5 using a yeast two-hybrid system (44), which agrees well with our results. There are some surprising differences as the number of peptides identified for some proteins changed quite dramatically. For example, 47 of Ecm29 peptides from Rpt5-HB-tagged cells were identified; this was much higher than the five Ecm29 peptides from Rpn11-HB-tagged cells. This suggests that subcomplexes of the proteasome may exist in vivo or that some PIPs can be differentially enriched after cross-linking due to their physical locations. Therefore, using several different HB-tagged proteasome subunits will be beneficial to capture all proteasome-interacting proteins. The SILAC strategy was successfully used to quantitatively distinguish the proteasome complex and its interacting proteins from background proteins. Based on the validation results using co-IP analysis, we assigned the proteins with abundance ratios ⬎1.5 as putative specific PIPs (Table II). Interestingly most of the identified putative PIPs have relative abundance ratios ranging from 1.6 to 6, smaller than those of the known components involved in ubiquitin-proteasome pathways. The differences in the relative abundance ratios among the PIPs may be due to inherent differences in their interactions with the proteasome complex and their role in the ubiquitin-proteasome degradation pathways. Among the identified PIPs are a number of abundant cellular proteins such as heat shock proteins, elongation factors, ribosomal proteins, etc. Due to the large quantity of these proteins present in the cells, it is difficult to completely prevent their nonspecific purification, which is thus leading to a relatively small abundance ratio. Nevertheless these proteins are likely bona fide proteasome interactors because their relative abundance ratios were consistently above 1.5. Among the identified chaperone proteins, different classes were detected including Ssa, Ssb, and Kar2 members of the Hsp70


family and Sse1/Sse2, Hsc82, and Hsp82 (the budding yeast homologs of Hsp90), and all have relative abundance ratios ⬎2. This suggests that their interactions with the proteasome are specific. The co-IP experiment further confirmed that the Hsp70 family member Ssa1 in budding yeast associates with the proteasome specifically. In higher eukaryotes, it has been shown that Hsc70/Hsp70 members facilitate the delivery of aggregation-prone substrates for degradation by interacting with the proteasome through an adaptor protein (45, 46). In addition, Hsp90 family members have been suggested to play a role in the proteasome structural integrity and assembly through their interactions with the 26 S proteasome (47). Therefore, different chaperone members may play distinct roles in modulating protein degradation by the proteasome. Previously Tef1/2 (i.e. EF-1 ␣) was identified as a putative PIP by mass spectrometry (7). In this work, Tef1/2 interaction with the proteasome was also detected and determined to be specific because it had a ratio of 1.6. This agrees well with recent reports that the elongation factor 1 complex (␣␤␥) interacts with the 26 S proteasome (48 –50), and Tef1 possesses a chaperone-like activity and can interact with both ubiquitinated substrates and the proteasome subunit Rpt1 to promote co-translationally damaged protein degradation (49, 50). In addition to Tef1/2, several other translational elongation factors including Yef3/Hef3, Eft1/2, and Tef4 were also identified as specific PIPs by the SILAC ratios (⬎1.6) and further confirmed by co-IP analysis. Although these interactions have not been detected before, we suspect that these translational elongation factors may also be involved in the degradation of co-translationally damaged proteins, and they may serve as the molecular linkage between the pathways of protein synthesis and degradation. In summary, QTAX is a powerful integrated proteomic approach to capture and identify stable and transient proteinprotein interactions occurring in the cellular environment. In comparison with the existing methods, we have made significant technical improvements in two major aspects for the study of protein interactions using in vivo cross-linking. First, the new tandem affinity tag (HB tag) has demonstrated to be effective for purification of the cross-linked proteasome complexes and their interacting proteins under fully denaturing conditions, resulting in substantial reduction of nonspecific binding during the purification in comparison to one-step Ni-NTA purification. Second, quantitative mass spectrometry using the SILAC strategy in yeast cells has been used to study proteasome interactions, resulting in the effective differentiation of specific interactions from nonspecific interactions. The QTAX strategy described here has shown that in vivo crosslinking using formaldehyde effectively captures proteasome complexes and their interacting partners including some that have not been identified previously. Further genetic and biochemical studies of these interactors may help to reveal how the 26 S proteasome is integrated into various biological pathways and how different PIPs play a role in substrate

recognition and translocation to the 26 S proteasome for degradation. Acknowledgments—We thank Prof. Ralph Bradshaw for critical reading of the manuscript, Prof. A. L. Burlingame for allowing us to use the development version of the Protein Prospector software package, and Aenoch Lynn, June Snedecor, and Peter Baker for help on Protein Prospector. We also thank Karen LeBlanc for the purified proteasome complex for the co-IP experiments and members of the Huang and Kaiser laboratories for help during this study. * This work was supported by National Institutes of Health Grants GM-74830 (to L. H.) and GM-66164 (to P. K.), Department of the Army Grant PC-041126 (to L. H.), and California Breast Cancer Research Program Grant 11NB-0177 (to P. K.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ** To whom correspondence should be addressed: Depts. of Physiology & Biophysics and Developmental & Cell Biology, Medical Science I, D233, University of California, Irvine, CA 92697-4560. E-mail: [email protected]. REFERENCES 1. Baumeister, W., Walz, J., Zuhl, F., and Seemuller, E. (1998) The proteasome: paradigm of a self-compartmentalizing protease. Cell 92, 367–380 2. Pickart, C. M. (2001) Mechanisms underlying ubiquitination. Annu. Rev. Biochem. 70, 503–533 3. Pickart, C. M., and Cohen, R. E. (2004) Proteasomes and their kin: proteases in the machine age. Nat. Rev. Mol. Cell. Biol. 5, 177–187 4. Voges, D., Zwickl, P., and Baumeister, W. (1999) The 26S proteasome: a molecular machine designed for controlled proteolysis. Annu. Rev. Biochem. 68, 1015–1068 5. Smalle, J., and Vierstra, R. D. (2004) The ubiquitin 26s proteasome proteolytic pathway. Annu. Rev. Plant Biol. 55, 555–590 6. Kloetzel, P. M. (2001) Antigen processing by the proteasome. Nat. Rev. Mol. Cell. Biol. 2, 179 –187 7. Verma, R., Chen, S., Feldman, R., Schieltz, D., Yates, J., Dohmen, J., and Deshaies, R. J. (2000) Proteasomal proteomics: identification of nucleotide-sensitive proteasome-interacting proteins by mass spectrometric analysis of affinity-purified proteasomes. Mol. Biol. Cell 11, 3425–3439 8. Leggett, D. S., Hanna, J., Borodovsky, A., Crosas, B., Schmidt, M., Baker, R. T., Walz, T., Ploegh, H., and Finley, D. (2002) Multiple associated proteins regulate proteasome structure and function. Mol. Cell 10, 495–507 9. Sone, T., Saeki, Y., Toh-e, A., and Yokosawa, H. (2004) Sem1p is a novel subunit of the 26 S proteasome from Saccharomyces cerevisiae. J. Biol. Chem. 279, 28807–28816 10. Schmidt, M., Haas, W., Crosas, B., Santamaria, P. G., Gygi, S. P., Walz, T., and Finley, D. (2005) The HEAT repeat protein Blm10 regulates the yeast proteasome by capping the core particle. Nat. Struct. Mol. Biol. 12, 294 –303 11. Hershko, A., and Ciechanover, A. (1998) The ubiquitin system. Annu. Rev. Biochem. 67, 425– 479 12. Saeki, Y., Saitoh, A., Toh-e, A., and Yokosawa, H. (2002) Ubiquitin-like proteins and Rpn10 play cooperative roles in ubiquitin-dependent proteolysis. Biochem. Biophys. Res. Commun. 293, 986 –992 13. Elsasser, S., Chandler-Militello, D., Muller, B., Hanna, J., and Finley, D. (2004) Rad23 and Rpn10 serve as alternative ubiquitin receptors for the proteasome. J. Biol. Chem. 279, 26817–26822 14. Verma, R., Oania, R., Graumann, J., and Deshaies, R. J. (2004) Multiubiquitin chain receptors define a layer of substrate selectivity in the ubiquitin-proteasome system. Cell 118, 99 –110 15. Elsasser, S., and Finley, D. (2005) Delivery of ubiquitinated substrates to protein-unfolding machines. Nat. Cell Biol. 7, 742–749 16. Deveraux, Q., Ustrell, V., Pickart, C., and Rechsteiner, M. (1994) A 26 S protease subunit that binds ubiquitin conjugates. J. Biol. Chem. 269, 7058 –7061 17. van Nocker, S., Sadis, S., Rubin, D. M., Glickman, M., Fu, H., Coux, O.,


377


18.

19.

20. 21.

22.

23. 24.

25.

26.

27. 28.

29.

30.

31. 32.

33.

Wefes, I., Finley, D., and Vierstra, R. D. (1996) The multiubiquitin-chainbinding protein Mcb1 is a component of the 26S proteasome in Saccharomyces cerevisiae and plays a nonessential, substrate-specific role in protein turnover. Mol. Cell. Biol. 16, 6020 – 6028 Saeki, Y., Sone, T., Toh-e, A., and Yokosawa, H. (2002) Identification of ubiquitin-like protein-binding subunits of the 26S proteasome. Biochem. Biophys. Res. Commun. 296, 813– 819 Lam, Y. A., Lawson, T. G., Velayutham, M., Zweier, J. L., and Pickart, C. M. (2002) A proteasomal ATPase subunit recognizes the polyubiquitin degradation signal. Nature 416, 763–767 Aebersold, R., and Mann, M. (2003) Mass spectrometry-based proteomics. Nature 422, 198 –207 Ho, Y., Gruhler, A., Heilbut, A., Bader, G. D., Moore, L., Adams, S. L., Millar, A., Taylor, P., Bennett, K., Boutilier, K., Yang, L., Wolting, C., Donaldson, I., Schandorff, S., Shewnarane, J., Vo, M., Taggart, J., Goudreault, M., Muskat, B., Alfarano, C., Dewar, D., Lin, Z., Michalickova, K., Willems, A. R., Sassi, H., Nielsen, P. A., Rasmussen, K. J., Andersen, J. R., Johansen, L. E., Hansen, L. H., Jespersen, H., Podtelejnikov, A., Nielsen, E., Crawford, J., Poulsen, V., Sorensen, B. D., Matthiesen, J., Hendrickson, R. C., Gleeson, F., Pawson, T., Moran, M. F., Durocher, D., Mann, M., Hogue, C. W., Figeys, D., and Tyers, M. (2002) Systematic identification of protein complexes in Saccharomyces cerevisiae by mass spectrometry. Nature 415, 180 –183 Gavin, A., Bosche, M., Krause, R., and Grandi, P., Marzioch, M., Bauer, A., Schultz, J., Rick, J. M., Michon, A. M., Cruciat, C. M., Remor, M., Hofert, C., Schelder, M., Brajenovic, M., Ruffner, H., Merino, A., Klein, K., Hudak, M, Dickson, D., Rudi, T., Gnau, V., Bauch, A., Bastuck, S., Huhse, B., Leutwein, C., Heurtier, M. A., Copley, R. R., Edelmann, A., Querfurth, E., Rybin, V., Drewes, G., Raida, M., Bouwmeester, T., Bork, P., Seraphin, B., Kuster, B., Neubauer, G., and Superti-Furga, G. (2002) Functional organization of the yeast proteome by systematic analysis of protein complexes. Nature 415, 141–147 Jackson, V. (1999) Formaldehyde cross-linking for studying nucleosomal dynamics. Methods 17, 125–139 Kuo, M. H., and Allis, C. D. (1999) In vivo cross-linking and immunoprecipitation for studying dynamic protein:DNA associations in a chromatin environment. Methods 19, 425– 433 Schmitt-Ulms, G., Legname, G., Baldwin, M. A., Ball, H. L., Bradon, N., Bosque, P. J., Crossin, K. L., Edelman, G. M., DeArmond, S. J., Cohen, E., and Prusiner, S. B. (2001) Binding of neural cell adhesion molecules (N-CAMs) to the cellular prion protein. J. Mol. Biol. 314, 1209 –1225 Schmitt-Ulms, G., Hansen, K., Liu, J., Cowdrey, C., Yang, J., DeArmond, S. J., Cohen, F. E., Prusiner, S. B., and Baldwin, M. A. (2004) Timecontrolled transcardiac perfusion cross-linking for the study of protein interactions in complex tissues. Nat. Biotechnol. 22, 724 –731 Orlando, V., Strutt, H., and Paro, R. (1997) Analysis of chromatin structure by in vivo formaldehyde cross-linking. Methods 11, 205–214 Vasilescu, J., Guo, X., and Kast, J. (2004) Identification of protein-protein interactions using in vivo cross-linking and mass spectrometry. Proteomics 4, 3845–3854 Otsu, M., Omura, F., Yoshimori, T., and Kikuchi, M. (1994) Protein disulfide isomerase associates with misfolded human lysozyme in vivo. J. Biol. Chem. 269, 6874 – 6877 Hall, D. B., and Struhl, K. (2002) The VP16 activation domain interacts with multiple transcriptional components as determined by protein-protein cross-linking in vivo. J. Biol. Chem. 277, 46043– 46050 Guthrie, C., and Fink, G. R. (1991) Guide to Yeast Genetics and Molecular Biology, Academic Press, Inc., San Diego, CA Reed, S. I., Hadwiger, J. A., and Lorincz, A. T. (1985) Protein kinase activity associated with the product of the yeast cell division cycle gene CDC28. Proc. Natl. Acad. Sci. U. S. A. 82, 4055– 4059 Petracek, M. E., and Longtine, M. S. (2002) PCR-based engineering of

378


yeast genome. Methods Enzymol. 350, 445– 469 34. Manivasakam, P., Weber, S. C., McElver, J., and Schiestl, R. H. (1995) Micro-homology mediated PCR targeting in Saccharomyces cerevisiae. Nucleic Acids Res. 23, 2799 –2800 35. Chalkley, R. J., Baker, P. R., Huang, L., Hansen, K. C., Allen, N. P., Rexach, M., and Burlingame, A. L. (2005) Comprehensive analysis of a multidimensional liquid chromatography mass spectrometry dataset acquired on a quadrupole selecting, quadrupole collision cell, time-of-flight mass spectrometer. II. New developments in Protein Prospector allow for reliable and comprehensive automatic analysis of large datasets. Mol. Cell Proteomics 4, 1194 –1204 36. Cronan, J. E. J. (1990) Biotination of proteins in vivo. A post-translational modification to label, purify, and study proteins. J. Biol. Chem. 265, 10327–10333 37. Ong, S. E., Blagoev, B., Kratchmarova, I., Kristensen, D. B., Steen, H., Pandey, A., and Mann, M. (2002) Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. Mol. Cell. Proteomics 1, 376 –386 38. Blagoev, B., Kratchmarova, I., Ong, S. E., Nielsen, M., Foster, L. J., and Mann, M. (2003) A proteomics strategy to elucidate functional proteinprotein interactions applied to EGF signaling. Nat. Biotechnol. 21, 315–318 39. Gruhler, A., Olsen, J. V., Mohammed, S., Mortensen, P., Faergeman, N. F., Mann, M., and Jensen, O. N. (2005) Quantitative phosphoproteomics applied to the yeast pheromone signaling pathway. Mol. Cell. Proteomics 4, 310 –327 40. Ong, S. E., Kratchmarova, I., and Mann, M. (2003) Properties of 13Csubstituted arginine in stable isotope labeling by amino acids in cell culture (SILAC). J. Proteome Res. 2, 173–181 41. Peng, J., Schwartz, D., Elias, J. E., Thoreen, C. C., Cheng, D., Marsischky, G., Roelofs, J., Finley, D., and Gygi, S. P. (2003) A proteomics approach to understanding protein ubiquitination. Nat. Biotechnol. 21, 921–926 42. Wohlschlegel, J. A., Johnson, E. S., Reed, S. I., and Yates, J. R., III (2004) Global analysis of protein sumoylation in Saccharomyces cerevisiae. J. Biol. Chem. 279, 45662– 45668 43. Denison, C., Rudner, A. D., Gerber, S. A., Bakalarski, C. E., Moazed, D., and Gygi, S. P. (2005) A proteomic strategy for gaining insights into protein sumoylation in yeast. Mol. Cell. Proteomics 4, 246 –254 44. Rezvani, K., Mee, M., Dawson, S., McIlhinney, J., Fujita, J., and Mayer, R. J. (2003) Proteasomal interactors control activities as diverse as the cell cycle and glutaminergic neurotransmission. Biochem. Soc. Trans. 31, 470 – 473 45. Luders, J., Demand, J., and Hohfeld, J. (2000) The ubiquitin-related BAG-1 provides a link between the molecular chaperones Hsc70/Hsp70 and the proteasome. J. Biol. Chem. 275, 4613– 4617 46. Wiederkehr, T., Bukau, B., and Buchberger, A. (2002) Protein turnover: a CHIP programmed for proteolysis. Curr. Biol. 12, R26 –R28 47. Imai, J., Maruya, M., Yashiroda, H., Yahara, I., and Tanaka, K. (2003) The molecular chaperone Hsp90 plays a role in the assembly and maintenance of the 26S proteasome. EMBO J. 22, 3557–3567 48. Tokumoto, T., Kondo, A., Miwa, J., Horiguchi, R., Tokumoto, M., Nagahama, Y., Okida, N., and Ishikawa, K. (2003) Regulated interaction between polypeptide chain elongation factor-1 complex with the 26S proteasome during Xenopus oocyte maturation. BMC Biochem. 4, 6 –15 49. Chuang, S. M., Chen, L., Lambertson, D., Anand, M., Kinzy, T. G., and Madura, K. (2005) Proteasome-mediated degradation of cotranslationally damaged proteins involves translation elongation factor 1A. Mol. Cell. Biol. 25, 403– 413 50. Chuang, S. M., and Madura, K. (2005) Saccharomyces cerevisiae Ubconjugating enzyme Ubc4 binds the proteasome in the presence of translationally-damaged proteins. Genetics 4, 1477–1484