Proteomics 2014, 14, 19–23
19
DOI 10.1002/pmic.201300315
TECHNICAL BRIEF
Single-step protease cleavage elution for identification of protein–protein interactions from GST pull-down and mass spectrometry Lin Luo, Nathan P. King, Jeremy C. Yeo, Alun Jones and Jennifer L. Stow Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, Australia
The study of protein–protein interactions is a major theme in biological disciplines. Pull-down or affinity-precipitation assays using GST fusion proteins have become one of the most common and valuable approaches to identify novel binding partners for proteins of interest (bait). Non-specific binding of prey proteins to the beads or to GST itself, however, inevitably complicates and impedes subsequent analysis of pull-down results. A variety of measures, each with inherent advantages and limitations, can minimise the extent of the background. This technical brief details and tests a modification of established GST pull-down protocols. By specifically eluting only the bait (minus the GST tag) and the associated non-specific binding proteins with a simple, single-step protease cleavage, a cleaner platform for downstream protein identification with MS is established. We present a proof of concept for this method, as evidenced by a GST pull-down/MS case study of the small guanosine triphosphatase (GTPase) Rab31 in which: (i) sensitivity was enhanced, (ii) a reduced level of background was observed, (iii) distinguishability of non-specific contaminant proteins from genuine binders was improved and (iv) a putative new protein–protein interaction was discovered. Our protease cleavage step is readily applicable to all further affinity tag pull-downs.
Received: July 25, 2013 Revised: October 10, 2013 Accepted: October 28, 2013
Keywords: GST pull-down / LC-MS/MS / PreScission protease / Protease cleavage elution / Technology
Additional supporting information may be found in the online version of this article at the publisher’s web-site
The GST pull-down has proven to be a powerful technique for the study of protein–protein interactions since its advent in the early 1990s [1], transforming the field. The system takes advantage of the high affinity and specificity of GST to reduced GSH, which is immobilised to a solid support. Specific antibodies are thus not required, making it an ideal alternative to immunoprecipitation, and in many cases a more
Correspondence: Dr Lin Luo, Institute for Molecular Bioscience, Queensland Bioscience Precinct, Building 80, The University of Queensland, Services Road, Brisbane, QLD 4072, Australia E-mail: [email protected] Fax: +61 7 3346 2101 Abbreviations: EEA1, early endosome antigen 1; GTP, guanosine triphosphate; LPS, lipopolysaccharide; PCE, protease cleavage elution C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
appropriate option. GST-fusion proteins are easily and efficiently expressed in bacteria, and can be purified in a single step [2]. GST pull-downs are typically employed as an initial in vitro discovery tool for seeking novel interactors of a particular protein of interest, or as a follow-up assay to test a suspected interaction [3]. Despite its wide use, the traditional application of pull-down assays has several limitations, in particular, interference from non-specific binding of contaminating proteins to the selected affinity matrix (i.e. agarose, sepharose or magnetic beads) [4]. In recent years, several approaches have been engaged to eliminate non-specific binding. Preclearance of the prey lysate with “empty” GSH beads is a standard step that removes some of the non-specific protein burden [3]. While this process risks removing any genuine target proteins that have a high affinity for the matrix, it is still commonly used. Another tactic is blocking the beads with a BSA solution before incubation with prey proteins [5]. The www.proteomics-journal.com
20
L. Luo et al.
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
PreScission protease recognition sequence:
Bead GSH
n
tio
u
g
el
lin
i Bo
n se tio ea elu ot Pr age ) CE av (P cle
benefits of this are somewhat levelled by the BSA (and BSA oligomers) contaminating the eluate. High stringency purification methods (high salt buffers, numerous washes) not only reduce the level of background, but also risk removal of low abundant and low affinity partner proteins [6]. SILAC [7] exhibits lower rates of non-specific binding, but the cost can be prohibitive. Once bait and prey proteins have been captured on a GSH matrix, the choice of elution method is an important consideration. The simplest way to release proteins for analysis is by boiling in reducing SDS-PAGE sample buffer. This not only provides total recovery of bound proteins, but also results in a high background of non-specific proteins and contamination by intact and degraded GST-fusion bait. On-bead tryptic digestion is a conventional approach applied in affinity precipitation studies, in particular when the bait is difficult to elute from the matrix [8, 9]. This technique is not widely used with GST pull-down experiments, as tryptic peptides derived from the bait can inhibit detection of less abundant peptides. More selective elution of captured proteins greatly improves the quality of proteomic data obtained from pull-downs. Proteins can be released by competitive displacement with reduced GSH, but then GST bait is present abundantly in the eluate. Tandem affinity purification approaches make use of multiple protein tags [10, 11]. A cleaner end result can be achieved, but tandem affinity purification is a relatively complex procedure, requiring two columns and successive purifications. Protein yields can decrease throughout the multi-step procedure and it is not ideal for detecting transient and low affinity interactions. Non-specific proteins can be filtered out post-MS based on known bead proteomes [12]. This approach however discounts the possibility that sticky proteins can also be valid binding partners, and non-specific proteins still contaminate the samples processed for MS. Commercially available vectors for protein purification generally feature a protease cleavage site for removal of the affinity tag. The commonly used pGEX-6P GST expression vectors (#28–9546–48, 50, 51; GE Healthcare, Sydney, Australia) feature a gene encoding GST from the parasite Schistosoma japonicum [13] and a cleavage site for the human rhinovirus 3C protease [14], marketed by GE Healthcare as PreScission protease. These vectors and the PreScission protease have been used extensively for recombinant protein production and purification [15–18]. Here, we introduce a novel application for these vectors and their integrated PreScission protease cleavage site: a simple, accessible, and powerful modification to existing GST pulldown protocols that can significantly improve the quality and integrity of data obtained from downstream MS analysis. A conceptual overview is presented in Fig. 1. Elution by boiling in SDS-PAGE sample buffer releases the GST fusion protein and any bound proteins, plus contaminants (including Escherichia coli proteins, if the bait was bacterially expressed) and non-specific protein from lysates from the beads. In contrast, selective elution by a protease only detaches the GSTfree bait and its interacting proteins.
Proteomics 2014, 14, 19–23
LEVLFQ GP
Bead
Bead
Affinity-tagged bait
Partially degraded affinity-tagged bait
PCE cleaved bait PCE cleaved, partially degraded bait Genuine binding partner
Genuine binding partner
Contaminant protein bound to affinity matrix Contaminant protein bound to GSH
Figure 1. A workflow comparison of conventional boiling method and the PCE method. The conventional boiling method releases all proteins from the affinity matrix. In contrast, the PCE method selectively releases only the tagless bait and specific binding partners.
To test the performance of this proposed protease cleavage elution (PCE) method, we compared it to conventional elution by boiling, using the mammalian Rab31 small GTPase (also known as Rab22b) as the model bait protein. Rab31 is a relatively uncharacterised member of the Rab5 GTPase subfamily [19], which also includes Rab21, Rab22a and Rab24. Rab31 is associated with post-Golgi trafficking, but its functions and complement of interacting proteins are as yet unknown. Rab31 has been shown to bind to early endosome antigen 1 (EEA1), with higher affinity compared to Rab5, due to its ability to bind to both Rab-binding domains of EEA1 [20]. Rab31 is highly expressed in immune cells, including the murine RAW 264.7 macrophage cell line [21], which was used in this study. Mouse Rab31 complementary DNA was cloned into the BamHI/EcoRI sites of pGEX-6P-1, and the GST-Rab31 fusion protein was expressed in E. coli, purified, and bound to GSH-sepharose 4B beads (#17–0756–05; GE Healthcare) by standard methods [22]. Six independent GST pull-downs from lipopolysaccharide (LPS)-activated RAW 264.7 mouse macrophage extracts using GST-Rab31 bait, and two GST control pull-downs, were performed as previously described [22]. Briefly, cell lysates were pre-cleared by incubation with GSH-sepharose beads. GTP-loaded GST-Rab31 sepharose beads were incubated with LPS-activated cell lysate for 1 h at 4⬚C with agitation. We used MicroSpin columns (#27– 3565–01; GE Healthcare) for all of the pull-downs. Beads were washed with ice-cold wash buffer (50 mM Tris, 150 mM NaCl, 1% NP-40, 1 mM PMSF, 1 mM DTT, pH 7.4). Elution was www.proteomics-journal.com
21
Proteomics 2014, 14, 19–23
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
A
Boiling elution Ctrl
#1
#2
Protease cleavage elution (PCE) #3
Ctrl
#1
EEA1
1
2
GST-Rab31 trimer
3
4
GST-Rab31 dimer
5
6
GST-Rab31
#2
#3 kDa 250 150 100 75
7
8
50
9
10
37
GST-Rab31 degradation GST
25 20
PCE-cleaved Rab31 12
11
15 10
Lane
1
2
3
4
Boiling elution
B
5
6
7
Post-cleavage beads
8
9
10
PCE
Input Control Rab31 Control Rab31 Control Rab31 EEA1 Immunoblot Actin GST-Rab31 Coomassie stain
Lane
C Relative intensity (%)
100
1
2
EEA1
3
4
5
6
7
GST PCE-cleaved Rab31
Actin
80 60 40 20
Bo il st B ed Po -cle oile con a st v d tro PC -cle age Rab l E av c 31 PC -cle age ont E- av R rol cl ed ab ea c 3 ve on 1 d tro Bo Rab l Po 3 ile st B d c 1 Po -cle oile on st av d tro PC -cle age Rab l E av c 31 PC -cle age ont E- av R rol cl ed ab ea c 3 ve on 1 d tro R l ab 31
0
Po
achieved conventionally by boiling in 3× SDS-PAGE sample buffer for 5 min, or by PCE for 1 h with (GST-tagged) PreScission protease (#27-0843-01; GE Healthcare), according to the manufacturer’s instructions. Eluted proteins were resolved on 7.5–15% gradient SDS-PAGE gels and stained overnight with colloidal CBB G250. Bands were excised from the stained gel and were processed as described previously [22]. Identification of proteins was performed by LC-MS/MS (nano-LCMS/MS) essentially as previously described [23], and detailed in Supporting Information. Our SDS-PAGE analysis revealed several advantages of PCE over the conventional boiling elution method. First, single-step proteolytic elution eliminated degraded GST-bait proteins between 25 to 40 kDa (Fig. 2A, lanes 3–5). Second, PCE also shifts the major band representing the Rab31 bait down from 37 to 20 kDa (lanes 3–5). This effectively avoids masking of potential genuine binding partners of a similar molecular mass to the GST-fusion bait protein itself. In addition, potential masking of proteins in a similar size range to the PCE-derived GST-free bait can be easily overcome by concomitantly processing the corresponding gel fragments from boiled elutions. Third, PCE also removes interference from GST oligomers (Fig. 2A). Such oligomers are absent in the PCE samples. Next, we used immunoblotting to compare the elution methods by analysing selected proteins pulled down by Rab31. Both elution methods yielded EEA1, a known Rab31 effector (Fig. 2B, also Fig. 1). Quantitative analysis of the Western blot showed that EEA1 was enriched more than fivefold in the PCE samples (Fig. 2C). Detection of cytoskeletal proteins, such as actin, even in high amounts, in MS outputs of GST pull-downs is usually overlooked due to its notoriety for being a ‘sticky’ contaminant. Our results from PCE showed an enrichment of actin in the GST-Rab31 pull-down elution (Fig. 2B, lane 7) but not in the GST control (Fig. 2B, lane 6; Fig. 2C). Therefore, our data suggest that actin is a genuine binding partner of Rab31. This is not surprising, since the closely related Rab5 GTPase is known to bind actin [24]. We concurrently analysed trypsin digestions of GST-Rab31 pull-down elutions by LC-MS/MS. Representative proteins are presented in Table 1, a comprehensive dataset is presented as Supporting Information Table 1. The MS results revealed an overall reduced background of non-specific proteins from the lysate. Included are previously reported matrix tethering metabolic enzymes, cytoskeleton proteins and ribosomal proteins [12] (e.g. mitochondrial aldehyde dehydrogenase, tubulin and 40S ribosomal protein S10), as well as bacterial proteins from the GST fusion protein purification procedure (e.g. E. coli 60 kDa chaperonin; Table 1 and Supporting Information Table 1). Notwithstanding, protein chaperones, such as DnaK, that non-specifically interact with bait protein cannot be eliminated by any affinity purification approach (Supporting Information Table 1). Moreover, the results demonstrated an enrichment of the genuine binding partner of Rab31, EEA1, with a higher MS protein
Figure 2. Identification of mammalian Rab31 binding partners by comparison of boiling elution and PCE. (A) Eight pull-downs from LPS-activated macrophage extracts with GST or GST-Rab31 bait were performed. Four samples were eluted by conventional boiling, and the other four samples were eluted by PCE. Samples were resolved on a 7–15% gradient SDS-PAGE gel and stained with colloidal CBB G250. Protein bands of note are indicated to the left. Excised gel pieces, labelled 1–12, were analysed by MS. (B) Western immunoblots of conventional boiling, post-PCE beads and PCE samples. The known Rab31 interactor EEA1 is pulled down by GST-Rab31 but not the GST controls. Actin is present in all elutions except the GST control eluted by protease cleavage, providing evidence that actin, although an acknowledged sticky protein, binds specifically to Rab31. The baits were visualised by CBB G250 staining. (C) Quantification of immunoblots by densitometry analysis. The band intensities were quantified using the inbuilt Western blot macro in ImageJ 1.4 (http://rsbweb.nih.gov/ij/), and are presented as percentages of protein band intensity relative to the corresponding baits as a representative from three independent experiments.
www.proteomics-journal.com
22
L. Luo et al.
Proteomics 2014, 14, 19–23
Table 1. Representative MS analysis of GST-Rab31 pull-downs by boiling elution and PCE
Boiled elution
PCE
Early endosome antigen 1 (Mus musculus) Uniprot ID: Q8BL66 Pull-down #1 #2 MS score 8.6 8.0 Number of peptides (95% confidence) 5 4 Sequence coverage (95% confidence) 5.7% 4.8%
#3 5.2 3 3.7%
#1 59.9 36 31.4%
#2 20.9 11 11.2%
#3 16.1 8 8.6%
Aldehyde dehydrogenase, mitochondrial (M. musculus) Uniprot ID: P47738 Pull-down #1 #2 MS score 10.0 6.0 Number of peptides (95% confidence) 5 3 Sequence coverage (95% confidence) 12.9% 8.7%
#3 6.0 4 8.7%
#1 0 0 0%
#2 2.0 1 3.3%
#3 0 0 0%
60 kDa chaperonin (Escherichia coli) Uniprot ID: P0A6F5 Pull-down #1 MS score 5.1 Number of peptides (95% confidence) 3 Sequence coverage (95% confidence) 9.7%
#3 4.0 3 9.7%
#1 0 0 0%
#2 0 0 0%
#3 0 0 0%
#3 0 0 0%
#1 24.2 17 42.1%
#2 2.6 2 7.7%
#3 6.0 6 15.5%
#2 7.7 5 12.8%
Actin, cytoplasmic 1, 2 (M. musculus) Uniprot ID: P60710, P63260 Pull-down #1 #2 MS score 0 0 Number of peptides (95% confidence) 0 0 Sequence coverage (95% confidence) 0% 0%
The MS score is a measurement of peptide confidence from the ProteinPilot scoring algorithm: Score = −log(1 − (percent confidence/100)). For example, a score of 2 = 99% confidence, whereas a score of 4 = 99.99% confidence.
identification score, more identified peptides and a larger EEA1 amino acid coverage (Table 1). Cytoplasmic actin (42 kDa) was not detected in any of the three boiled elutions (Table 1), since it was completely masked by the GST-Rab31 bait protein band on the gel (Fig. 2A). We thus show that PCE of samples can also be used to identify bona fide binding partners that may otherwise be disregarded as ‘sticky’ proteins. Here, we describe and validate a simple, single-step proteolytic elution method for improved identification of protein– protein interactions using the classical GST pull-down technique coupled to MS. Our results illustrate the advantages of using PCE as a protocol modification, summarised as follows: (i) There is significantly less background due to the retention of GST on the beads after elution. (ii) Sample pollution from degraded bait and GST oligomers is eliminated. (iii) This is a useful tool for detecting authentic binders that are also acknowledged “sticky” proteins. (iv) This is a reproducible method and is ideal for highthroughput analyses. (v) This strategy is versatile and uncomplicated and can be applied to alternative vectors and other proteases. Indeed, vectors can easily be modified to incorporate a protease cleavage site for these and other purposes, including immunoprecipitation. PCE thus has the capacity to effectively diminish both false-positive and false-negative results, without compromis C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ing sensitivity. We recommend the adoption of this method modification for investigations of novel protein–protein interactions by means of affinity tag pull-downs. We acknowledge the Mass Spectrometry Facility and the Facility for Life Science Automation at the Institute for Molecular Bioscience, The University of Queensland. This work was supported by funding from an Australian National Health and Medical Research Council Program Grant (606788) (JLS). The authors declare no conflicts of interest.
References [1] Kaelin, W. G., Pallas, D. C., Decaprio, J. A., Kaye, F. J. et al., Identification of cellular proteins that can interact specifically with the T/E1A-binding region of the retinoblastoma gene product. Cell 1991, 64, 521–532. [2] Tolia, N. H., Joshua-Tor, L., Strategies for protein coexpression in Escherichia coli. Nat. Meth. 2006, 3, 55–64. [3] Sambrook, J., Russell, D. W., Detection of protein-protein interactions using the GST fusion protein pulldown technique. Protocol 3, Cold Spring Harbor Protocols, USA, 2006, pp. 18.55–18.59. [4] ten Have, S., Boulon, S., Ahmad, Y., Lamond, A. I., Mass spectrometry-based immuno-precipitation proteomics—the user’s guide. Proteomics 2011, 11, 1153–1159. [5] Vos, L. J., Famulski, J. K., Chan, G. K. T., hZwint-1 bridges the inner and outer kinetochore: identification of the kinetochore localization domain and the hZw10-interaction domain. Biochem. J. 2011, 436, 157–168.
www.proteomics-journal.com
Proteomics 2014, 14, 19–23 [6] Berggard, T., Linse, S., James, P., Methods for the detection and analysis of protein-protein interactions. Proteomics 2007, 7, 2833–2842. [7] Ong, S. E., The expanding field of SILAC. Anal. Bioanal. Chem. 2012, 404, 967–976. [8] Fukuyama, H., Ndiaye, S., Hoffmann, J., Rossier, J. et al., On-bead tryptic proteolysis: an attractive procedure for LC-MS/MS analysis of the Drosophila caspase 8 protein complex during immune response against bacteria. J. Proteomics 2012, 75, 4610–4619. [9] Fonovic, M., Verhelst, S. H. L., Sorum, M. T., Bogyo, M., Proteomics evaluation of chemically cleavable activity-based probes. Mol. Cell. Proteomics 2007, 6, 1761–1770. [10] Miteva, Y. V., Budayeva, H. G., Cristea, I. M., Proteomicsbased methods for discovery, quantification, and validation of protein-protein interactions. Anal. Chem. 2013, 85, 749–768.
23 man tissue transglutaminase. Protein Expr. Purif. 2013, 87, 41–46. [16] Jerlstrom-Hultqvist, J., Stadelmann, B., Birkestedt, S., Hellman, U. et al., Plasmid vectors for proteomic analyses in Giardia: Purification of virulence factors and analysis of the proteasome. Eukaryot. Cell 2012, 11, 864–873. [17] Al-Sheikh, A., Ahmad, S., Khan, Z. U., Modification of an expression vector for efficient recombinant production and purification of mitogillin of Aspergillus fumigatus expressed in Escherichia coli. Protein Expr. Purif. 2011, 76, 90–96. [18] Liu, X., He, Z., Zhou, M., Yang, F. et al., Purification and characterization of recombinant extracellular domain of human HER2 from Escherichia coli. Protein Expr. Purif. 2007, 53, 247–254. [19] Hutagalung, A. H., Novick, P. J., Role of Rab GTPases in membrane traffic and cell physiology. Physiol. Rev. 2011, 91, 119–149.
[11] Rigaut, G., Shevchenko, A., Rutz, B., Wilm, M. et al., A generic protein purification method for protein complex characterization and proteome exploration. Nat. Biotechnol. 1999, 17, 1030–1032.
[20] Lodhi, I. J., Chiang, S.-H., Chang, L., Vollenweider, D. et al., Gapex-5, a Rab31 guanine nucleotide exchange factor that regulates Glut4 trafficking in adipocytes. Cell Metab. 2007, 5, 59–72.
[12] Trinkle-Mulcahy, L., Boulon, S., Lam, Y. W., Urcia, R. et al., Identifying specific protein interaction partners using quantitative mass spectrometry and bead proteomes. J. Cell Biol. 2008, 183, 223–239.
[21] Diekmann, Y., Seixas, E., Gouw, M., Tavares-Cadete, F. et al., Thousands of Rab GTPases for the cell biologist. PLoS Comput. Biol. 2011, 7, e1002217.
[13] Smith, D. B., Johnson, K. S., Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 1988, 67, 31–40. [14] Walker, P. A., Leong, L. E. C., Ng, P. W. P., Tan, S. H. et al., Efficient and rapid affinity purification of proteins using recombinant fusion proteases. Bio-Technology 1994, 12, 601–605. [15] Roy, I., Smith, O., Clouthier, C. M., Keillor, J. W., Expression, purification and kinetic characterisation of hu-
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[22] Brymora, A., Valova, V. A., Robinson, P. J., Protein-protein interactions identified by pull-down experiments and mass spectrometry. Curr. Protoc. Cell Biol. 2004, 22, 17.5.1–17.5.51. [23] Lewis, R. J., Yang, A., Jones, A., Rapid extraction combined with LC-tandem mass spectrometry (CREM-LC/MS/MS) for the determination of ciguatoxins in ciguateric fish flesh. Toxicon 2009, 54, 62–66. [24] Lanzetti, L., Palamidessi, A., Areces, L., Scita, G. et al., Rab5 is a signalling GTPase involved in actin remodelling by receptor tyrosine kinases. Nature 2004, 429, 309–314.
www.proteomics-journal.com
19
DOI 10.1002/pmic.201300315
TECHNICAL BRIEF
Single-step protease cleavage elution for identification of protein–protein interactions from GST pull-down and mass spectrometry Lin Luo, Nathan P. King, Jeremy C. Yeo, Alun Jones and Jennifer L. Stow Institute for Molecular Bioscience, The University of Queensland, Brisbane, Queensland, Australia
The study of protein–protein interactions is a major theme in biological disciplines. Pull-down or affinity-precipitation assays using GST fusion proteins have become one of the most common and valuable approaches to identify novel binding partners for proteins of interest (bait). Non-specific binding of prey proteins to the beads or to GST itself, however, inevitably complicates and impedes subsequent analysis of pull-down results. A variety of measures, each with inherent advantages and limitations, can minimise the extent of the background. This technical brief details and tests a modification of established GST pull-down protocols. By specifically eluting only the bait (minus the GST tag) and the associated non-specific binding proteins with a simple, single-step protease cleavage, a cleaner platform for downstream protein identification with MS is established. We present a proof of concept for this method, as evidenced by a GST pull-down/MS case study of the small guanosine triphosphatase (GTPase) Rab31 in which: (i) sensitivity was enhanced, (ii) a reduced level of background was observed, (iii) distinguishability of non-specific contaminant proteins from genuine binders was improved and (iv) a putative new protein–protein interaction was discovered. Our protease cleavage step is readily applicable to all further affinity tag pull-downs.
Received: July 25, 2013 Revised: October 10, 2013 Accepted: October 28, 2013
Keywords: GST pull-down / LC-MS/MS / PreScission protease / Protease cleavage elution / Technology
Additional supporting information may be found in the online version of this article at the publisher’s web-site
The GST pull-down has proven to be a powerful technique for the study of protein–protein interactions since its advent in the early 1990s [1], transforming the field. The system takes advantage of the high affinity and specificity of GST to reduced GSH, which is immobilised to a solid support. Specific antibodies are thus not required, making it an ideal alternative to immunoprecipitation, and in many cases a more
Correspondence: Dr Lin Luo, Institute for Molecular Bioscience, Queensland Bioscience Precinct, Building 80, The University of Queensland, Services Road, Brisbane, QLD 4072, Australia E-mail: [email protected] Fax: +61 7 3346 2101 Abbreviations: EEA1, early endosome antigen 1; GTP, guanosine triphosphate; LPS, lipopolysaccharide; PCE, protease cleavage elution C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
appropriate option. GST-fusion proteins are easily and efficiently expressed in bacteria, and can be purified in a single step [2]. GST pull-downs are typically employed as an initial in vitro discovery tool for seeking novel interactors of a particular protein of interest, or as a follow-up assay to test a suspected interaction [3]. Despite its wide use, the traditional application of pull-down assays has several limitations, in particular, interference from non-specific binding of contaminating proteins to the selected affinity matrix (i.e. agarose, sepharose or magnetic beads) [4]. In recent years, several approaches have been engaged to eliminate non-specific binding. Preclearance of the prey lysate with “empty” GSH beads is a standard step that removes some of the non-specific protein burden [3]. While this process risks removing any genuine target proteins that have a high affinity for the matrix, it is still commonly used. Another tactic is blocking the beads with a BSA solution before incubation with prey proteins [5]. The www.proteomics-journal.com
20
L. Luo et al.
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
PreScission protease recognition sequence:
Bead GSH
n
tio
u
g
el
lin
i Bo
n se tio ea elu ot Pr age ) CE av (P cle
benefits of this are somewhat levelled by the BSA (and BSA oligomers) contaminating the eluate. High stringency purification methods (high salt buffers, numerous washes) not only reduce the level of background, but also risk removal of low abundant and low affinity partner proteins [6]. SILAC [7] exhibits lower rates of non-specific binding, but the cost can be prohibitive. Once bait and prey proteins have been captured on a GSH matrix, the choice of elution method is an important consideration. The simplest way to release proteins for analysis is by boiling in reducing SDS-PAGE sample buffer. This not only provides total recovery of bound proteins, but also results in a high background of non-specific proteins and contamination by intact and degraded GST-fusion bait. On-bead tryptic digestion is a conventional approach applied in affinity precipitation studies, in particular when the bait is difficult to elute from the matrix [8, 9]. This technique is not widely used with GST pull-down experiments, as tryptic peptides derived from the bait can inhibit detection of less abundant peptides. More selective elution of captured proteins greatly improves the quality of proteomic data obtained from pull-downs. Proteins can be released by competitive displacement with reduced GSH, but then GST bait is present abundantly in the eluate. Tandem affinity purification approaches make use of multiple protein tags [10, 11]. A cleaner end result can be achieved, but tandem affinity purification is a relatively complex procedure, requiring two columns and successive purifications. Protein yields can decrease throughout the multi-step procedure and it is not ideal for detecting transient and low affinity interactions. Non-specific proteins can be filtered out post-MS based on known bead proteomes [12]. This approach however discounts the possibility that sticky proteins can also be valid binding partners, and non-specific proteins still contaminate the samples processed for MS. Commercially available vectors for protein purification generally feature a protease cleavage site for removal of the affinity tag. The commonly used pGEX-6P GST expression vectors (#28–9546–48, 50, 51; GE Healthcare, Sydney, Australia) feature a gene encoding GST from the parasite Schistosoma japonicum [13] and a cleavage site for the human rhinovirus 3C protease [14], marketed by GE Healthcare as PreScission protease. These vectors and the PreScission protease have been used extensively for recombinant protein production and purification [15–18]. Here, we introduce a novel application for these vectors and their integrated PreScission protease cleavage site: a simple, accessible, and powerful modification to existing GST pulldown protocols that can significantly improve the quality and integrity of data obtained from downstream MS analysis. A conceptual overview is presented in Fig. 1. Elution by boiling in SDS-PAGE sample buffer releases the GST fusion protein and any bound proteins, plus contaminants (including Escherichia coli proteins, if the bait was bacterially expressed) and non-specific protein from lysates from the beads. In contrast, selective elution by a protease only detaches the GSTfree bait and its interacting proteins.
Proteomics 2014, 14, 19–23
LEVLFQ GP
Bead
Bead
Affinity-tagged bait
Partially degraded affinity-tagged bait
PCE cleaved bait PCE cleaved, partially degraded bait Genuine binding partner
Genuine binding partner
Contaminant protein bound to affinity matrix Contaminant protein bound to GSH
Figure 1. A workflow comparison of conventional boiling method and the PCE method. The conventional boiling method releases all proteins from the affinity matrix. In contrast, the PCE method selectively releases only the tagless bait and specific binding partners.
To test the performance of this proposed protease cleavage elution (PCE) method, we compared it to conventional elution by boiling, using the mammalian Rab31 small GTPase (also known as Rab22b) as the model bait protein. Rab31 is a relatively uncharacterised member of the Rab5 GTPase subfamily [19], which also includes Rab21, Rab22a and Rab24. Rab31 is associated with post-Golgi trafficking, but its functions and complement of interacting proteins are as yet unknown. Rab31 has been shown to bind to early endosome antigen 1 (EEA1), with higher affinity compared to Rab5, due to its ability to bind to both Rab-binding domains of EEA1 [20]. Rab31 is highly expressed in immune cells, including the murine RAW 264.7 macrophage cell line [21], which was used in this study. Mouse Rab31 complementary DNA was cloned into the BamHI/EcoRI sites of pGEX-6P-1, and the GST-Rab31 fusion protein was expressed in E. coli, purified, and bound to GSH-sepharose 4B beads (#17–0756–05; GE Healthcare) by standard methods [22]. Six independent GST pull-downs from lipopolysaccharide (LPS)-activated RAW 264.7 mouse macrophage extracts using GST-Rab31 bait, and two GST control pull-downs, were performed as previously described [22]. Briefly, cell lysates were pre-cleared by incubation with GSH-sepharose beads. GTP-loaded GST-Rab31 sepharose beads were incubated with LPS-activated cell lysate for 1 h at 4⬚C with agitation. We used MicroSpin columns (#27– 3565–01; GE Healthcare) for all of the pull-downs. Beads were washed with ice-cold wash buffer (50 mM Tris, 150 mM NaCl, 1% NP-40, 1 mM PMSF, 1 mM DTT, pH 7.4). Elution was www.proteomics-journal.com
21
Proteomics 2014, 14, 19–23
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
A
Boiling elution Ctrl
#1
#2
Protease cleavage elution (PCE) #3
Ctrl
#1
EEA1
1
2
GST-Rab31 trimer
3
4
GST-Rab31 dimer
5
6
GST-Rab31
#2
#3 kDa 250 150 100 75
7
8
50
9
10
37
GST-Rab31 degradation GST
25 20
PCE-cleaved Rab31 12
11
15 10
Lane
1
2
3
4
Boiling elution
B
5
6
7
Post-cleavage beads
8
9
10
PCE
Input Control Rab31 Control Rab31 Control Rab31 EEA1 Immunoblot Actin GST-Rab31 Coomassie stain
Lane
C Relative intensity (%)
100
1
2
EEA1
3
4
5
6
7
GST PCE-cleaved Rab31
Actin
80 60 40 20
Bo il st B ed Po -cle oile con a st v d tro PC -cle age Rab l E av c 31 PC -cle age ont E- av R rol cl ed ab ea c 3 ve on 1 d tro Bo Rab l Po 3 ile st B d c 1 Po -cle oile on st av d tro PC -cle age Rab l E av c 31 PC -cle age ont E- av R rol cl ed ab ea c 3 ve on 1 d tro R l ab 31
0
Po
achieved conventionally by boiling in 3× SDS-PAGE sample buffer for 5 min, or by PCE for 1 h with (GST-tagged) PreScission protease (#27-0843-01; GE Healthcare), according to the manufacturer’s instructions. Eluted proteins were resolved on 7.5–15% gradient SDS-PAGE gels and stained overnight with colloidal CBB G250. Bands were excised from the stained gel and were processed as described previously [22]. Identification of proteins was performed by LC-MS/MS (nano-LCMS/MS) essentially as previously described [23], and detailed in Supporting Information. Our SDS-PAGE analysis revealed several advantages of PCE over the conventional boiling elution method. First, single-step proteolytic elution eliminated degraded GST-bait proteins between 25 to 40 kDa (Fig. 2A, lanes 3–5). Second, PCE also shifts the major band representing the Rab31 bait down from 37 to 20 kDa (lanes 3–5). This effectively avoids masking of potential genuine binding partners of a similar molecular mass to the GST-fusion bait protein itself. In addition, potential masking of proteins in a similar size range to the PCE-derived GST-free bait can be easily overcome by concomitantly processing the corresponding gel fragments from boiled elutions. Third, PCE also removes interference from GST oligomers (Fig. 2A). Such oligomers are absent in the PCE samples. Next, we used immunoblotting to compare the elution methods by analysing selected proteins pulled down by Rab31. Both elution methods yielded EEA1, a known Rab31 effector (Fig. 2B, also Fig. 1). Quantitative analysis of the Western blot showed that EEA1 was enriched more than fivefold in the PCE samples (Fig. 2C). Detection of cytoskeletal proteins, such as actin, even in high amounts, in MS outputs of GST pull-downs is usually overlooked due to its notoriety for being a ‘sticky’ contaminant. Our results from PCE showed an enrichment of actin in the GST-Rab31 pull-down elution (Fig. 2B, lane 7) but not in the GST control (Fig. 2B, lane 6; Fig. 2C). Therefore, our data suggest that actin is a genuine binding partner of Rab31. This is not surprising, since the closely related Rab5 GTPase is known to bind actin [24]. We concurrently analysed trypsin digestions of GST-Rab31 pull-down elutions by LC-MS/MS. Representative proteins are presented in Table 1, a comprehensive dataset is presented as Supporting Information Table 1. The MS results revealed an overall reduced background of non-specific proteins from the lysate. Included are previously reported matrix tethering metabolic enzymes, cytoskeleton proteins and ribosomal proteins [12] (e.g. mitochondrial aldehyde dehydrogenase, tubulin and 40S ribosomal protein S10), as well as bacterial proteins from the GST fusion protein purification procedure (e.g. E. coli 60 kDa chaperonin; Table 1 and Supporting Information Table 1). Notwithstanding, protein chaperones, such as DnaK, that non-specifically interact with bait protein cannot be eliminated by any affinity purification approach (Supporting Information Table 1). Moreover, the results demonstrated an enrichment of the genuine binding partner of Rab31, EEA1, with a higher MS protein
Figure 2. Identification of mammalian Rab31 binding partners by comparison of boiling elution and PCE. (A) Eight pull-downs from LPS-activated macrophage extracts with GST or GST-Rab31 bait were performed. Four samples were eluted by conventional boiling, and the other four samples were eluted by PCE. Samples were resolved on a 7–15% gradient SDS-PAGE gel and stained with colloidal CBB G250. Protein bands of note are indicated to the left. Excised gel pieces, labelled 1–12, were analysed by MS. (B) Western immunoblots of conventional boiling, post-PCE beads and PCE samples. The known Rab31 interactor EEA1 is pulled down by GST-Rab31 but not the GST controls. Actin is present in all elutions except the GST control eluted by protease cleavage, providing evidence that actin, although an acknowledged sticky protein, binds specifically to Rab31. The baits were visualised by CBB G250 staining. (C) Quantification of immunoblots by densitometry analysis. The band intensities were quantified using the inbuilt Western blot macro in ImageJ 1.4 (http://rsbweb.nih.gov/ij/), and are presented as percentages of protein band intensity relative to the corresponding baits as a representative from three independent experiments.
www.proteomics-journal.com
22
L. Luo et al.
Proteomics 2014, 14, 19–23
Table 1. Representative MS analysis of GST-Rab31 pull-downs by boiling elution and PCE
Boiled elution
PCE
Early endosome antigen 1 (Mus musculus) Uniprot ID: Q8BL66 Pull-down #1 #2 MS score 8.6 8.0 Number of peptides (95% confidence) 5 4 Sequence coverage (95% confidence) 5.7% 4.8%
#3 5.2 3 3.7%
#1 59.9 36 31.4%
#2 20.9 11 11.2%
#3 16.1 8 8.6%
Aldehyde dehydrogenase, mitochondrial (M. musculus) Uniprot ID: P47738 Pull-down #1 #2 MS score 10.0 6.0 Number of peptides (95% confidence) 5 3 Sequence coverage (95% confidence) 12.9% 8.7%
#3 6.0 4 8.7%
#1 0 0 0%
#2 2.0 1 3.3%
#3 0 0 0%
60 kDa chaperonin (Escherichia coli) Uniprot ID: P0A6F5 Pull-down #1 MS score 5.1 Number of peptides (95% confidence) 3 Sequence coverage (95% confidence) 9.7%
#3 4.0 3 9.7%
#1 0 0 0%
#2 0 0 0%
#3 0 0 0%
#3 0 0 0%
#1 24.2 17 42.1%
#2 2.6 2 7.7%
#3 6.0 6 15.5%
#2 7.7 5 12.8%
Actin, cytoplasmic 1, 2 (M. musculus) Uniprot ID: P60710, P63260 Pull-down #1 #2 MS score 0 0 Number of peptides (95% confidence) 0 0 Sequence coverage (95% confidence) 0% 0%
The MS score is a measurement of peptide confidence from the ProteinPilot scoring algorithm: Score = −log(1 − (percent confidence/100)). For example, a score of 2 = 99% confidence, whereas a score of 4 = 99.99% confidence.
identification score, more identified peptides and a larger EEA1 amino acid coverage (Table 1). Cytoplasmic actin (42 kDa) was not detected in any of the three boiled elutions (Table 1), since it was completely masked by the GST-Rab31 bait protein band on the gel (Fig. 2A). We thus show that PCE of samples can also be used to identify bona fide binding partners that may otherwise be disregarded as ‘sticky’ proteins. Here, we describe and validate a simple, single-step proteolytic elution method for improved identification of protein– protein interactions using the classical GST pull-down technique coupled to MS. Our results illustrate the advantages of using PCE as a protocol modification, summarised as follows: (i) There is significantly less background due to the retention of GST on the beads after elution. (ii) Sample pollution from degraded bait and GST oligomers is eliminated. (iii) This is a useful tool for detecting authentic binders that are also acknowledged “sticky” proteins. (iv) This is a reproducible method and is ideal for highthroughput analyses. (v) This strategy is versatile and uncomplicated and can be applied to alternative vectors and other proteases. Indeed, vectors can easily be modified to incorporate a protease cleavage site for these and other purposes, including immunoprecipitation. PCE thus has the capacity to effectively diminish both false-positive and false-negative results, without compromis C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
ing sensitivity. We recommend the adoption of this method modification for investigations of novel protein–protein interactions by means of affinity tag pull-downs. We acknowledge the Mass Spectrometry Facility and the Facility for Life Science Automation at the Institute for Molecular Bioscience, The University of Queensland. This work was supported by funding from an Australian National Health and Medical Research Council Program Grant (606788) (JLS). The authors declare no conflicts of interest.
References [1] Kaelin, W. G., Pallas, D. C., Decaprio, J. A., Kaye, F. J. et al., Identification of cellular proteins that can interact specifically with the T/E1A-binding region of the retinoblastoma gene product. Cell 1991, 64, 521–532. [2] Tolia, N. H., Joshua-Tor, L., Strategies for protein coexpression in Escherichia coli. Nat. Meth. 2006, 3, 55–64. [3] Sambrook, J., Russell, D. W., Detection of protein-protein interactions using the GST fusion protein pulldown technique. Protocol 3, Cold Spring Harbor Protocols, USA, 2006, pp. 18.55–18.59. [4] ten Have, S., Boulon, S., Ahmad, Y., Lamond, A. I., Mass spectrometry-based immuno-precipitation proteomics—the user’s guide. Proteomics 2011, 11, 1153–1159. [5] Vos, L. J., Famulski, J. K., Chan, G. K. T., hZwint-1 bridges the inner and outer kinetochore: identification of the kinetochore localization domain and the hZw10-interaction domain. Biochem. J. 2011, 436, 157–168.
www.proteomics-journal.com
Proteomics 2014, 14, 19–23 [6] Berggard, T., Linse, S., James, P., Methods for the detection and analysis of protein-protein interactions. Proteomics 2007, 7, 2833–2842. [7] Ong, S. E., The expanding field of SILAC. Anal. Bioanal. Chem. 2012, 404, 967–976. [8] Fukuyama, H., Ndiaye, S., Hoffmann, J., Rossier, J. et al., On-bead tryptic proteolysis: an attractive procedure for LC-MS/MS analysis of the Drosophila caspase 8 protein complex during immune response against bacteria. J. Proteomics 2012, 75, 4610–4619. [9] Fonovic, M., Verhelst, S. H. L., Sorum, M. T., Bogyo, M., Proteomics evaluation of chemically cleavable activity-based probes. Mol. Cell. Proteomics 2007, 6, 1761–1770. [10] Miteva, Y. V., Budayeva, H. G., Cristea, I. M., Proteomicsbased methods for discovery, quantification, and validation of protein-protein interactions. Anal. Chem. 2013, 85, 749–768.
23 man tissue transglutaminase. Protein Expr. Purif. 2013, 87, 41–46. [16] Jerlstrom-Hultqvist, J., Stadelmann, B., Birkestedt, S., Hellman, U. et al., Plasmid vectors for proteomic analyses in Giardia: Purification of virulence factors and analysis of the proteasome. Eukaryot. Cell 2012, 11, 864–873. [17] Al-Sheikh, A., Ahmad, S., Khan, Z. U., Modification of an expression vector for efficient recombinant production and purification of mitogillin of Aspergillus fumigatus expressed in Escherichia coli. Protein Expr. Purif. 2011, 76, 90–96. [18] Liu, X., He, Z., Zhou, M., Yang, F. et al., Purification and characterization of recombinant extracellular domain of human HER2 from Escherichia coli. Protein Expr. Purif. 2007, 53, 247–254. [19] Hutagalung, A. H., Novick, P. J., Role of Rab GTPases in membrane traffic and cell physiology. Physiol. Rev. 2011, 91, 119–149.
[11] Rigaut, G., Shevchenko, A., Rutz, B., Wilm, M. et al., A generic protein purification method for protein complex characterization and proteome exploration. Nat. Biotechnol. 1999, 17, 1030–1032.
[20] Lodhi, I. J., Chiang, S.-H., Chang, L., Vollenweider, D. et al., Gapex-5, a Rab31 guanine nucleotide exchange factor that regulates Glut4 trafficking in adipocytes. Cell Metab. 2007, 5, 59–72.
[12] Trinkle-Mulcahy, L., Boulon, S., Lam, Y. W., Urcia, R. et al., Identifying specific protein interaction partners using quantitative mass spectrometry and bead proteomes. J. Cell Biol. 2008, 183, 223–239.
[21] Diekmann, Y., Seixas, E., Gouw, M., Tavares-Cadete, F. et al., Thousands of Rab GTPases for the cell biologist. PLoS Comput. Biol. 2011, 7, e1002217.
[13] Smith, D. B., Johnson, K. S., Single-step purification of polypeptides expressed in Escherichia coli as fusions with glutathione S-transferase. Gene 1988, 67, 31–40. [14] Walker, P. A., Leong, L. E. C., Ng, P. W. P., Tan, S. H. et al., Efficient and rapid affinity purification of proteins using recombinant fusion proteases. Bio-Technology 1994, 12, 601–605. [15] Roy, I., Smith, O., Clouthier, C. M., Keillor, J. W., Expression, purification and kinetic characterisation of hu-
C 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
[22] Brymora, A., Valova, V. A., Robinson, P. J., Protein-protein interactions identified by pull-down experiments and mass spectrometry. Curr. Protoc. Cell Biol. 2004, 22, 17.5.1–17.5.51. [23] Lewis, R. J., Yang, A., Jones, A., Rapid extraction combined with LC-tandem mass spectrometry (CREM-LC/MS/MS) for the determination of ciguatoxins in ciguateric fish flesh. Toxicon 2009, 54, 62–66. [24] Lanzetti, L., Palamidessi, A., Areces, L., Scita, G. et al., Rab5 is a signalling GTPase involved in actin remodelling by receptor tyrosine kinases. Nature 2004, 429, 309–314.
www.proteomics-journal.com
