Proteomic reactors and their applications in biology

Authors

  • Hu Zhou,

    1.  Ottawa Institute of Systems Biology (OISB), University of Ottawa, ON, Canada
    2.  Department of Biochemistry, Microbiology and Immunology, University of Ottawa, ON, Canada
    3.  Shanghai Institute of Materia Medica, Chinese Academy of Sciences, China
    Search for more papers by this author
  • Zhibin Ning,

    1.  Ottawa Institute of Systems Biology (OISB), University of Ottawa, ON, Canada
    2.  Department of Biochemistry, Microbiology and Immunology, University of Ottawa, ON, Canada
    Search for more papers by this author
  • Fangjun Wang,

    1.  Ottawa Institute of Systems Biology (OISB), University of Ottawa, ON, Canada
    2.  Department of Biochemistry, Microbiology and Immunology, University of Ottawa, ON, Canada
    3.  Key Lab of Separation Sciences for Analytical Chemistry, Dalian Institute of Chemical Physics, China
    Search for more papers by this author
  • Deeptee Seebun,

    1.  Ottawa Institute of Systems Biology (OISB), University of Ottawa, ON, Canada
    2.  Department of Biochemistry, Microbiology and Immunology, University of Ottawa, ON, Canada
    Search for more papers by this author
  • Daniel Figeys

    1.  Ottawa Institute of Systems Biology (OISB), University of Ottawa, ON, Canada
    2.  Department of Biochemistry, Microbiology and Immunology, University of Ottawa, ON, Canada
    Search for more papers by this author

  • Note
    Hu Zhou and Zhibin Ning contributed equally to this review

D. Figeys, Department of Biochemistry, Microbiology and Immunology, University of Ottawa, 451 Smyth Road, Ottawa, ON, Canada K1H 8M5
Fax: +1 613 562 5655
Tel: +1 613 562 5800 ext 8674
E-mail: dfigeys@uottawa.ca

Abstract

Proteomic analysis requires the combination of an extensive suite of technologies including protein processing and separation, micro-flow HPLC, MS and bioinformatics. Although proteomic technologies are still in flux, approaches that bypass gel electrophoresis (gel-free approaches) are dominating the field of proteomics. Along with the development of gel-free proteomics, came the development of devices for the processing of proteomic samples termed proteomic reactors. These microfluidic devices provide rapid, robust and efficient pre-MS sample procession by performing protein sample preparation/concentration, digestion and peptide fractionation. The proteomic reactor has advanced in two major directions: immobilized enzyme reactor and ion exchange-based proteomic reactor. This review summarizes the technical developments and biological applications of the proteomic reactor over the last decade.

Abbreviations
IEBPR

ion exchange-based proteomic reactor

IMER

immobilized enzyme reactor

SAX

strong anion exchange

SCX

strong cation exchange

WAX

weak anion exchange

WCX

weak cation exchange

Introduction

Proteomic studies rely heavily on a number of different techniques (sample processing, HPLC, MS, bioinformatics) that enable the identification and quantitation of thousands of proteins. Originally, at the core of proteomics was 2D gel electrophoresis which permitted the separation of thousands of protein variants (spots). It was expected that the combination of 2D gel electrophoresis and MS would allow for the identification of thousands of proteins. However, for different technical reasons, including the reproducibility of 2D gel electrophoresis and the labor involved in gel cutting, this was not achieved on a regular basis. Instead, gel-free approaches, such as multidimensional protein identification technology, were proposed as alternatives to 2D gel electrophoresis. Although thousands of proteins are routinely identified using ‘gel-free’ approaches, these technologies still have some major drawbacks in terms of analysis time, automation, sensitivity and robustness. Furthermore, hydrophobic proteins (like membrane proteins), post-translationally modified proteins, low-abundance proteins and minute amount of samples still remain major challenges to current proteomic technologies. These challenges are the driving force behind efforts to develop technologies for the integration of cell/tissue protein extraction, preconcentration, protein fractionation, protein reduction/alkylation, protein digestion and peptide fractionation. Recently, microfluidic devices, such as the proteomic reactor, that integrate multiple protein processing steps into a single miniaturize device were introduced. The proteomic reactor is a microfluidic device that was originally designed to process a small amount of sample, especially for protein digestion. Almost all of the pre-MS analysis procedures can be performed on the reactor, including protein concentration, desalting, buffer exchange, reduction, alkylation and digestion, resulting in a reduction in sample loss and an improved limit of detection. Two major strategies were applied in the development of the proteomic reactor: the first strategy is based on immobilized enzyme reactor (IMER), whereas the second strategy is based on ion exchange-based proteomic reactor (IEBPR). The recent developments of these proteomic reactors and their applications are summarized in the following sections.

Initial reports

The initial reports on IMER for proteomic study were from Wang et al. [1] who presented a microfluidic-ESI MS-interfaced device, which was used for integrated tryptic digestion, separation and electrospray of proteins. Protein digestion was performed in a microfluidic chip with immobilized trypsin beads packed either within the sample inlet reservoir or in a packed bed. The resulting coverage of the amino acid sequence ranged from 92% for cytochrome c to 71% for BSA. The first report on IEBPR was by Ethier et al. [2] who presented the development of a single microfluidic device, termed the proteomic reactor, which consisted of a small bed of packed strong cation-exchange material (SCX) for sample processing. As shown in Fig. 1, proteomic samples are acidified and loaded onto the SCX reactor at a low pH (< 3), and the Eppendorf tube and filtration membrane can be used as alternatives to the column. Most proteins are positively charged at such a low pH and absorbed onto the reactor material while the nonionic detergents and contaminants are readily washed away. The dithiothreitol and iodoacetamide reagents are introduced into the reactors for protein reduction and alkylation, respectively. The previously loaded trypsin is then activated by increasing the pH to 8 for protein digestion. The resulting peptides are eluted by a LC-MS/MS compatible buffer. Ethier et al. demonstrated that the proteomic reactor is at least 10× more sensitive than current gel-free methodologies with one protein identified per 440 pg of protein lysate injected onto the reactor. Furthermore, as few as 300 cells can be directly introduced onto the proteomic reactor and analyzed by MS with 17 proteins identified. Although only a few proteins were identified from 300 cells, the improvement in analytical performances was also beneficial when more cells were utilized. More developments in IMER and IEBPR have been reported since these initial reports.

Figure 1.

 Flow chart of the proteomic reactor. Sample preparation: the proteomes are extracted from yeast, cell or tissue. The proteomes can be prefractionated using size-exclusion chromatography or subcellular fractionation. The proteomes are then processed on the proteomic reactor in the column, in the tube or on the filtration membrane. Finally, the resulting peptides are analyzed using LC-MS/MS.

Manipulation of enzyme activity

The development of an IMER-based proteomic reactor requires maintaining the catalytic activity of the immobilized enzyme and having a high mass transfer rate of the substrate between the mobile and stationary phases. These two factors will determine the efficiency of the IMER. Several groups have focused their research on improving or maintaining the enzyme catalytic activity during the immobilization procedures. Freije et al. [3] reported an integrated protein analysis platform using immobilized, acetylated trypsin for enhancing digestion efficacy. Trypsin acetylation was performed in solution by gradual, stepwise addition of 1 m acetic acid N-hydroxysuccinimide ester. Freije et al. [3] reported that the acetylation of immobilized trypsin resulted in enhanced enzyme stability and catalytic activity, with fewer interfering tryptic autolysis products. Sim et al. [4] proposed the first application of a temperature-controllable microreactor for sample preparation by fabricating a heater and a temperature sensor within the microreactor. Thermal denaturation of the sample protein was performed at 85 °C for 1 min, followed by 10 min protein digestion at 37 °C [4]. Percy and Schriemer [5] described rheostatic control of solvent composition of tryptic digestion in an immobilized enzyme reactor. Programmable solvent waveforms were used to adjust the solvent composition (0–45% acetonitrile) of a binary gradient system. The organic solvent compositions can be changed to regulate the fragment length of the tryptic digest products from full digestion (0 missed cleavages) to no digestion (intact protein) [5]. Liu et al. demonstrated that an inflation bulb-driven microfluidic reactor can be applied for IR-assisted proteolysis. This device contained an inflation bulb-driving system, a simple cross-polymethyl methacrylate microchip and a temperature-controllable IR radiation system [6]. IR radiation improved proteolysis efficiency, reducing the digestion time from 12 h (using conventional in-solution digestion) to 5 min [6]. Yamaguchi et al. [7] demonstrated a protease-immobilized microreactor using biofunctional cross-linker agents, paraformaldehyde and glutaraldehyde, to immobilize trypsin and chymostrypsin to poly(tetrafluoroethylene) microtube. They obtained promising results from this microreactor using standard proteins, and the numbers of identified peptides and Km value (representing the binding affinity between enzyme and substrate) were similar to in-solution digestion. However, performances for more complex samples and lower concentrations were not assessed. Spross and Sinz [8] described a capillary trypsin IMER by immobilizing trypsin onto a poly(glycidyl methacrylate-co-acrylamide-co-ethylene glycol dimethycrylate) monolith using the glutaraldehyde technique. Good performances were obtained using protein standards even when one protein was present at 1000× the level of the other proteins in a simple protein mixture.

By contrast, in the IEBPR, the enzyme is not immobilized and can be introduced using two different approaches. The first method consists of mixing the enzyme (e.g. trypsin) and the proteomic sample in a ratio ranging from 1 : 4 to 1 : 10 at low pH. The sample and enzyme mixture can then be loaded onto the proteomic reactor where the enzyme can be activated by changing the pH [2,9–13]. Alternatively, the enzyme (trypsin) can be infused into the reactor at a concentration of 2 μg·μL−1 when the previously loaded proteomic sample is ready for digestion [14,15]. Proteomic reactors have been designed using strong cation exchange (SCX) and strong anion exchange (SAX) material, which affects the loading pH and therefore the activity of the enzymes. For example, in the SAX proteomic reactor, Glu-C and chymotrypsin digestion have lower identification performances than in the SCX reactor, which is likely due to their partial denaturation at the high pH (pH 12) [12]. By contrast, other enzymes, such as trypsin, are not affected by the different conditions required by the SAX and SCX proteomic reactors [12]. In earlier versions of the IEBPR, digestion was performed by changing the pH to 8.0 using 100 mm Tris/HCl (pH 8.0) [2,9–11]. Since then, it has been realized that the salt concentration affects the binding of proteins and the elution of peptides. Instead, in later versions, the salt concentration of the digestion buffer was reduced to 20 mm of ammonia bicarbonate [12–14] or as low as 10 mm ammonia bicarbonate [15], which was suitable for peptide elution and fractionation. Digestion can be performed either at room temperature [2,9,11,15] or at 37 °C [12–14]; however, the performance of the IEBPR at different temperatures has not been assessed.

Improvements of supporting material

To date, most implementations of IMER have used support material based on poly(vinylidene fluoride) porous membrane, monolithic material, metal-ion-based or nanoparticle-based material and surfaces that can be regenerated. Gao et al. described a miniaturized membrane reactor for proteolytic digestion with online ESI-MS identification. In this device, microfluidic channels were fabricated on a poly(dimethylsiloxane) substrate, whereas the IMER consisted of trypsin adsorbed on a poly(vinylidene fluoride) porous membrane [16]. Based on the large surface area-to-volume ratio of porous membrane media, the authors reported a much higher catalytic activity. The authors reported peptide identification for cytochrome c using as little as 0.04 pmol; however, the performance of the device using real biological samples was not tested. The analyte residence time inside the membrane, the analyte concentration and the reaction temperature need to be controlled for optimal digestion performance [16]. Pereira-Medrano et al. [17] demonstrated a novel glass/poly(dimethylsiloxane) micro-immobilized enzyme reactor with enzymes covalently immobilized onto poly(acrylic acid) plasma-modified surfaces for membrane proteomics analysis. To their credit, the authors tested their devices using real biological samples derived from bacterial membrane fractions of Synechocystis sp. PCC 6803. They reported a lower number of identified proteins (2/3) but much faster processing time compared with in-solution digestion. Duan et al. developed a monolithic enzymatic microreactor in a fused-silica capillary by in situ polymerization of acrylamide, glycidyl methacrylate and ethylene dimethacrylate in the presence of a binary porogenic mixture of dodecanol and cyclohexanol. This was followed by an ammonia solution treatment, glutaraldehyde activation and trypsin immobilization [18]. Although interesting performances were reported using this device, it still needs to be tested using real biological samples. Lin and Skinner [19] described the preparation and characterization of methacrylate monolithic enzyme reactors using fused-silica capillaries. Wu et al. [20] developed titania- and alumina-based poly(dimethylsiloxane) microfluidics enzymatic reactors for rapid protein digestion. Microfluidics with microchannel and stainless steel tubing were fabricated using poly(dimethylsiloxane) casting and O2-plasma treatment to generate a layer of silanol groups on the channel wall. The silanol groups on the microchannel wall can then be used for covalent attachment via the hydroxyl groups of trypsin encapsulated in titania or alumina sol matrix [20]. Slightly better results were obtained with this device than with in-solution digestion for the analysis of BSA. Unfortunately, the device was not tested using real biological samples. Liu et al. reported a microchip reactor coated with a gold nanoparticle network in which trypsin was entrapped. This design provides efficient online proteolysis of low-level proteins and complex extracts originating from mouse macrophages [21]. Liu et al. reported roughly the same number of identified proteins using their device compared with in-solution digestion of the same amount of protein. Shui et al. reported a novel design of a proteolytic nanoreactor using synthesized 3D nanopore-based mesoporous silica for sample enrichment, purification and efficient proteolysis. The versatility of this technique meets the practical demands of rapid and comprehensive proteomic analysis of protein mixtures [22]. In particular, they reported 15× more identified proteins on their device compared with in-solution digestion for a nuclear protein extract isolated from mouse liver cells. Qiao et al. [23] reported a nanoreactor based on cyano-functionalized mesoporous silicate for tryptic digestion of proteins within the mesochannels. They tested their device using a protein extract from human liver tissue and reported 165 proteins identified using 5 μg of protein extract. Bílkováet al. [24] reported an easily replaceable protease microreactor for microfluidic devices using magnetic particles coated with poly(N-isopropylacrylamide), polystyrene, poly(2-hydroxyethyl methacrylate-co-ethylene dimethacrylate), poly(glycidyl methacrylate), [(2-amino-ethyl)hydroxymethylen]biphosphonic acid or alginic acid with immobilized trypsin. They obtained limited proteolysis by changing the residence time of the proteins within the microreactor. Li et al. [25] reported a regenerable protease microreactor based on metal-ion-chelated adsorption of enzyme: small microspheres (200 nm in diameter) with strong magnetism were synthesized and their surfaces modified to chelate metal ions. Ionic copper and trypsin were subsequently adsorbed onto the microsphere surfaces and the beads were incorporated into the channel of the microreactor. The researchers managed to identify a few proteins from a single fraction of a rat liver extract fractionated by RPLC. Using a similar approach, Ma et al. [26] developed a metal-ion chelate-immobilized enzyme reactor (that can be regenerated) in which a metal chelating agent is attached to an organic–inorganic hybrid silica monolith in a capillary. Once activated by ionic copper, the capillary can be used to immobilize trypsin and to process proteomic samples [26]. Moreover, the capillary can be stripped of trypsin using EDTA and reactivated with ionic copper. The authors reported faster digestion and a slightly higher number of proteins identified using this approach compared with in-solution digestion for processing of rat liver extracts. Wang et al. described that a bioreactor can be made by inserting a piece of trypsin-immobilized glassfiber into the needle of a syringe. The syringe can then be used as a flow-through bioreactor and can be regenerated by simply changing the core piece of glassfiber [27]. The device was only tested using standard proteins.

In 2006, the ion-exchange-based proteomic reactor was developed by the Figeys’ lab [2]. Since then, ion-exchange materials have improved. Ion-exchange chromatography materials are generally classified as either SCX, weak cation exchange (WCX), SAX or weak anion exchange (WAX). The extremely narrow tolerance in pH range of WCX and WAX cannot be used to capture proteins with various pI values. At the core of the proteomic reactor is the ability to modify the electrostatic interactions between charged amino acid side chains (or N-/C-terminus) and the surface charge of ion-exchange materials. SCX is the most commonly used ion-exchange technique in proteomic research [28–37]. At low pH, proteins and peptides often carry multiple charges, whereas SCX material carries negative charges. The ionic bonds between proteins and SCX material can be modulated by the pH and the ionic strength leading to the capture and release of proteins from the SCX. There are two major forms of beads in ion-exchange chromatography, i.e. silica- and polymer-based. In the earlier version of the proteomic reactor, silica-based SCX beads were packed into a silica capillary tube for sample processing [2,4,10,11]. However, silica-based sorbents are usually stable within a narrow pH range of ∼ 2–7.5, but unstable at higher pH. By contrast, the polymer-based sorbents (e.g. styrene-divinylbenzene) are stable within a much wider pH range of 0–14 [38]. Therefore, polymer-based ion-exchange beads were used in later studies [12–14]. Beside SCX beads, SAX beads were also applied in their studies [12]. Furthermore, the SCX monolithic matrix, prepared by direct in situ polymerization of ethylene glycol methacrylate phosphate and bisacrylamide in a trinary porogenic solvent consisting dimethylsulfoxide, dodecanol and N,N′-dimethylformamide, was utilized in this proteomic reactor [15]. A polymer-based multiplexed microfluidic proteomic reactor was also developed by Tian et al. which was used for parallel analysis of small amounts of immunopurified protein samples. Its design allowed for the simultaneous processing of multiple samples on the same device. Each reactor was made of SCX beads that were packed and restricted into a 1-cm microchannel by two integrated pillar frits [39]. The device is manufactured from a combination of low-cost hard cyclic olefin copolymer thermoplastic and elastomeric thermoplastic materials, for example, styrene, ethylene or butylenes, using rapid hot-embossing replication techniques with a polymer-based stamp [39]. Zhou et al. [40] developed a simplified and user-friendly reactor compatible with bench-top centrifuges to analyze membrane proteins. The centrifugal proteomic reactor [40] retains the key characteristics of the proteomic reactor such as a small volume of slurry, as little as 10 μL can be used, small loading amount, 20 μg, and rapid processing time, 2.5 h. Moreover, the centrifugal proteomic reactor has several operational advantages over the previously described column-based proteomic reactor [2,10–14,39,41]. These advantages include: (a) improved membrane protein preparation; (b) simplified and user-friendly experimental procedures; and (c) integration of all the steps using a bench-top centrifuge instead of using a pressurized vessel which can be performed in any proteomic laboratory [40].

Integration with protein and peptide separation/fractionation

Already, some IMER have been incorporated into proteomic workflows. For example, Yuan et al. [42] established an integrated proteomic analysis platform by combining protein separation (mixed WAX and WCX resin or size exclusion chromatography) with online digestion using a trypsin micro-immobilized enzyme reactor. Protein identification was performed by microflow (μ)-RPLC with ESI-MS/MS. Although experiments were performed using real and complex biological samples, the number of proteins identified remains limited because of the sensitivity of the mass spectrometer utilized (LCQDuo). In addition, Sun et al. [43] demonstrated an integrated device for protein fractionation with RPLC, online sample buffer exchange to remove the organic solvent and adjust the pH, protein enrichment and digestion by using a membrane interface and a monolithic hybrid silica-based IMER. Jiang and Lee [44] reported an integrated platform by online coupling of the μ-enzyme reactor with μ-membrane chromatography to perform trypsin digestion and peptide separation/identification. The immobilized trypsin-containing μ-enzyme reactor was used to perform rapid protein digestion. The μ-membrane chromatography used a membrane media sandwiched between the poly(dimethylsiloxane) microchannels as the stationary phase for the separation of peptides based on differences in their hydrophobicity. The device was only used to process standard proteins. Hou et al. [9] reported a 96-well plate proteomic reactor that was coupled with protein fractionation by size-exclusion chromatography for large-scale identification of proteins. MCF7 cell lysate (400 μg) was separated into 35 protein fractions by size-exclusion chromatography followed by an on-plate reactor digestion, resulting in a total of 2683 unique identified proteins with a 1% false-positive rate [9]. Zhou et al. [11] reported processing 15 protein fractions (from the post-nuclear supernatant of HUH7 cell homogenate separated by a 10–40% continuous sucrose gradient) on the proteomic reactor. They reported the identification of > 1100 phosphopeptides using this approach. Zhou et al. [13] combined the proteomic reactor (SAX and SCX) with a step pH gradient elution approach (for the SAX reactor: pH from 12 to 2.5; for the SCX reactor: pH from 2.5 to 12) for the identification of lower abundance proteins.

Incorporation of multiple enzymes

For some applications, having the highest sequence coverage is much more important than the number of identified proteins. To achieve higher sequence coverage in protein identification, multiple enzymatic digestions have been utilized to analyze total cell/tissue lysates or affinity-purified protein complexes. Lin and Skinner [19] demonstrated that a polymer-based monolithic enzyme reactor, made of enzymes (trypsin and Glu-C) immobilized on porous methacrylate, can be used for online protein digestion. This online reactor led to 87% sequence coverage of cytochrome c, which is 13% higher coverage than with in-solution digestion [19]. Liuni et al. reported a microfluidic reactor for proteolysis with on-chip ESI-MS analysis. On-chip digestion was performed on a wide (1.5 cm), shallow (10 mm) reactor ‘well’ that is functionalized with pepsin–agarose [45]. Zhou et al. [12] reported that SAX beads can be also used in the proteomic reactor and that multiple enzymes (trypsin, chymotrypsin and Glu-C) can enhance protein identification and sequence coverage to provide confident protein identification. Combination of the identified peptides with three different enzymes (trypsin, chymotrypsin and Glu-C) resulted in 82% sequence coverage for the SAX reactor and 99% sequence coverage for the SCX reactor [12]. Ma et al. [46] presented an integral membrane protein analysis method by coupling formic-acid-assisted solubilization and pepsin-based IMER (pepsin-IMER). To date, limited numbers of enzymes (trypsin, Glu-C, chemotrypsin and pepsin) have been incorporated into IMER and IEBPR. More enzymes, such as proteinase K, Arg-C and so on should be considered as new candidates for protein digestion. Table 1 summarizes the parameters that need to be considered in the development of both the IEBPR and IMER.

Table 1.   Summary of the parameters utilized in proteomic reactor. NA, not applicaple.
ParameterIon exchange-based proteomic reactorImmobilized enzyme reactor
Supporting material
 Porous membraneNAApplicable
 Regeneration of enzymeNA
 Monolithic/polymer-based
 Silica-basedNA
Buffer compatibility
 Ionic detergentNot compatibleNA
 Noionic detergentNA
 Salt concentrationHigh salt concentraion is not compatibleNA
 UreaAs high as 8 mNA
 Organic solvent0–100% organic solventOrganic solvent composition can be used to regulate enzyme activity
 pH tolerancePolymer based: pH 1–14NA
Multiple enzyme
 Trypsin
 Glu-C
 Chymotrypsin
 PepsinNA
Fractionation
 Protein fractionation
  Size-exclusion chromatography
  Subcellur fractionationNA
  RPLCNA
  Weak anion exchange/weak cation exchangeNA
  μ-membrane chromatographyNA
 Peptide fractionationStep pH gradientNA
Other
 TemperatureRoom temperature, 37 °CSensitive
 IR assistantNA
 Enzyme activity manupinationNAAcetylation
 Processing timeSeveral minutes to hours2–3 h

Applications of proteomic reactor

The critical parameter for the evaluation of IMER and IEBPR is their successful application for analyzing real biological samples. Ma et al. [46] reported a facile membrane protein profiling method: membrane proteins were solubilized by formic acid, online digested by pepsin-IMER and analyzed by SCX and μ-RPLC-ESI-MS/MS. Using this method, 235 unique proteins were identified from a rat liver microsome membrane protein fraction, and 39% (91/235) were annotated as membrane proteins with one or more transmembrane domains [46]. Zhou et al. [40] used the centrifugal proteomic reactor for membrane protein analysis, resulting in the identification of 945 plasma membrane proteins and 955 microsomal membrane proteins, of which 63% and 47% were predicted as bona fide membrane proteins, respectively. Vasilescu et al. [14] demonstrated that the proteomic reactor permits the analysis of affinity-purified proteins by LC-MS/MS, resulting in 27 ubiquitination sites to be precisely mapped on 21 proteins and the identification of 58 candidate ubiquitinated proteins. Zhou et al. [14] developed a glycoproteomic reactor that combined protein concentration and purification, disulfide bond reduction, peptide-N-glycosidase-mediated 18O-labeling and deglycosylation, alkylation, tryptic digestion and pH-based fractionation. In total 82 unique glycopeptides, representing 41 unique glycoproteins, were identified from as little as 5 μL of human plasma. Zhou et al. used subcellular fractionation and a novel phosphoproteomic reactor that combined a SCX proteomic reactor and phosphopeptide enrichment by Ti-IMAC for the analysis of the subcellular phosphoproteome. They reported the identification of 1141 unique phosphopeptides from subcellular fractions from HuH7 cells [11]. Zhou et al. demonstrated, using step pH gradients on the SAX and SCX proteomic reactors, that 1106 and 685 proteins, respectively, could be identified from yeast lysate. By contrast, only 613 proteins were identified using the conventional in-solution method (2D LC-MS/MS based on 10 salt steps) [13]. Usually, yeast proteins with a codon adaptation index of < 0.2 are defined as low-abundance proteins. Forty-eight and 28% of the proteins identified in the pH fractions from the SAX and SCX proteomic reactors have a codon adaptation index < 0.2. By contrast, only 22% of the proteins identified from the 2D LC-MS/MS runs have a codon adaptation index < 0.2 [13]. These results indicated that the proteomic reactor with pH fractionation is a powerful and promising tool for identifying low-abundance proteins from minute amounts of sample [13]. Tian et al. [15] demonstrated a novel rare-cell proteomic reactor using a SCX monolithic matrix that can achieve protein identification and quantification efficiently from only 50 000 human embryonic stem cells. More than 2200 proteins were identified from 50 000 human embryonic stem cells using this rare-cell proteomic reactor. Tian et al. [39] developed a multiplexed microfluidic proteomic reactor, and its limit of detection can reach < 2 ng of protein. This microfabricated proteomic reactor allowed for the simultaneous processing of multiple samples on the same devices. For example, immunopurified samples from the yeast protein Htz1 from untagged control, Htz1-TAP and Htz1-TAP swr1Δ were analyzed in parallel on this multiplexed microfluidic reactor, resulting in the identification of 26 Htz1-interacting proteins [39]. Applications of proteomic reactors for the processing of complex biological samples remain limited to some specific fields; however, these microfluidic devices have already shown advantages over conventional in-solution digestion methods. Furthermore, the proteomic reactor has the potential to integrate almost all the steps of proteomic workflows. Table 2 shows a brief summary of the applications of both types of proteomic reactors (IEBPR and IMER).

Table 2.   Summary of the biological applications of proteomic reactor. NA, not applicable.
ApplicationsIon exchange-based proteomic reactorImmobilized enzyme reactor
Membrane protein
 Sample loading amount20 μg [40]NA
 Identified proteins945 (plasma membrane);235 [46]
 955 (microsomal membrane) [40]
 Membrane protein percentage63% (plasma membrane);39% [46]
 47% (microsomal membrane) [40]
Post-translational modifications
 Ubiquitination
  Sample loading amountFrom immunopurification of ubiquitinated proteins [10]NA
  Identified proteins79 [10]NA
  Sites27 (from 21 ubiquitinated proteins) [10]NA
 Glycosylation
  Sample loading amount5 μL of human plasma [14]NA
  Identified proteins122 [14]NA
  Sites63 (from 41 glycoproteins)[14]NA
 Phosphorylation
  Sample loading amountFrom subcellular fractions[11]NA
  Identified proteins621[11]NA
  Sites1141 (from 621 phosphoproteins) [11]NA
Low-abundance protein
 Sample loading amount50 μg [13]NA
 Identified proteins685 (by strong cation exchange reactor);NA
1106 (by strong anion exchange reactor) [13]
 Low-abundance protein percentage28% (by strong cation exchange reactor);NA
48% (by strong anion exchange reactor) [13]
Rare cells and quantification
 Sample loading amount50 000 human embryonic stem cells [15]NA
 Quantified proteins2200 [15]NA
Protein–protein interaction
 Sample loading amountFrom yeast modified chromatin immunopurification (mChIP) [39]NA
 Identified proteins27 [39]NA

Other types of proteomic reactors

Wisniewski et al. [47] reported the filter-aided sample preparation method to process proteomic samples, which is also called a proteomic reactor by the authors. It is described as a universal sample preparation method for proteome analysis. The core innovation of filter-aided sample preparation resides in the use of the same filter to remove detergent, to exchange buffers and to filter the undigested proteins as well as trypsin. Specifically, a 10K filter device is used as the core reactor, and most of the proteins can be retained by the filter, whereas the detergent, as well as the other small molecules (dithiothreitol and iodoacetamide), flows through the device. The digestion is then performed in the filter device, and the digested peptides are eluted, whereas the protease and un-digested proteins are retained by the filter. Because high concentrations of urea and detergents can be easily removed using this approach, it is useful for processing membrane proteins. Filter-aided sample preparation combines the advantages of in-gel and in-solution digestion for protein identification. It was also combined with a StageTip SAX fractionation method for the analysis of tissue membrane [48]. They reported the identification of more than 4000 proteins from mouse hippocampus. Filter-aided sample preparation can handle milligram amounts of protein and can be coupled to different fractionation methods [49]. It is has been used for the post-translational modification analysis of formalin-fixed and paraffin-embedded [50] tissues and normal tissues [51].

Kadiyala et al. [48] proposed a single-tube sample preparation method which might provide another form of in-tube reactor. Perfluorooctanoic acid was employed to replace the commonly used nonvolatile surfactants, such as SDS, and volatile triethyl phosphine was used for the cysteine reduction. The replacement of surfactants and other chemicals with volatile compounds resulted in a single-tube sample preparation for the subsequent MS analysis. The method does not employ clean-up devices and membrane filters, which can minimize sample loss and can be useful when dealing with minute samples.

Perspective

Research on proteomic reactors has benefited from the development of supporting materials, the integration of protein/peptide fractionations, the incorporation of multiple enzymes and coupling with other techniques. However, only a few reports have tested proteomic reactors for analyzing real biological samples including yeast, human cell lines and human embryonic stem cells. These initial reports are very promising with applications in hydrophobic membrane protein analysis [47], post-translational modification analyses (such as ubiquitination [10], glycosylation [14] and phosphorylation [11]), low-abundance protein identification [13] and protein identification and quantification in minute amounts of proteomic samples [15] (Fig. 2). We foresee that more integrated proteomic systems based on the proteomic reactor will be developed. The widespread application of proteomic reactors will likely require the development of commercial versions of these devices. The end goal is to develop systems in which the user only needs to introduce a cell/tissue lysate and all the steps for processing and analyzing the sample would be fully automated.

Figure 2.

 Summary of proteomic reactors. Either SAX or SCX beads (‘yin’ and ‘yang’) can be used in the proteomic reactor. The supporting material can be silica-based, polymer-based or monolithic. The proteomic reactor is coupled with protein fractionation (like size-exclusion chromatography or subcellular fractionation) or peptide fractionation (using step pH/salt gradients). Multiple enzymes (trypsin, chymotrypsin and Glu-C) are used in the device for protein digestion. The device was employed in post-translational modifications analysis (such as ubiquitination, glycosylation and phosphorylation), SILAC-base quantitation (dimethyl, 18O labellng, label-free quantitation will be further integrated into the device), minute sample amount protein analysis (rare cell proteomic reactor), low-abundance protein analysis and hydrophobic membrane protein identification.

Acknowledgements

DF would like to acknowledge a Canada Research Chair in Proteomics and Systems Biology. HZ and ZN would like to acknowledge the post-doctorate fellow award from the CIHR training program in neurodegenerative lipidomics. The authors would also like to acknowledge editorial help from Dr Ettore Appella and Dr Lisa Jenkins.

Advertisement