NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Zhu MX, editor. TRP Channels. Boca Raton (FL): CRC Press/Taylor & Francis; 2011.

Cover of TRP Channels

TRP Channels.

Show details

Chapter 5Proteomic Analysis of TRPC Channels

, , , , and .

5.1. INTRODUCTION

Ca2+ is a ubiquitous and fundamental signaling component that is utilized by cells to regulate a diverse range of critical cellular functions. Typically, cells respond to a Ca2+ signal that is generated inside the cell in response to activation of a wide variety of cell surface receptors, including those involved in neurotransmitter, hormonal, and sensory signaling. In most cases, the initial Ca2+ signal generated in the cell is a specific increase in cytoplasmic [Ca2+] ([Ca2+]i) resulting from release of Ca2+ from internal Ca2+ stores (mainly the endoplasmic reticulum [ER]) or entry of Ca2+ from the external medium across the plasma membrane. Both routes involve movement of Ca2+ through Ca2+ channels that are localized within these cellular membranes. While intracellular Ca2+ release from ER occurs via channels activated by inositol 1,4,5-trisphosphate (IP3), cyclic ADP-ribose, or Ca2+ itself, Ca2+ influx across the plasma membrane is achieved via numerous types of Ca2+ channels, including voltage-gated Ca2+ channels and store-operated Ca2+ channels, as well as a variety of ligand-gated cation channels, although the type of channels can vary depending on the cell type.13Among these, the transient receptor potential (TRP) superfamily of ion channels have been described to be involved in a diverse array of signaling mechanisms that regulate critical sensory functions in cells as well as other processes such as secretion, proliferation, neuronal guidance, cell death, and development.

This article will focus on the TRPC subfamily of Ca2+-permeable channels that are activated in response to stimulation of G-protein-coupled receptors linked to phosphatidylinositol 4,5-bisphosphate (PIP2) hydrolysis.4,5 Ca2+ entry via TRPC channels regulates key cellular functions in a variety of cell types, including those from skeletal and cardiac muscle, exocrine, endocrine, endothelial, neuronal, epithelial, and smooth muscle cells.6 Given the importance of these channels in cell functions, it is critical to resolve the mechanism(s) involved in the regulation of the TRPC channel function. TRPC channels can be activated by several different mechanisms following neurotransmitter stimulation of cells. Diacylglycerol generated from PIP2 hydrolysis serves as a messenger to activate TRPC3, TRPC6, and TRPC7. Ca2+ itself has been shown to directly activate TRPC4 and TRPC5. Furthermore, TRPC3, TRPC5, and TRPC6 are dynamically recruited to the plasma membrane following stimulation.4 TRPC1 and TRPC6 have also been reported to be activated by cell stretching.5 Another mechanism proposed for TRPC1, TRPC3, and TRPC4 is by internal Ca2+ store depletion. In this mechanism, release of Ca2+ from intracellular Ca2+ stores is sensed by the ER-Ca2+ sensor protein STIM1, which translocates to peripheral regions of the cells where it interacts with and activates specific channels. Examples of these channels include CRAC channels (a major component of which is Orai1) and TRPC channels (e.g., TRPC1, TRPC4, and TRPC3). Indeed, STIM1–TRPC1 complex is formed when cells are stimulated.6 Orai1 has also been detected in this complex.6 Thus, the assembly of store-operated channels is a highly complicated, spatially/temporally coordinated process that involves several different functionally distinct proteins. It is highly likely that there are other as yet unidentified components of this complex that are critical for the activation and regulation of store-operated Ca2+ entry (SOCE).

Much of the initial insights into TRPC protein interactions have been derived from studies with the Drosophila TRP channel, which belongs to the TRPC subfamily and was the first TRP channel to be identified.3 The channel is localized in the Drosophila eye where it has a critical role in phototransduction. It was shown that this TRP channel resides in a multiprotein signalplex. Both TRP–TRP interactions and TRP interactions with other proteins in the signaling complex are important for proper channel activity and regulation of phototransduction.3 The Drosophila TRP forms a dynamic complex with scaffolding (e.g., INAD) and signaling proteins (e.g., phospholipase and calmodulin). INAD forms the core of this complex because it has the ability, via multiple PDZ domains, to bind to numerous signaling proteins and serve as a platform for their interaction with TRP and regulation of the channel function. Critical protein sequences, conserved in TRP channel families, appear to be involved in these various specific protein–protein interactions. These include the coiled–coiled domain (TRP–TRP interaction) as well as the ankyrin-like repeat region (TRP-signaling protein interactions), calmodulin-and lipid-binding domains, as well as other less well-characterized protein sequences. Because mammalian TRPC proteins share many of the same structural signatures with the Drosophila TRP channel, it was hypothesized that these proteins also share the property of homomeric or heteromeric interactions with other TRPC channels and signaling proteins. It is now well established that a number of proteins interacting with mammalian TRPC channels are crucial for their function and regulation.7 Qualitative and quantitative differences in the protein components associated with different TRPC channels have not yet been clearly described.

Initial studies for assessing protein–protein interactions involving TRPC channels mainly utilized yeast two-hybrid analysis, GST-fusion protein interactions, coimmunoprecipitations (co-IP), and immunolocalizations. Although the search for a PDZ-domain containing scaffolding protein did not lead to identification of an INAD-like protein, several other scaffolding proteins have been identified that interact with specific TRPC members. These include NHERF (TRPC4), Homer (TRPC1, TRPC3), junctaphilin (TRPC3), and RACK1 (TRPC3).4,8 Other proteins that have been noted to be associated with a number of TRPC channels include key Ca2+ signaling proteins such as plasma membrane Ca2+-ATPases (PMCA), G proteins, phospholipase C (PLC), sarco/endoplasmic reticulum Ca2+ ATPases (SERCA), and IP3 receptors (IP3Rs). Complicating our understanding of these protein–protein interactions is the detection of these interacting proteins residing in other organelles and structures within the cell, e.g., ER,9 cytoskeletal scaffolding structures,3 and subpopulations of mitochondria (peripheral) located near the plasma membrane.10 Thus, defining a physiological relevance for these interactions has been a challenge. Furthermore, it is also clear that the subcellular environment of the TRP proteins is dynamic. For example, there is an increase in the interaction of TRPC3 with IP3Rs and G proteins following stimulation of cells. This interaction appears to be required for plasma membrane localization of TRPC3 and its function. The complex also modulates IP3R function.11 Homer interactions with TRPC3, TRPC1, as well as IP3Rs are dynamically regulated by cell stimulation, which impacts on the channel function.7 Association of TRPC5 with the exocyst complex is involved in its trafficking.12 Partitioning of TRPC1 into lipid raft domains increases upon store depletion, where it colocalizes and associates with STIM1 following store depletion.13,14 Further, interaction of TRPC1 with caveolin-1 (Cav1) is required for plasma membrane localization of the channel. However, in response to stimulation, the channel dissociates from Cav1 and binds to STIM1.15 Clearly, in addition to the basic “resting” complex of proteins associated with TRPC channels, the changes triggered by activation also need to be elucidated.

Attempts to delineate the protein–protein interactions in TRPC channel complexes using such techniques as co-IP and immunolocalization have revealed an extensive list of proteins associated with TRPC channels.7 However, there are inherent limitations in these methods because these are primarily based on hypothesis-driven approaches. Hypothesis-driven approaches require prior insight to identify candidate proteins that are then screened for interactions with the target (TRPC) protein based on other observations, e.g., functional data. Although this might be a useful approach for confirming “proposed” interacting proteins, a “discovery-based” method can be expected to yield a much more comprehensive and unbiased survey of interacting proteins. New applications of existing technologies (mass spectrometry and separation sciences combined with high-throughput data processing) have allowed the possibility of discovery-driven approaches designed to provide an unbiased and complete spectrum of binding partners without the need for prior preliminary data. Discovery-driven approaches used before, e.g., yeast two-hybrid and GST-fusion protein interaction analysis, have their own critical drawbacks, i.e., only binary interactions are detected, there is a high false-positive rate, and the high-throughput data management methods, including genome and bioinformatics data, are not readily available.

In this article, we describe how a discovery-based proteomic approach can be employed to identify novel protein–protein interactions associated with TRPC channels. We will also discuss attempts made in our laboratory in this direction including the pitfalls that we encountered. Most of the previously reported studies with TRPC channels report partial or selective analysis of proteins. In particular, many studies have been done with heterologous expression systems that have their own major drawbacks. We will discuss these and also compare them with studies involving endogenous proteins from tissues. The studies we have carried out represent the only high-throughput proteomic analyses of endogenous TRPC family members. Our results confirmed many previously proposed protein interactions with TRPC1 and TRPC3, as well as identified new, previously unreported interactions.16 Because physiological responses involving these channels require changes in the molecular components associated with them, the method used for proteomic analysis should allow quantitative assessment of the relative levels of proteins in these complexes. In this report, we give an overview of new methodologies that are being applied to quantify the proteins in the signalplex in the resting state and also applications to quantitatively assess changes in the proteome associated with channel activation.

5.2. PROTEOMIC ANALYSIS OF TRPC-ASSOCIATED PROTEIN COMPLEX

5.2.1. Biochemical Considerations

5.2.1.1. Proteomic Analysis of Membrane Proteins

Unlike soluble proteins, special biochemical considerations are required for proteomic analysis of integral membrane proteins, owing to the fact that these proteins are hydrophobic and associated with a specific lipid environment. These considerations also apply to analysis of ion channel proteins such as TRPC channels, which are primarily associated with plasma membrane Ca2+ signaling mechanisms. The main points that need to be taken into account in the experimental strategy used for proteomic analysis of membrane proteins are as follows:

  • 1. Membrane proteins need to be solubilized using detergents in order to release them from the lipid environment. It is imperative that the conditions used preserve the functional integrity of the protein. For example, protein interactions, such as those that occur via hydrophobic domains, can be altered by the type of a detergent. Of course, care should be taken to avoid denaturation or aggregation of the protein because this will also adversely affect the interacting proteins. Typically, nonionic detergents such as octylglucoside, which has a high critical micelle concentration, are suitable for this procedure. Inclusion of lipids (similar to the endogenous milieu) as well as glycerol (dehydrating reagent) helps to maintain the integrity of protein complexes.
  • 2. The availability of trypsin cleavage sites is usually low owing to the hydrophobic nature of integral membrane proteins. This is a hindrance for proteomic analyses because of the limitations of MS/MS analysis, i.e., large peptides are difficult to fragment using conventional collision-induced dissociation (CID). Peptides with an upper limit of 25–30 amino acid residues in length typically yield high-quality MS/MS spectra, although exceptions to this length guideline do indeed exist. Additionally, the ability to extract large peptide fragments from the polyacylamide gel (in-gel digests) is limited.17 Together, this results in a low yield of peptides that has a significant impact on the detection of proteins present in low abundance, such as plasma membrane ion channels.
  • 3. Ideally, an in-solution digest of proteins or protein complexes would yield the most thorough results in a shotgun proteomic analysis, as losses associated with sample handling and poor recovery from the gel matrix are minimized. However, many detergents, while efficient in solubilizing the target protein, are unsuitable for in-solution digest owing to incompatibility issues in the downstream HPLC and electrospray ionization-based mass spectroscopy.
  • 4. Integral membrane proteins of the plasma membrane are usually present in low abundance. This drawback requires alternate enrichment strategies18 to achieve a sufficient material to analyze. A consequence of this scaling up, unfortunately, can also result in a proportional increase in nonspecifically bound proteins. An alternative approach would be to carry out an initial enrichment step prior to purification, e.g., one can start with a purified plasma membrane preparation rather than whole tissue or crude membrane fractions. However, availability of tissue could be a serious limitation to this. Despite such preliminary enrichment steps, control experiments are required to sort out putative specific and nonspecific interactions with the target protein. Examples of appropriate control experiments for immunoaffinity-based isolations would be preclearing the lysate using buffer conditioned solid support (e.g., Protein A derivatized magnetic beads) and/or preclearing using an immobilized nonspecific immunoglobulin to minimize the detection of nonspecific interactions. Proteins identified as potential interactors in discovery-based assays still require further assessment using orthogonal biochemical and functional assays (see below).

The following sections will discuss methodologies that have been applied to successfully perform shotgun proteomic analysis on integral membrane proteins and that have been used for TRPC channels. Additionally, new procedures that can enhance proteomic discoveries will be introduced (e.g., new gel-free platforms using organic acid-cleavable detergents that have been developed to solubilize membranes18). These detergents perform their traditional role up to the HPLC step but can be removed by acid-induced cleavage to allow in-solution digest and analysis of the extracted target protein.

5.2.1.2. Use of Immunoprecipitation: Limitations and Advantages

Once solubilized from its membrane environment, the target protein needs to be isolated in sufficient quantities prior to proteomic analysis. One approach used for this is affinity purification. Although it is well established that with proper optimization detergents do not interfere with antibody-antigen interactions,19 the affinity and specificity that a particular antibody has toward its antigen vary considerably. Another major drawback is the commercial availability of reliable antibodies. Most commercially available antibodies are developed for use as diagnostic reagents, e.g., for Western blotting or immunohistochemical staining. These antibodies must be evaluated for their properties as preparative affinity reagents. Another critical concern is that the immunoprecipitation (IP) reagents (e.g., beads and the Fc portion of the antibody) provide numerous sites for nonspecific binding to occur. Thus, it is important to design the IP protocol to minimize these problems, as well as to control and account for them. Further, because the concentration of the antibody is relatively high compared to the protein sample to be tested, one can predict a high level of contamination from IgG peptides. One modification that has been successfully applied is covalent conjugation of the antibody to the matrix. However, one must demonstrate that the antibody–matrix cross-linking has not impaired the ability of the antibody to capture and release the appropriate bait protein. Thus, it is possible to selectively resolve biologically relevant protein–protein interactions from experimentally introduced contaminants.19

If an ideal antibody targeted toward the native protein is not available, a heterologously expressed epitope-tagged protein can be used. In this case, the antibody directed against the tag is used for affinity purification of the protein complex. However, if the tagged bait protein expression level is too high, the assembly of the protein complex may not occur in a physiological/stoichiometric manner.20 Overexpression systems also lead to “backup” of membrane trafficking pathways, resulting in a large amount of protein in other intracellular compartments, including ER and Golgi, or in extreme cases trapped in inclusion bodies. In the latter case, the protein is often resistant to solubilization. Isolation of proteins by IP would include proteins associated with the target localized in ER and Golgi, and these are generally chaperones, heat shock, and other trafficking-related proteins that might not be relevant to the mature functional protein that ultimately assembles in the plasma membrane.

Finally, another important consideration while implementing IP for isolation of a membrane protein complex is the stringency of the wash. This wash is performed after immobilization of the protein on the beads, prior to elution. The purpose of this procedure is to wash off no-specifically/loosely attached proteins so that a reproducible core group of proteins that are tightly associated with the target protein can be retained. Specificity of this protein complex can be determined by comparing samples obtained using control IgG with those obtained using antibody targeted to the target protein. Cells in which the target protein has been knocked down are also a very useful control, although this would largely be determined by the efficiency of the knockdown. In all these cases, the wash conditions critically determine the outcome. If the wash is too mild (i.e., lacking in ionic strength or detergent), too many nonspecific proteins will remain associated with the target protein. Conversely, if the wash buffer is too stringent (e.g., a RIPA buffer with a high SDS concentration), it can result in disruption of transient and/or low affinity (but nonetheless important) interactions between the target protein and putative interacting proteins.19 Thus, a considerable effort is required to ascertain that the IP and wash conditions are yielding a consistent and reproducible array of proteins. Polyacrylamide gel electrophoresis (PAGE) and silver staining can be used to examine the panel of proteins in the sample at every step.

5.2.1.3. Biochemical Confirmation of Proteomic Data

The experimental approaches discussed above are aimed toward controlling critical variables that can impact immunopurification of membrane proteins for proteomic analysis. Despite these efforts, it is important to remember that the proteins detected in the final sample after proteomic analysis can only be defined as “putative interacting” proteins until they are confirmed using biochemical, molecular, or functional techniques. This confirmation is especially important when identification of a protein is based on a single peptide. Although, in principle, a single unique peptide is sufficient to identify a protein,21 the reliability of the identification increases exponentially with the number of peptides used to identify the protein. Some of the more commonly used techniques to confirm the identifications detected by MS/MS include Western blotting, direct yeast two-hybrid complement and interactions, immunolocalization/colocalization, cross-linking, and multiepitope-ligand cartography (for a review, see Ref. 19). The putative interacting protein can also be modulated in cells (overexpression or knockdown), and the effect on the target protein function can be assessed. Together, these methods provide a comprehensive assessment of the proteins that associate with the target protein.

5.3. OVERVIEW OF TRPC PROTEOMIC ANALYSIS

5.3.1. Previous Proteomic Studies of TRPC Proteins

Despite the extensive amount of work on deducing the protein–protein interactions involving TRPC proteins using traditional methods such as yeast two-hybrid analysis and co-IP, there is very little published work regarding proteomic analysis of TRPC channels. In one of the few publications, TRPC5 and TRPC6 immunocomplexes were analyzed by spot-picking after SDS-PAGE.22 After sequencing using MS/MS and subsequent identification, a partial list of interacting proteins was obtained from this approach, including several cytoskeletal proteins such as spectrin, myosin, actin, drebrin, tubulin, and neurabin. Additionally, endocytic vesicle-associated proteins such as clathrin, adapter-related protein complex AP-2, and a dynamin-l-like protein were identified, as well as the plasmalemmal Na+/K+-ATPase. Although several important observations were achieved in this study, two important shortcomings were evident. (1) The analysis was not based on a comprehensive approach since only select bands were analyzed. (2) More importantly, no control experiments were done to eliminate nonspecific binding. In many cases, quantitative comparisons between a common control and TRPC-IP peptide are needed to distinguish between specific and nonspecific interactions.

Another study used proteomic techniques to map autophosphorylation sites on TRPM6 and TRPM7.23 This paper, while somewhat unrelated to the determination of protein–protein interactions in TRPC channels, nevertheless shows how proteomic techniques can be applied to better understand protein modifications (e.g., posttranslation modifications) that might affect function.

5.3.2. Choice of Proteomic Techniques in the Analysis of TRPC Proteins

We will discuss two principal methodologies that can be utilized for initial discovery-based proteomic analysis. The first, geLC–MS/MS, utilizes one-dimensional (1D) SDS-PAGE to separate the protein components of the TRPC channel complex isolated by affinity purification (Figure 5.1a). This 1D PAGE is followed by sectioning (~1 mm each, see Figure 5.1b) the entire length of each gel lane as opposed to selected bands. Because the whole lane is sampled, visualization by protein staining is optional, although routinely employed as documentation. Individual gel bands are subjected to in-gel enzymatic proteolysis and subsequent extraction of the peptide hydrolysate, followed by tandem mass spectral analysis of the peptides and submission of the data to a database search engine in order to determine the identity of the peptides and infer the identity of their precursor proteins. Some of the benefits of the geLC–MS/MS approach are the facile ability of working with detergent-containing samples and the significant resolving power (theoretical plates of separation) of SDS-PAGE. Importantly, all peptides produced from enzymatic digestion of proteins in a given gel slice remain in a single fraction, unless the underlying precursor protein spans a slice junction. This fact substantially simplifies inference of the correct precursor protein during subsequent bioinformatic analysis of the data because the approximate molecular weight of the precursor can be determined from molecular weight standards. Limitations of the geLC–MS/MS approach are principally related to the partial and variable recovery of peptides in the extraction step.

FIGURE 5.1. Overview of the biochemical strategy of isolating “specific TRPC1 or TRPC3 binding partners” from the rat brain.

FIGURE 5.1

Overview of the biochemical strategy of isolating “specific TRPC1 or TRPC3 binding partners” from the rat brain. (a) After isolating rat brain crude membranes and solubilizing the membranes with octylglucoside, immunoprecipitates were (more...)

The second approach, termed two-dimensional liquid chromatography–mass spectrometry (LC/LC–MS/MS), begins with solution phase digestion of the affinity purified mixture of proteins in aggregate, followed by peptide fractionation using, for example, strong cation exchange (SCX) and reversed-phase (RP) liquid chromatographies coupled with tandem mass spectral analysis. Alternate chemistries can and have been employed as variations on the multidimensional LC theme. Two-dimensional LC can be performed while directly coupled to the mass spectrometer,24,25 although there are advantages to decoupling the chromatographic steps. Performing offline SCX allows the practitioner to have the ability to increase sample loading capacity, employ organic modifiers in the SCX solvents, and take advantage of increased chromatographic resolution by employing a linear elution gradient as compared to step elution mandated by online 2D LC–MS/MS.

5.3.3. Instrumentation and Details of MS/MS Analysis

Electrospray ionization (ESI) is an ideal interface to couple LC systems directly online to modern tandem mass spectrometers, e.g., ion traps, ion trap-Orbitrap, and triple quadrupoles (QqQ). As the name implies, nanoflow LC utilizes submicroliter-per-minute flow rates; 200–400 nL min−1 is typical. At the time the TRPC experiments were performed, our laboratory utilized an automated 1D HPLC system that was constructed from Shimadzu LC-VP Series components (Kyoto, Japan), which allowed the use of both an autosampler and an HPLC at submicroliter-per-minute flow rates. For the uninitiated, it is difficult to accommodate the system volume of an autosampler with nanoflow rates used with 300-μm ID HPLC columns. The optimal flow rate for autosampler use greatly exceeds that required for microbore HPLC. Conversely, at flow rates optimal for microbore HPLC, autosampler loading times would be prohibitively long. Our system utilized three LC-10ADVP pumps with microflow control kits, an SIL-10ADVP automatic injector, an SCL-10AVP Controller, and a Cheminert CN2 nanovolume switching valve (Valco Instruments Company Inc., Houston, TX, USA).

One of the pumps (Pump C) delivered Buffer A at 40 μL min−1 to the autosampler. Samples were selected and loaded onto a peptide CapTrap (0.5 mm × 2 mm, PLRP-S 5μ 100Å; Michrom BioResource Inc., Auburn, CA, USA), which was connected to the nanovolume switching valve. This valve was used as a solvent selector for trapping, desalting, and loading samples onto the reverse-phase column. The other two LC-10ADVP pumps (A and B) were operated in conventional mode developing a linear gradient (10% B [3 min], a linear gradient of 10–60% B [40 min], 60–80% B [10 min], 80% B [2 min]) at 10 μL min−1. The LC eluant was connected to a micro splitter valve (Upchurch Scientific Inc., Oak Harbor, WA, USA) to deliver an operating flow rate of 400 nL min−1 to the switching valve. After extensive washing with Buffer A, the LC gradient was brought in-line, and the peptide samples were eluted sequentially from the Peptide CapTrap onto a PicoTip fused silica HPLC column (BetaBasic C18, 0.075 × 100 mm, 360-μm OD 15-μm spray tip) and into a ThermoFinnigan LCQ Classic ESI-ion trap mass spectrometer (San Jose, CA, USA).

Mobile phase buffers were RP-A, water/acetonitrile/formic acid = 95.0/5/0.1 (v/v/v); RP-B, water/acetonitrile/formic acid = 20/80/0.1 (v/v/v); and RP-C, water/formic acid = 100/0.1 (v/v). The LCQ was operated in positive ion mode with a dynamic exclusion set to repeat count = 1, exclusion duration = 0.5 min, exclusion mass width = 3 amu. Spectra were acquired in a data-dependent manner with the top five most intense ions in the precursor scan selected for CID. Normalized collision energy was 35.0, and the minimum precursor ion intensity required was 2× baseline at equilibrium.

For quantitative measurements, our laboratory utilizes a true, splitless, nanoflow HPLC system (Eksigent Technologies, Inc., Dublin, CA, USA). The HPLC columns, buffers, and gradient are the same as described above. In order to insure ion beam stability under all LC conditions, we use an Advion Nanomate (Ithaca, NY, USA) ESI ion source that is mounted on an LTQ-Orbitrap mass spectrometer (ThermoFisher Scientific, San Jose, CA, USA). This instrument is operated in data-dependent mode. Survey MS scans are acquired in the Orbitrap in profile mode with the resolution set to a value of 60,000. Up to five of the most intense ions per scan are fragmented and analyzed in the linear ion trap.

5.3.4. Software Requirements for MS/MS Proteomic Analysis

The Center for Information Technology (CIT) at the National Institutes of Health maintains a computer cluster hosting the Mascot Search Engine (Matrix Science, Inc., Boston, MA, USA). The NIH Mascot cluster consists of four nodes: one head node and three computational nodes. Each node is a dual-core dual-socket (four cores total) 2.6-GHz Opteron with 8 GB of RAM, running 64-bit Centos 5.4. The nodes are connected by 1-Gb/s ethernet to each other and to the fileserver on which the data reside. File storage is on a Netapp FAS 960 file server. CIT performs regularly scheduled updates of the genomic sequence libraries used in MS-based protein identification. Individual users remotely connect to the cluster using the Mascot Daemon. In addition to defining the configuration of search parameters, the Daemon automates conversion of the raw MS/MS data into a concatenated list of mass versus intensity pairs for each data file (.mgf), transmission of the .mgf files to the search engine, and management of the search result files.

The following parameters were used to search the TRPC data described herein: database = SwissProt (derived from UniProt); taxonomy = All; enzyme = trypsin; maximum number of missed cleavages = 2; peptide charge states = +1 - +4; mass measurement = monoisotopic; low resolution ms: peptide mass tolerance = 1.2 Da; fragment ion mass tolerance = 0.6 Da; high resolution ms: peptide mass tolerance = 50 ppm, fragment ion mass tolerance = 0.6 Da; fixed modification = carbamidomethyl-Cysteine; variable modification = methionine oxidation (Met). Subsequent searches were performed to detect common posttranslational modifications (e.g., protein N-acetylation) and chemical artifacts (e.g., pyro-Glutamic acid formation and pyro-carbamidomethyl-Cysteine from N-terminal Gln and camCys, respectively). In these searches, taxonomy was restricted to Rattus. While a detailed discussion of the scoring algorithm used by the Mascot search engine is beyond the scope of this paper, it is important to point out that because the number of rat proteins is very small compared to the total number of protein sequences in the SwissProt database, the rate of false-positive identifications will be greatly increased. Practitioners are well advised to check the more general mammalian taxonomy search results to compare with Rattus taxonomy and to manually inspect new identifications made as a result of subsequent searches on a subset of total database records.

5.4. VALIDATION OF PROTEOMIC DATA

5.4.1. Experimental Design to Compensate for Nonspecific Interactions

To illustrate didactically how various experimental parameters affect a membrane protein proteomic study, a brief outline of our initial work on the analyses of TRPC1 and TRPC3 proteomes follows. In our initial attempts to increase the copy number of TRPCs in our system and to simplify the extraction step, we transfected HEK293 cells with FLAG-tagged TRPC1 or TRPC3 and used untransfected HEK293 cells as a control. After obtaining a crude membrane fraction, octylglucoside solubilizates were immunoprecipitated with anti-FLAG-linked agarose beads and the IP fraction (of FLAG-TRPC3 eluted with free FLAG peptide) tested for Ca2+ permeability in proteoliposomes. After confirming an active FLAG-TRPC3 complex,16 we proceeded to scale up the IP, and to maximize the chances of retaining as many proteins as possible, we used a low stringency IP wash buffer (TBS which consisted of 50 mM Tris HCl, 150 mM NaCl, pH 7.4) before eluting the IP fraction. Subsequent MS/MS analysis of the IP fractions of both FLAG-TRPC1 and FLAG-TRPC3 revealed an extremely large set of proteins for both (>5000). To obtain a more manageable set of proteins, we repeated the experiment, but instead of washing with mild TBS, we washed the IP fraction with a more stringent IP wash buffer that contained (among other things) 0.5 M NaCl and 0.5% NP-40 (for a complete formulation, see Ref. 16). MS/MS of these IP fractions resulted in a more manageable number of proteins (944 for TRPC1 and 256 for TRPC3). However, it became evident that a sizeable number of the identified proteins were a result of overexpression (Table 5.1). A relatively high percentage of proteins associated with protein synthesis and trafficking, including those in the ER and Golgi, were detected. As discussed above, this is a common problem with overexpression systems because the immunopurification pulls down all proteins, including those in intracellular compartments.20 Fully functional TRPC channels are normally present in a complex where each component is assembled in a specific stoichiometric ratio with available natural binding partners. Overabundance of newly synthesized or partially assembled TRPC proteins will result in exposure to chaperone proteins (for eventual destruction) or their congregation in sorting vesicles waiting for natural binding partners to become available. In either case, these other protein interactions will be higher relative to the interactions of the mature protein within its functional milieu, resulting in difficulty in detecting the latter interactions. On the basis of these observations and considerations, we next pursued a more realistic proteomic assessment using native TRPC channels obtained from natural tissue sources. In native systems, TRPC channels have a low turnover. Thus, we can expect most of the protein to be in its mature state in the plasma membrane.

TABLE 5.1

TABLE 5.1

Proteomic Results using overexpression of FLAG-tagged TRpcs

We used rat brain for our initial studies with native TRPC channels because this is a well-characterized tissue source that has been previously shown to substantially express TRPC1 and TRPC3.26 The procedure we utilized was to homogenize quickfrozen rat brains followed by differential centrifugation to obtain a crude membrane fraction. The membrane fraction was solubilized using a buffer containing octylglucoside/lipid/glycerol mixture, and the solubilized fraction was used to immunoprecipitate either endogenous TRPC1 or TRPC3 by their respective antibodies. To account for nonspecific interactions, we precleared the solubilizates with Protein A Sepharose CL-4B beads and also performed a control IP using rabbit IgG (because the native anti-TRPC antibodies are derived from rabbit). The flowchart of our final experimental design is shown in Figures 5.1a and 5.1b; this figure depicts a multistep approach that included IP, SDS-PAGE followed by in-gel digest, extraction, separation, and analysis by MS/MS. This approach required considerable scale-up of the protein preparation because of the lower amount of TRPC proteins in a native system. We determined that a 10-fold higher amount of crude membrane protein from brain (as compared to membranes from cells expressing FLAG-tagged protein) was needed to obtain an IP sample that was concentrated enough for detection of proteins after in-gel digest and extraction. Despite these additional efforts that were required, the approach was successful, and significantly more functionally relevant proteins were identified in the TRPC sample with a lower percentage of irrelevant interactions (see Table 5.2 and compare with Table 5.1).

TABLE 5.2

TABLE 5.2

Functional Overlap of the TRPC1 and TRPC3 Proteomes

5.4.2. Criteria for the Determination of Valid Identification

5.4.2.1. Importance of Multipeptide Identification

Our evaluation strategy to assign a candidate protein as a specific TPPC-interacting protein is based on matching sequenced peptide fragments from the MS/MS analysis to known proteins via database searching. This process, termed protein inference, has been reviewed extensively, and a generalized list of guidelines to help eliminate false positives has evolved (for a review, see Refs. 27 and 28). In our analysis, several layers of filtering mechanisms were used. For example, known contaminating proteins (e.g., IgG’s and keratins) were initially filtered out because these proteins are obvious artifacts. After initial filtering of known contaminants, candidate proteins then had to meet the following criteria to be putatively identified as a relevant protein:

  • 1. The candidate protein has two or more sibling peptides (multihit), and it is present only in the TRPC-IPs.
  • 2. Multihit proteins appearing in both the TRPC-IPs and control IgG-IP (which also had one or more shared peptides that could be used for quantitative assessment) must demonstrate at least a twofold quantitative increase between TRPC-IP and control-IP.
  • 3. Proteins identified by a single peptide could be assigned as a TRPC-interacting protein only if additional verification (e.g., Western blotting or previous functional studies) could confirm the relevance or presence of the protein in TRPC function or regulation.

In all cases, the existence of at least one unique (distinct) peptide (that could only be present in the candidate’s protein sequence) was required. To analyze a family of proteins, cluster analysis of shared peptides was performed using MassSieve29 in order to determine if unique sequences were present in individual family members. Only family members with at least one unique sequence were included into any of the “specific TRPC interacting protein” category (the selection strategy is detailed in Figure 5.2).

FIGURE 5.2. Flowchart of considerations used to assign a candidate protein a “specific TRPC1 or TRPC3 binding partner.

FIGURE 5.2

Flowchart of considerations used to assign a candidate protein a “specific TRPC1 or TRPC3 binding partner.” After filtering of obvious artifacts such as keratins and immune proteins, “Multiple Peptide Hits” proteins were (more...)

5.4.2.2. Quantitative Analysis of Shared Peptides

If a protein was identified in both the control and TRPC runs and a common peptide sequence was used to identify both, quantitative data could be used to differentiate between specific and nonspecific binding. The stability of the LC–MS/MS ion current makes it useful to retrieve label-free quantitative information such as retention time, peak intensity, and integrated area from the MS1 data, which affords the ability to compare the same putative precursor protein between experiments. For quantification, there are multiple software tools with the capability to integrate MS1 data corresponding to identified peptides from MS2 scans. We used an in-house software tool, DBParser, to extract retention time and peak intensity from MS1 raw data based on the precursor mass identified by Mascot. The peak area and the number of scans were calculated from the selected ion chromatogram. Mass tolerance for extraction of the ion current is user-selected and instrument-dependent. In the case of the LCQ classic employed in these studies, a mass tolerance of ±0.6 Da was used. In addition to peptide sequence, m/z, Ions Score, Homology Score, and Identity Score from Mascot search, quantification reports include retention time, peak intensity, area, and number of scans. Quantitative differences of the same protein identified in both the control and TRPC samples were analyzed by comparing the sum intensity ratio (TRPC/control) of a unique peptide present in both runs (when possible). If a quantitative ratio of more than twofold resulted, we interpreted this as a quantitative difference and not resulting from a mere nonspecific binding event. When possible, Western blot analysis of equal amounts of control-IP and TRPC3-IP fractions was analyzed to confirm quantitative differences.

5.4.3. TRPC3 Proteome

In our initial attempt to elucidate the functional components and regulation of TRPC-associated Ca2+ influx, we utilized a solubilization–reconstitution approach that we had previously used to assess Ca2+ permeability pathways (as gauged by 45Ca2+ uptake assays) in plasma membranes from salivary glands.30 FLAG-TRPC3 immunocomplexes were released using a free FLAG peptide and subsequently reconstituted into proteoliposomes and function gauged by 45Ca2+ uptake assays. Immunopurified FLAG-TRPC3 displayed 1-oleoyl-2-acetyl-sn-glycerol (OAG) -stimulated Ca2+ influx following reconstitution in a proteoliposomal system. Western blots verified the presence of FLAG-TRPC3 in the reconstituted vesicles. Proteoliposomes containing FLAG-TRPC3 displayed an approximately twofold higher 45Ca2+ uptake compared to proteoliposomes prepared from immunoprecipitates obtained from control nontransfected HEK293 cells.16 Further, OAG, but not the inactive analog 1,3-dioctanoyl-sn-glycerol, increased 45Ca2+ uptake in the FLAG-TRPC3 proteoliposomes. Interestingly, a TRPC3 mutant with mutation in its pore region displayed reduced 45Ca2+ uptake into proteoliposomes. These data demonstrate that our solubilization and IP conditions yield a functional TRPC3 protein. Despite the preservation of function, proteomic studies of FLAG-TRPC3 revealed the presence of a considerable number of artificially induced interactions in the immuno-purified FLAG-TRPC3 complex (see table above). Thus, a native anti-TRPC3 antibody was used to immunoprecipitate TRPC3 from solubilized rat brain crude membranes using the same conditions we have previously used that allowed retention of function. Proteins in the TRPC3 and control (using rabbit IgG) immunoprecipitates were analyzed, and analysis of the sequence data revealed the presence of 76 specific TRPC3-associated proteins (as determined by the selection criteria described above). Figure 5.3 shows the functional distribution of the proteins contained in the TRPC3 proteome.

FIGURE 5.3. Functional classifications of the (a) TRPC1 and (b) TRPC3 proteomes constructed from proteins deemed to be specific TRPC1 or TRPC3 binding partners as described in Figure 5.

FIGURE 5.3

Functional classifications of the (a) TRPC1 and (b) TRPC3 proteomes constructed from proteins deemed to be specific TRPC1 or TRPC3 binding partners as described in Figure 5.2. (From Lockwich, T., J. et al., Journal of Proteome Research 7(3), 979-989, (more...)

5.4.4. TRPC1 Proteome

As described above, we have also used a native anti-TRPC1 antibody (from rabbit) to immunoprecipitate TRPC1 from solubilized rat brain crude membranes using the same experimental approach that preserved the TRPC3 function. Proteins in the TRPC1 and control (using rabbit IgG) immunoprecipitates were separated by SDS-PAGE and subsequent gel slices trypsinized as outlined above. After the extraction of the peptides, the peptide fragments were separated and sequenced using HPLC and MS/MS techniques, respectively. Analysis of the sequence data revealed the presence of 59 specific TRPC1-associated proteins that can be initially included into the TRPC1 proteome, based on multiple peptide identification (and not present in the control). Additionally, eight proteins present in both the TRPC1 and control immunoprecipitates were determined to have quantitative differences between the TRPC1-IP fraction and the control (IgG) nonspecific binding fraction. After biological confirmation, eight additional proteins initially identified by one peptide in the TRPC1-IP (but not in the control IP) are also included in the TRPC1 proteome (making a total of 75 validated proteins associated with TRPC1). Figure 5.3 shows the functional distribution of the proteins contained in the TRPC1 proteome.

5.4.5. Functional Overlap of the TRPC1 and TRPC3 Proteomes

Interestingly, we observed many of the same proteins in both the TRPC1 and TRPC3 proteomes (approximately 50%). Moreover, when we analyzed the functional similarities among common proteins, a very asymmetrical distribution was evident (Table 5.2). The functional comparison between the two reveals a greater overlap in certain functional groups, e.g., protein involved in endocytosis. Hybrid TRPC1/TRPC3 channels have been proposed to exist,8 and, consistent with this observation, TRPC3 co-immunoprecipitates with TRPC1-IP from rat brain extracts (unpublished observations). It is tempting therefore to speculate that, in certain functional areas, TRPC1 and TRPC3 coexist as heteromultimers and thus share the same protein partners. In contrast, the functional areas where little overlap in binding partners is observed could indicate independent roles for TRPC1 and TRPC3. Further studies will be needed to elucidate the complexes where TRPC1 and TRPC3 interact independently as opposed to complexes where both are present. To guide us in this attempt, one approach is to examine the distribution of the channels in thin sections of the rat brain. Mapping these regions together with localization of specific proteins might provide clues as to the relevance of the findings obtained from TRPC1 and TRPC3 proteomic analysis.

5.5. QUANTITATIVE PROTEOMIC TECHNIQUES

5.5.1. SILAC (Stable Isotope Labeling by Amino Acids in Cell Culture)

The plasma membrane is a dynamic system, and the composition of membrane proteins can dramatically change after contact with other cells, signal molecules, or simply the maturation of a cell. Some proteins are only temporarily exposed to the cell surface, but, nevertheless, these transient changes are important to understanding how signals are relayed to the interior of the cell to elicit responses. SILAC is a technique that exploits the fact that mammalian cells cannot synthesize a number of amino acids.31 Isotopically labeled analogs of these amino acids can be synthesized and are commercially available. Either naturally occurring or “heavy” amino acids (added to amino acid deficient cell culture media) can then be incorporated into all proteins as they are synthesized. Because there is no chemical difference between the light and heavy amino acids, two groups of cells can be grown that should behave exactly alike. To ensure complete incorporation of the heavy amino acid into the experimental cell culture group, at least six passages of cell culture are allowed to proceed before testing for complete incorporation. The experimental cell population can then be treated in a specific way, such as activation by thapsigargin or specific agonists. Equal amounts of protein from both sets of cell culture conditions can then be mixed and fractionated, based on binding to a target protein (in this case, IP). Changes in the levels of binding partners (between the control and treated cell cultures) can be quantitatively analyzed at the level of the peptide mass spectrum or peptide fragment mass spectrum. This type of analysis can determine if the amount of a binding partner decreases, remains the same, or increases as a result of the specific treatment.32 Such an approach can be used for detecting dynamic changes in the TRPC complex. For example, control cells can be labeled with naturally occurring lysine and arginine, while stimulated (or otherwise treated) cells can be labeled with heavy amino acids 13C6-lysine and 13C6,15N4-arginine. Solubilization and immunopurification of the protein of interest followed by MS analysis should reveal important quantitative changes on the interacting partners. Such studies have not yet been carried out with TRPC channels and should provide invaluable data as to the regulatory components of these channels.

5.5.2. Quick LC–MS (Quantitative IP Combined with Knockout)

One of the most difficult aspects of proteomic analysis when methods such as IP are used is the presence of nonspecifically bound proteins. Despite experimental designs to minimize these interactions (e.g., preclearing of the extract and use of control antibodies), interference invariably results from proteins that nonspecifically bind to any of the components used to isolate the target protein. The prevailing method to eliminate nonspecific interactions is to compare proteins identified in a specific extraction with proteins extracted from a lysate that does not contain the target protein. Clearly, this is not possible in cases that target ubiquitously expressed proteins. Therefore, our previous methods to correct for nonspecific binding can only be partially effective. Recently, a new strategy has been developed to better discriminate specifically and nonspecifically bound members of an extracted protein complex. This new strategy, named QUICK,33 combines SILAC with RNA interference (RNAi) to knock down target protein expression in one of the cell cultures. After IP of equal protein samples and subsequent analysis by LC–MS/MS, doublets should be observed, which appear either in equal amounts (nonspecific) or in much greater amount in the non-RNAi treated culture (indicative of a specific interaction). One obvious drawback with QUICK is that this technique cannot be used with native tissue because it requires cell culture. These and other limitations due to the nonspecific interactions need to be resolved in future studies.

5.6. CONCLUSION

The recent application of high-throughput informational processing combined with more established techniques such as HPLC and mass spectrometry has allowed the emergence of proteomics, a powerful new tool in determining protein–protein interactions. Previously undreamed of capabilities that allow the complete identification of a complex mixture of proteins are now possible. With this new capability come important limitations, however. The constant reassessment of data to confirm validity and specificity is a necessary caveat that must be respected to allow the field of proteomics to evolve into a mainstream biochemical technique. Through our experiences analyzing the proteomes of several components of the Ca2+ entry pathway, we have identified several important areas that have drastic effects on the quality of the results. The first critical area is experimental design. Although the use of epitope-tagged proteins has revolutionized the ease of amplifying and extracting a target protein, it does promote interactions that are not physiologically relevant. In our estimation, it is advisable to scale up a natural system rather than to attempt a proteomic analysis on an overexpressed epitope-tagged system. Another important consideration is prefractionation. As discussed earlier, membrane proteins are usually present in low copy numbers, and therefore, every attempt should be made to concentrate the target protein before the actual proteomic analysis. For example, in our studies on the TRPC1 and TRPC3 proteomes, a crude membrane fraction from the rat brain homogenate was secured before proceeding to solubilization. Other variations of this fractionation strategy (i.e., isolating lipid raft domains before analysis) have allowed others a suitable starting point for proteomic analyses of membrane proteins.34

For shotgun proteomic analyses involving immunoprecipitation to isolate the target protein and associated proteins, important consideration should be given to the stringency of the washing buffer used prior to elution. As reported earlier, a wash buffer too low in ionic strength and detergent concentration results in an overly complex immunocomplex, whereas an overly harsh wash buffer can strip away important interactions. The selection of an appropriate wash buffer must be determined experimentally. In our opinion, a good starting point would be a functionally active complex that includes the minimal number of components. The ability to control antibody leaching into the system should also be an important consideration. Incorporation of covalently conjugated antibody–matrix systems in our proteomic projects has resulted in improved detection of natural protein interactions with the target protein. Other new and developing methods, as well as improved bioinformatics tools, should provide greater ease and specificity for carrying out successful proteomic methodology.

We hope that this review of our proteomic analysis of components of the Ca2+ entry pathway can help others to contemplate a proteomic approach to other membrane proteins. Although a proteomic analysis consists of many steps (complete with their inherent pitfalls), we feel that the discovery-based results obtained from such an approach are more than an ample reward for the risk and effort needed to obtain them.

REFERENCES

1.
Birnbaumer L. The TRPC class of ion channels: a critical review of their roles in slow, sustained increases in intracellular Ca2+ concentrations. Annual Review of Pharmacology and Toxicology. 2009;49:395. [PubMed: 19281310]
2.
Ramsey I.S, Delling M, Clapham D.E. An introduction to TRP channels. Annual Review of Physiology. 2006;68:619. [PubMed: 16460286]
3.
Montell C. TRP channels in Drosophila photoreceptor cells. Journal of Physiology. 2005;567:45. [PMC free article: PMC1474165] [PubMed: 15961416]
4.
Ambudkar I.S. Trafficking of TRP channels: determinants of channel function. Handbook of Experimental Pharmacology. 2007;179:541. [PubMed: 17217078]
5.
Spassova M.A, Hewavitharana T, Xu W, Soboloff J, Gill D.L. A common mechanism underlies stretch activation and receptor activation of TRPC6 channels. Proceedings of the National Academy of Science of the United States of America. 2006;103:16–586. [PMC free article: PMC1637625] [PubMed: 17056714]
6.
Ambudkar I.S, Ong H.L, Liu X, Bandyopadhyay B.C, Cheng K.T. TRPC1: the link between functionally distinct store-operated calcium channels. Cell Calcium. 2007;42:213. [PubMed: 17350680]
7.
Kiselyov K, Kim J.Y, Zeng W, Muallem S. Protein-protein interaction and function TRPC channels. Pflügers Archiv. 2005;451:116. [PubMed: 16044307]
8.
Ambudkar I.S, Ong H.L. Organization and function of TRPC channelosomes. Pflügers Archiv. 2007;455:187. [PubMed: 17486362]
9.
Delmas P, Wanaverbecq N, Abogadie F.C, Mistry M, Brown D.A. Signaling microdomains define the specificity of receptor-mediated InsP3 pathways in neurons. Neuron. 2002;34:209. [PubMed: 11970863]
10.
Park M.K, Ashby M.C, Erdemli G, Peterson O.H, Tepikin A.V. Perinuclear, perigranular and sub-plasmalemmal mitochondria have distinct functions in the regulation of cellular calcium transport. EMBO Journal. 2001;20:1863. [PMC free article: PMC125431] [PubMed: 11296220]
11.
Singh B.B, Lockwich T.P, Bandyopadhyay B.C. et al. VAMP2-dependent exocytosis regulates plasma membrane insertion of TRPC3 channels and contributes to agonist-stimulated Ca2+ influx. Molecular Cell. 2004;15(4):635. [PubMed: 15327778]
12.
Bezzerides V.J, Ramsey I.S, Kotecha S, Greka A, Clapham D.E. Rapid vesicular translocation and insertion of TRP channels. Nature Cell Biology. 2004;6:709. [PubMed: 15258588]
13.
Lockwich T.P, Lui X, Singh B.B, Jadlowiec J, Weiland S, Ambudkar I.S. Assembly of Trp1 in a signaling complex associated with caveolin-scaffolding lipid raft domains. Journal of Biological Chemistry. 2000;275:11–934. [PubMed: 10766822]
14.
Lockwich T, Singh B.B, Liu X, Ambudkar I.S. Stabilization of cortical actin induces internalization of transient receptor potential 3 (Trp3)-associated caveolar Ca2+ signaling complex and loss of Ca2+ influx without disruption of Trp3-inositol triphosphate receptor association. Journal of Biological Chemistry. 2001;276:42–401. [PubMed: 11524429]
15.
Pani B, Ong H.L, Brazer S.C, Liu X, Rauser K, Singh B.B, Ambudkar I.S. Activation of TRPC1 by STIM1 in ER-PM microdomains involves release of the channel from its scaffold caveolin-1. Proceedings of the National Academy of Science of the United States of America. 2009;106:20–087. [PMC free article: PMC2785296] [PubMed: 19897728]
16.
Lockwich T, Pant J, Makusky A. et al. Analysis of TRPC3-interacting proteins by Tandem mass spectrometry. Journal of Proteome Research. 2007;7:979. [PubMed: 18205297]
17.
Tan S, Tan H.T, Chung C.M. Membrane proteins and membrane proteomics. Proteomics. 2008;8:3924. [PubMed: 18763712]
18.
Lu B, McClatchy D.B, Kim J.Y, Yates J.R III. Strategies for shotgun identification of integral membrane proteins by tandem mass spectrometry. Proteomics. 2008;8:3947. [PubMed: 18780349]
19.
Zhou M, Veenstra T.D. Proteomic analysis of protein complexes. Proteomics. 2007;7:2688. [PubMed: 17639601]
20.
Kabbani N. Proteomics of membrane receptors and signaling. Proteomics. 2008;8:4146. [PubMed: 18763708]
21.
Rohrbough J.G, Breci L, Merchant N, Miller S, Haynes P.A. Verification of single-peptide protein identifications by the application of complementary database search algorithms. Journal of Biomolecular Techniques. 2006;17(5):327. [PMC free article: PMC2291803] [PubMed: 17122065]
22.
Goel M, Sinkins W, Keightley A, Kinter M, Schilling W.P. Proteomic analysis of TRPC5- and TRPC6-binding partners reveals interaction with the plasmalemmal Na+/K+-ATPase. Pflügers Archiv. 2005;451(1):87. [PubMed: 16025302]
23.
Clark K, Middlebeek J, Morrice N.A, Figdor C.G, Lasonder E, van Leewen F.N. Massive autophosphorylation of the Ser/Thr-rich domain controls protein kinase activity of TRPM6 and TRPM7. PLoS ONE. 2008;3(3):e1876. doi:10.1371/journal.pone.0001876. [PMC free article: PMC2267223] [PubMed: 18365021]
24.
Masuda J, Maynard D.M, Nishimura M, Ueda T, Kowalak J.A, Markey S.P. Fully automated micro- and nanoscale one- or two-dimensional high-performance liquid chromatography system for liquid chromatography-mass spectrometry compatible with non-volatile salts for ion exchange chromatography. Journal of Chromatography A. 2005;1063:57. [PubMed: 15700457]
25.
Maynard D.M, Masuda J, Yang X.Y, Kowalek J.A, Markey S.P. Characterizing complex peptide mixtures using a multi-dimensional liquid chromatography–mass spectrometry system: Saccharomyces cerevisiae as a model system. Journal of Chromatography B. 2004;810:69. [PubMed: 15358309]
26.
Eder P, Poteser M, Groschner K. TRPC3: a multifunctional, pore-forming signalling molecule. In: Flockerzi V, Nilius B, editors. Transient Receptor Potential (TRP) Channels. Vol. 179. Berlin Heildelberg: Springer; 2007. p. 77. [PubMed: 17217051]
27.
Nesvizhskii A.I. Protein identification by tandem mass spectrometry and sequence database searching. In: Matthiesen R, editor. Methods in Molecular Biology. Vol. 367. Totowa, NJ: Humana Press; 2006. p. 87. [PubMed: 17185772]
28.
Lu B, Xu T, Park S.K, Yates J.R III. Shotgun protein identification and quantification by mass spectrometry. In: Reinders J, Sickman A, editors. Proteomics, Methods in Molecular Biology. Vol. 564. Totowa, NJ: Humana Press; 2009. p. 261. [PubMed: 19544028]
29.
Slotta D.J, McFarland M.A, Makusky A.J, Markey S.P. Indianapolis, IN: (American Society for Mass Spectrometry); Jun, 2007. 2007. MassSieve: A new visualization tool for mass spectrometry-based proteomics. Paper presented at 55th Annual Conference on Mass Spectrometry and Allied Topics; pp. 4–7.
30.
Lockwich T, Chauthaiwale J, Ambudkar S.V, Ambudkar I.S. Reconstitution of a passive Ca2+-transport pathway from the basolateral plasma membrane of rat parotid gland acinar cells. Journal of Membrane Biology. 1995;148:277. [PubMed: 8747559]
31.
Ong S, Blagoev B, Kratchmarova I. et al. Stable isotope labeling by amino acids in cell culture, SILAC, as a simple and accurate approach to expression proteomics. Molecular and Cellular Proteomics. 2002;1(5):376. [PubMed: 12118079]
32.
Mann M. Functional and quantitative proteomics using SILAC. Nature Reviews. Molecular Cell Biology. 2006 Dec.(7):952. [PubMed: 17139335]
33.
Selbach M, Mann M. Protein interaction screening by quantitative immunoprecipitation combined with knockdown (QUICK) Nature Methods. 2006;3:981. [PubMed: 17072306]
34.
Josic D, Clifton J.G. Mammalian plasma membrane proteomics. Proteomics. 2007;7:3010. [PubMed: 17654460]
Copyright © 2011 by Taylor and Francis Group, LLC.
Bookshelf ID: NBK92813PMID: 22593959

Views

  • PubReader
  • Print View
  • Cite this Page

Related information

  • PMC
    PubMed Central citations
  • PubMed
    Links to PubMed

Similar articles in PubMed

See reviews...See all...

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...