Chapter 13
Plasma Membrane Proteomics Erik Alexandersson, Niklas Gustavsson, Katja Bernfur, Per Kjellbom, and Christer Larsson
Abstract Proteins residing in the plasma membrane have key functions in transport, signal transduction, vesicle trafficking and many other important processes. To better understand these processes it is necessary to reveal the identity of plasma membrane proteins and to monitor modifications and regulation of their expression. This chapter is an overview of the methods used in plant plasma membrane proteomic studies and the results obtained so far. It focuses on studies using mass spectrometry for identification and includes aspects of plasma membrane fractionation, extraction and washing treatments, assessment of purity, separation methods for plasma membrane proteins and choice of techniques for protein cleavage. Finally, the results of plasma membrane proteomic studies are compared and problems with contaminating proteins are discussed.
13.1
Introduction
The plasma membrane is the interface between the cell and its surroundings. This is a position with a central structural role in the cell, and it implies a range of important functions, such as transport of compounds into and out of the cell, communication between the cell exterior and interior, and defence against Erik Alexandersson, Department of Biochemistry, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, 221 00 Lund, Sweden, E-mail:
[email protected] Niklas Gustavsson, Department of Biochemistry, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, 221 00 Lund, Sweden Katja Bernfur, Department of Biochemistry, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, 221 00 Lund, Sweden Per Kjellbom, Department of Biochemistry, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, 221 00 Lund, Sweden Christer Larsson Department of Biochemistry, Center for Chemistry and Chemical Engineering, Lund University, P.O. Box 124, 221 00 Lund, Sweden 186 J. Šamaj and J. Thelen (eds.), Plant Proteomics © Springer 2007
13 Plasma Membrane Proteomics
187
invading pathogens. These functions are thought to be fulfilled by transport proteins, signal transduction components, and by proteins involved in membrane trafficking. In Arabidopsis, there are over 26,000 annotated genes and, according to the “ARAMEMNON” database, about one-quarter of the proteins these genes encode are predicted to have transmembrane domains. The proportion of proteins with predicted transmembrane domains is approximately the same in rice – a monocot plant (Schwacke et al. 2003). The plasma membrane is probably the most diverse membrane of the cell, with a protein composition that varies with cell type, developmental stage and environment; it is likely to harbour thousands of proteins. For instance, receptor-like protein kinases (RLKs) alone, most of which probably reside in the plasma membrane, are represented by more than 600 genes in the Arabidopsis genome (reviewed by Shiu and Bleecker 2001). Several attempts using mass spectrometry (MS) have been made to identify proteins residing in the plant plasma membrane, primarily in Arabidopsis but also to some extent in rice (Tanaka et al. 2004). According to the Arabidopsis sub-cellular proteomic database, SUBA (www.suba.bcs.uwa.edu.au; Heazlewood et al. 2005), close to 600 unique proteins have been identified by MS in eight major proteomic studies of the Arabidopsis plasma membrane (Borner et al. 2002; Elortza et al. 2003, 2006; Nühse et al. 2003; Santoni et al. 2003; Alexandersson et al. 2004; Dunkley et al. 2006; Nelson et al. 2006). There have also been a few studies on plasma membrane proteomics in plants with as yet un-sequenced genomes, e.g. spinach (Kjell et al. 2004), and on plasma membrane lipid rafts in tobacco (Mongrand et al. 2004). Recently, several integral proteins were identified in the plasma membrane of barley seed aleurone layers (Hynek et al. 2006). Although the rice genome has been sequenced, the number of plasma membrane proteomic studies in rice is still limited in comparison to Arabidopsis. Three types of proteomic studies on plasma membrane proteins can be distinguished: (1) studies on highly purified plasma membrane fractions aiming at general identification of proteins (Prime et al. 2000; Santoni et al. 2000, 2003; Marmagne et al. 2004; Alexandersson et al. 2004); (2) studies on highly purified plasma membranes targeting certain groups of proteins by taking advantage of their specific structural properties, such as glycosylphosphatidylinositol (GPI)anchored proteins or phosphoproteins (Borner et al. 2002; Elortza et al. 2003, 2006; Nühse et al. 2003, 2004); and (3) studies that quantify peptides by MSbased methods and compare the amounts of specific proteins between different cell fractions in order to establish their localisations [e.g. LOPIT (localisation of organelle proteins by isotope tagging); Dunkley et al. 2004, 2006; Nelson et al. 2006]. The latter type of study allows for the use of crude membrane fractions. Comparison of proteins in plasma membranes from stressed and nonstressed tissue – sometimes referred to as differential proteomics – has also been reported (Kawamura and Uemura 2003). These and other studies will be further discussed below.
188
13.2 13.2.1
E. Alexandersson et al.
Plasma Membrane Purification and Fractionation Plant Material
The first step in plasma membrane fractionation is the choice of plant material. So far, little attention has been devoted to the biological variation of samples. Thus, there exists a great variability in the plant material used in different studies, and in some studies the plant material used is not well defined. Furthermore, plasma membrane preparations originating from different plant organs, such as roots and leaves, are sometimes pooled, which has two obvious drawbacks: (1) the sample becomes unnecessarily complex with regard to the number of proteins (microarray studies show that gene expression in e.g. leaves and roots are quite distinct from each other, and this is most likely reflected in the protein composition), and (2) information about organ-specific protein localisation is lost. Concerning the plant material used for plasma membrane preparation, parallels can be drawn with the development of microarrays, where the initial focus was on technical advancements, only later becoming more centred on biological questions. In addition, more directed studies on proteins expressed in specific cell types or in response to environmental cues will increase the overall number of plasma membrane proteins identified.
13.2.2
Cell Fractionation
The plasma membrane constitutes only a minor part of the total membrane component of plant cells, which are normally dominated by chloroplasts with their extensive thylakoid membrane (green tissue) or by mitochondria with their internal membrane system (non-green tissue). These larger organelles may be removed by low speed centrifugation (typically 10,000 g for 10 min) and plasma membranes may then be harvested by subsequent high speed centrifugation of the supernatant. The resulting microsomal fraction will, however, still be dominated by membranes originating from chloroplasts and/or mitochondria depending on the tissue used, and will in addition contain membrane vesicles derived from all other types of organelles present in this tissue. The isolation of reasonably pure plasma membranes from this very complex mixture is thus a demanding task. So far, only two techniques have produced plasma membrane fractions of a purity sufficient for proteomic analyses. These techniques are “partitioning in aqueous two-phase systems” (for latest update, see Larsson et al. 1994) and “preparative free-flow electrophoresis” (Sandelius et al. 1986); the two methods are compared and reviewed in Sandelius and Morré (1990). Notably, both methods separate membrane vesicles according to surface properties, and preparations with a purity of 90% or better may be obtained. Since preparative free-flow electrophoresis requires rather sophisticated equipment whereas partitioning in aqueous two-phase
13 Plasma Membrane Proteomics
189
systems can be performed with standard laboratory equipment, the latter technique has been more often employed. Indeed, partitioning in aqueous two-phase systems has been used to obtain plasma membrane preparations from a large number of plant species and tissues ranging from leaves, roots, and wood to cell cultures during the last two decades. Two-phase partitioning has also been the method of choice in Arabidopsis plasma membrane proteomic studies. The use of two-phase partitioning for animal plasma membranes has recently been reviewed by Schindler and Nothwang (2006).
13.2.3
Washing and Extraction Methods
When the plasma membrane fraction has been obtained it is often useful to wash the membranes in order to eliminate soluble contaminating proteins, or to enrich for integral proteins at the expense of peripheral proteins. This also reduces sample complexity and thus enables identification of proteins of lower abundance. Contaminating soluble proteins enclosed in plasma membrane vesicles can be removed by disrupting the vesicles via sonication or freezing and thawing (Palmgren et al. 1990), or by changing the sidedness (polarity) of the vesicles by treatment with the detergent Brij 58 (Johansson et al. 1995). Proteins that are peripherally attached to the plasma membrane can be removed by salt or alkaline treatments, or by organic solvent extraction (for a recent review, see Rolland et al. 2006). An obvious drawback with all membrane washings is the potential loss of water soluble proteins that actually reside in the plasma membrane and consequently are true peripheral plasma membrane proteins. As described in Sect. 13.5.5, quantitative methods such as LOPIT can probably help solve this problem.
13.2.3.1
Sonication/Freezing and Thawing
During sonication or freezing and thawing, membrane vesicles are broken and reformed, which releases proteins enclosed in these vesicles into the soluble phase. These treatments are often combined with salt washing to also remove loosely attached proteins.
13.2.3.1.1
Brij 58
Plasma membrane vesicles obtained by two-phase partitioning have a cytoplasmicside-in orientation. Treatment with the detergent Brij 58 changes the sidedness of the vesicles and thus effectively removes soluble proteins enclosed in the vesicles (Johansson et al. 1995). Brij 58 treatment was used in Alexandersson et al. (2004) in combination with salt washing.
190
13.2.3.1.2
E. Alexandersson et al.
Salt Wash
Ions abolish the electrostatic interactions of contaminating soluble proteins and peripheral membrane proteins as well as the polar headgroups of lipids. This method is used regularly to remove proteins attached to membranes. By varying the salt concentration, a more or less stringent removal of proteins is achieved. Low (0.2 M) to high (1 M) concentrations of salts (usually KCl or NaCl) are used. Salt washes were employed by Alexandersson et al. (2004) and Santoni et al. (1999).
13.2.3.1.3
Alkaline Treatment
Soluble, contaminating proteins and peripheral proteins can also be removed by treatment with sodium carbonate (Fujiki et al. 1982) or NaOH. For example, NaOH treatment of Arabidopsis plasma membrane fractions was carried out by Marmagne et al. (2004). Urea and NaOH treatment was combined to strip membranes of soluble proteins in a study focused on the identification of aquaporins residing in the Arabidopsis plasma membrane (Santoni et al. 2003). To maximise the number of identified aquaporins, which are integral proteins with six transmembrane α-helices, both carbonate and urea/NaOH treatments were tested, and it was concluded that urea/NaOH increased the number of identified aquaporins more, probably because it yielded less complex samples by a more stringent removal of peripheral proteins. Ephritikhine et al. (2004) compared the properties of the proteins identified by either NaOH or salt treatments and concluded that no major differences were seen for Arabidopsis plasma membrane fractions between these two methods. This was in contrast to plastid envelope and mitochondrial membranes, where NaOH treatment removed more peripheral proteins than NaCl treatment.
13.2.3.1.4 Organic Solvent Extraction Alternatively, instead of stripping the plasma membranes of peripheral proteins, integral proteins can be selectively extracted by organic solvents (Seigneurin-Berny et al. 1999). Chloroform/methanol is most commonly used, as reported by Marmagne et al. (2004).
13.2.3.1.5
Triton X-114 Fractionation
Using this technique, integral proteins are enriched in the lower, detergent-rich phase, and peripheral proteins are distributed to the upper, water phase (Bordier 1981). Triton X-114 fractionation was tested by Santoni et al. (2000).
13 Plasma Membrane Proteomics
13.2.3.1.6
191
Other Methods
Additional methods to eliminate contaminating proteins are continuously being developed. For instance, in barley aleurone layers the soluble protein β-amylase was recognised to contaminate plasma membrane preparations in spite of washing by salts and treatment with Triton X-100 or Brij 58. Hynek et al. (2006) therefore developed a method to exclude soluble proteins by employing batch reversed-phase chromatography. In this manner, it was possible to separate soluble and integral proteins, represented by β-amylase and H+-ATPases, respectively, by elution with different concentrations of 2-propanol. In another attempt to eliminate contaminating proteins, Nielsen et al. (2005) was able to identify a large number of plasma membrane proteins in mouse brain tissue by using a combination of high-salt, carbonate and urea washes, as well as digitonin, which improves the separation of animal plasma membranes from other membranes.
13.2.4
Assessment of Purity
Purity is usually assessed by determining the levels of different markers in the plasma fraction with e.g. the microsomal fraction as a reference fraction. The most commonly used marker for the plasma membrane has been H+-ATPase, and for the main contaminants in the microsomal fractions, chloroplast thylakoids and mitochondrial inner membranes, chlorophyll and cytochrome c oxidase are often used. Suitable marker proteins can be found for most membranes, and the levels can be determined either by measuring enzyme activities or by using antibodies. However, it should be realised that determining enrichment ratios of known plasma membrane proteins and depletion ratios of contaminating proteins in the plasma membrane fraction produces only relative values, and does not indicate whether a fraction is 70% or 90% pure. Absolute purity can be determined by using a specific stain for the plasma membrane: silicotungstic (or phosphotungstic) acid at pH 3 (Roland 1978; Morré 1990). Using this stain, plasma membrane vesicles appear as well stained membranes in electron micrographs, whereas other membranes are only faintly stained. Some simple morphometric procedure may then be used to calculate the proportion of well stained membranes. The purity of plasma membranes obtained by two-phase partitioning and free-flow electrophoresis was determined as 89–97% by this procedure (e.g. Kjellbom and Larsson 1984; Sandelius and Morré 1990; Morré 1990). MS can also be used to verify purity. As shown in Fig. 13.1, the enrichment or depletion of peptides unique to different marker proteins (or group of proteins) can be determined. This can be done by using stable isotope labelling such as isobaric tags for relative and absolute quantification (iTRAQ) reagent tags (Applied Biosystems, Framingham, MA) as in Fig. 13.1 or by using H218O or H216O during trypsinisation to incorporate 18O or 16O into the C-termini of peptides, as in Nelson et al. (2006). Note that the enrichment of plasma membrane proteins in the plasma membrane fraction will vary, as shown in Fig. 13.1 for two peptides representing two aquaporin isoforms, PIP2;1 and PIP2;7, and a peptide
Fig. 13.1 Assessment of plasma membrane purity using isobaric tags for relative and absolute quantification (iTRAQ). Tryptic peptide extracts from the plasma membrane fraction (PM), the intracellular membrane fraction (ICM), and the corresponding microsomal fraction (MF) were labeled with different iTRAQ reagent tags (Applied Biosystems, Framingham, MA) and mixed in a 1:1:1 ratio based on protein concentration. The mixed peptide extract was fractionated by strong cation exchange chromatography followed by reversed phase liquid chromatography, and fractions were collected directly on a matrix assisted laser desorption/ionisation (MALDI) sample plate. Each fraction was analysed by MALDI-time-of-flight (TOF) mass spectrometry (MS) and selected peptides were analysed by tandem MS (MS/MS). Upon MS/MS analysis of a peptide, the three different tags yield different reporter ions of m/z 115, 116 and 117, respectively. The intensities of the three reporter ions represent the concentration of the same peptide in each of the three membrane fractions (115–PM, 116–MF, 117–ICM). The PM and ICM fractions were obtained by aqueous two-phase partitioning of an Arabidopsis leaf MF. No washing procedure was used, which may explain the relatively high level of the chloroplast soluble enzyme ribulose 1,5-bisphosphate carboxylase/oxygenase (RuBisCo) activase. a MS-spectrum of one LCfraction in which a peptide with m/z 1,409.65 was detected. The peptide was selected for MS/MSanalysis. b MS/MS-spectrum of the peptide selected in a. Inset Mass region in which the iTRAQ reporter ions appear. The peptide was identified as amino acids 44–55 (GPSGSPWYGSDR) from LHC-II type 1, a chloroplast thylakoid protein depleted in the plasma membrane preparation as (continued)
13 Plasma Membrane Proteomics
193
representing four H+-ATPase isoforms. The reason may be different rates of turnover, which will result in different proportions of the proteins in the membranes of the endoplasmic reticulum (ER) and in the Golgi stacks. It may also be the result of endocytosis and exocytosis to recycle plasma membrane proteins, which seems to be an important regulatory mechanism also in plants and will result in different proportions of the proteins in endosomes (reviewed by Murphy et al. 2005; Šamaj et al. 2005).
13.3 13.3.1
Protein Separation Gel-Based Techniques
In early proteomic studies of the plant plasma membrane, two-dimensional gels were used to separate proteins (Prime et al. 2000; Santoni et al. 2000). However, few integral proteins were identified as it proved difficult to resolve these by non-ionic or zwitterionic detergent in the first iso-electrofocusing (IEF) step. Instead, onedimensional gels were used to increase the number of resolved integral proteins; an approach that was successfully used in several studies on cellular membranes (e.g. Millar and Heazlewood 2003; Alexandersson et al. 2004; Marmagne et al. 2004; Peltier et al. 2004). Further development of two-dimensional electrophoresis (2-DE) to resolve hydrophobic proteins was undertaken; e.g. Navarre et al. (2002) reported the successful identification of integral plasma membrane proteins in yeast by the use of the cationic detergent trimethyl ammonium bromide (CTAB) in the IEF step. Other types of modified two-dimensional systems have also been developed, such as BAC/SDS-PAGE, which involves gel electrophoresis in an acidic buffer system using the cationic detergent benzyldimethyl-n-hexadecylammonium chloride (16-BAC) in the first dimension and SDS-PAGE in the second dimension (Hartinger et al. 1996). Lately, different variants of 2-DE, in which SDS-PAGE is used in both the first and second dimensions (termed doubled SDS-PAGE or dSDS-PAGE), have been developed (Rais et al. 2004). In such systems, the acrylamide concentration, urea content and trailing ion used in the gels are altered between the two dimensions. Burré et al. (2006) performed a comparison of identified synaptic vesicle proteins resolved by either one-dimensional SDS-PAGE, BAC/SDS-PAGE or dSDS-PAGE and demonstrated that the three gel-based separation methods were partly complementary to each other regarding resolution and identification of integral proteins. shown by the low intensity of the 115 reporter ion. c Examples of plasma membrane proteins enriched in the plasma membrane fraction (protein ratios >1; aquaporin isoforms PIP2;1 and PIP2;7, and plasma membrane H+-ATPase isoforms AHA1, 2, 6 and 8), and proteins depleted in the plasma membrane preparation (protein ratios
