Microfluidic Flow Cell for Sequential Digestion of Immobilized

Article pubs.acs.org/ac

Microfluidic Flow Cell for Sequential Digestion of Immobilized Proteoliposomes Erik T. Jansson,† Carolina L. Trkulja,† Jessica Olofsson,†,¶ Maria Millingen,† Jennie Wikström,‡ Aldo Jesorka,† Anders Karlsson,‡ Roger Karlsson,‡ Max Davidson,‡ and Owe Orwar*,† †

Department of Chemical and Biological Engineering, Chalmers University of Technology, SE-412 96 Göteborg, Sweden Nanoxis AB, Lennart Torstenssonsgatan 5, SE-400 16 Göteborg, Sweden

‡

S Supporting Information *

ABSTRACT: We have developed a microfluidic flow cell where stepwise enzymatic digestion is performed on immobilized proteoliposomes and the resulting cleaved peptides are analyzed with liquid chromatography−tandem mass spectrometry (LC−MS/MS). The flow cell channels consist of two parallel gold surfaces mounted face to face with a thin spacer and feature an inlet and an outlet port. Proteoliposomes (50−150 nm in diameter) obtained from red blood cells (RBC), or Chinese hamster ovary (CHO) cells, were immobilized on the inside of the flow cell channel, thus forming a stationary phase of proteoliposomes. The rate of proteoliposome immobilization was determined using a quartz crystal microbalance with dissipation monitoring (QCM-D) which showed that 95% of the proteoliposomes bind within 5 min. The flow cell was found to bind a maximum of 1 μg proteoliposomes/cm2, and a minimum proteoliposome concentration required for saturation of the flow cell was determined to be 500 μg/mL. Atomic force microscopy (AFM) studies showed an even distribution of immobilized proteoliposomes on the surface. The liquid encapsulated between the surfaces has a large surface-tovolume ratio, providing rapid material transfer rates between the liquid phase and the stationary phase. We characterized the hydrodynamic properties of the flow cell, and the force acting on the proteoliposomes during flow cell operation was estimated to be in the range of 0.1−1 pN, too small to cause any proteoliposome deformation or rupture. A sequential proteolytic protocol, repeatedly exposing proteoliposomes to a digestive enzyme, trypsin, was developed and compared with a single-digest protocol. The sequential protocol was found to detect ∼65% more unique membrane-associated protein (p < 0.001, n = 6) based on peptide analysis with LC−MS/MS, compared to a single-digest protocol. Thus, the flow cell described herein is a suitable tool for shotgun proteomics on proteoliposomes, enabling more detailed characterization of complex protein samples.

T

typically used to search against a protein database, in order to obtain a list of proteins found in the sample.6,9−12 Proteomic studies of cytosolic proteins are rather straightforward and generally involve cell lysation and pelleting of membrane components followed by in-solution digestion of the supernatant. Accessing proteins associated with the lipid bilayers of the plasma membrane and organelles typically involves solubilization with detergents, organic solvents, or acids,11,13−16 which inevitably destroy their native structure. Avoiding protein denaturation is crucial for structure−function studies, e.g., in limited proteolysis,17,18 where the threedimensional structure and conformational changes of a protein can be probed and assessed via concentration- and time-limited enzymatic digestion of a protein followed by MS analysis of the resulting peptides. Here, we describe a microfluidic flow cell in which proteoliposomes directly derived from cultured cells are immobilized, thus creating a stationary phase of proteins associated with the membrane. The functionalization of the

here exist several analytical and biochemical methods used for handling and characterization of water-soluble proteins.1,2 Proteins associated with the lipid bilayers of the plasma membrane and organelles of the cell are, however, difficult to process and analyze, due to their strong hydrophobicity. This property also makes these proteins difficult to study in vitro without loss of structure or function. Only a limited number of membrane proteins have, to date, been characterized by X-ray crystallography, of which only a small fraction has been expressed in mammalian hosts.3 Mass spectrometry (MS) is a versatile tool for the identification of proteins and can be used for proteomic topdown and bottom-up analysis. The top-down approach is the suitable choice for the study of posttranslational modifications on single proteins, whereas the bottom-up approach handles protein identifications in high-complexity samples.4−8 For the identification of proteins with bottom-up proteomics, all proteins in a mixture are digested by a protease, where the subsequently created peptides are separated by one- or multidimensional liquid chromatography (LC), followed by analysis using tandem mass spectrometry (MS/MS). The complex spectral data set that results from this analysis is © 2012 American Chemical Society

Received: February 21, 2012 Accepted: May 30, 2012 Published: May 30, 2012 5582

dx.doi.org/10.1021/ac300519q | Anal. Chem. 2012, 84, 5582−5588

Analytical Chemistry

Article

Figure 1. (A) Overview of the workflow for digestion of proteoliposomes using the flow cell. (i) Membrane is extracted from cells, from which (ii) proteoliposomes are formed, followed by injection of the proteoliposome suspension into the flow cell channel. (iii) Proteoliposomes are immobilized on the inside of the flow cell channel, and a (iv) digestive enzyme is injected, commencing proteolysis of the proteins of the proteoliposomes. Finally, the peptides resulting from the digestion are eluted, while the proteoliposomes are still remaining immobilized in the flow cell channel. (B) The flow cell layout. Each flow cell contains six microfluidic channels; injection and elution can be facilitated with pipettes via the inlets and outlets. (C) Flow profile of the microfluidic channel. The inset depicts the diameter of a proteoliposome, which is 2 orders of magnitude smaller than the height of the channel. (D) Single-digestion protocol. Proteoliposome sample was incubated in one single channel. A solution of trypsin was injected, and after 120 min the generated peptides were eluted. (E) Sequential digestion protocol. Proteoliposome sample was incubated in one single channel. Solutions with trypsin and ambic were sequentially injected multiple times with increasing incubation time of trypsin; the injection of ambic was used as a washing step between trypsin incubations. Hence, a series of digested samples was collected over time from the same channel.

flow cell surfaces with proteoliposomes, in combination with a large area-to-volume ratio of the thin liquid film, provides fast transfer rates of chemicals to the proteoliposomes, which can be subjected to multiple rounds of solutions, without losing material or diluting the sample. Thus, the microfluidic platform can be used to label, modify, and digest membrane proteins in a sequential fashion in order to allow for dynamic modification of structure and function. Here, we demonstrate that, using sequential exposure to a digestive enzyme, a larger number of membrane-associated proteins can be detected compared to a single-exposure digestion.

listed is based on a calculation from manufacturing controlled adhesive coat weights using a density of 1.012 g/cm3. At each end of a channel, there is a port (inlet/outlet) for injecting and eluting fluids with, e.g., pipettes or tubing. The surface area exposed to the solution is 950 mm2, and the volume contained by each flow cell channel is ∼25 μL. The calculated volume based on the channel dimensions correlates well with experimentally determined channel volumes, by overfilling the channel and measuring the excess volume with manual pipetting. Loading of solutions is performed with injection of 100 μL of solution using a pipette at a volume flow of ca. 5 μL/ s, whereas washing steps are achieved with injection of 1 mL of solution at a volume flow of ca. 20 μL/s. The inlets and outlets of the flow cell are tapered; hence, ordinary pipette tips will seal when pressed down into the ports (Figure 1). Chemicals and Reagents. Cell culture medium (DMEM/ F12 with glutamine), fetal bovine serum, and Accutase were purchased from PAA. Zeocin was purchased from Invitrogen. The Pierce BCA protein assay kit was purchased from Thermo Scientific. Sequencing grade modified trypsin was purchased from Promega. All other chemicals were purchased from Sigma. Proteoliposome Preparation from RBCs. Human red blood cells (RBCs) were obtained from Sahlgrenska University

■

MATERIALS AND METHODS Flow Cell Design. The flow cell features six channels, each with the dimensions h = 50 μm, w = 5 mm, L = 95 mm, where h, w, and L are the height, width, and length of a channel, respectively, and consists of two closely spaced gold-layered polymer substrates [glycol-modified poly(ethylene terephthalate)], which are assembled using a double-adhesive tape where the geometrical shape of the channel has been punched out (Figure S-1 in the Supporting Information). The doubleadhesive tape (200 MP High Performance, 3M) is 50 μm thick, hence determining the height of the channel. The thickness 5583



Article

samples used for proteomic analysis, the total protein concentration was 0.8 mg/mL. Determination of Material Binding to Flow Cell Channel Surfaces. As a model sample for immobilization of proteoliposomes on the flow cell channel surfaces, 50−150 nm diameter proteoliposomes suspended in Tris buffer (10 mM Tris, 300 mM NaCl, pH 8) obtained from human RBCs by ultrasonication were used. The size distribution of the proteoliposomes had previously been determined by dynamic light scattering (data not shown). The amount of material adsorbed was determined by injecting a known concentration of proteoliposomes into the flow cell channel. Unbound proteoliposomes were eluted from the channel, and the absorbance at 214 nm was measured for both the original sample injected and the eluant. Atomic Force Microscopy. Atomic force microscopy (AFM) measurements were carried out on a PicoSPM instrument manufactured by Molecular Imaging, Inc. (Tempe, AZ, U.S.A.), equipped with a large area scanner (100 × 100 μm2). Sharpened, uncoated SiN cantilevers (MSCT-NONM) were obtained from Veeco (U.K.) and cleaned with SDS and Milli-Q water. A scanning probe image processor software (SPIP, Image Metrology, Denmark) was used to plane-fit and deconvolute the images. Quartz Crystal Microbalance with Dissipation Monitoring. Quartz crystal microbalance with dissipation monitoring (QCM-D) measurements were performed using a D300 from Q-Sense (Biolin Scientific AB, Västra Frölunda, Sweden) together with gold sensors, QSX 301 (Q-Sense). A shear oscillation is induced at the resonance frequency and at overtones 3, 5, and 7. The AT-cut crystal used had a fundamental resonance frequency of 4.95 MHz, and measurements were recorded at the third overtone (15 MHz). The instrument measures changes in frequency (f) and dissipation (D) at each overtone. Changes in frequency are mainly related to mass changes on the surface, whereas changes in dissipation are related to the viscoelastic properties of the adsorbed layer. For rigid films, the change in frequency for overtone n is proportional to the change in the total mass on the surface according to the Sauerbrey relation

Hospital, Göteborg, Sweden. The RBCs were washed several times with phosphate-buffered saline (140 mM NaCl, 3 mM KCl, 10 mM Na2HPO4, 2 mM KH2PO4, pH 7.4) to remove broken cells and leaked hemoglobin and then lysed with 5 mM Tris, 1 mM EDTA, pH 8 at 0 °C. The membranes were washed nine times with the lysis buffer to remove soluble protein. To remove peripheral membrane proteins, membranes were washed three times with 1 M KCl, one time with 10 mM NaOH, and finally one time with lysis buffer. Washed membranes (22 mg/mL) were stored at −80 °C until use. RBC membranes were thawed, diluted to 1 mg/mL (with 300 mM NaCl, 10 mM Tris, pH 8), and vesiculated by tipsonication using a Vibra Cell (model 501) from Sonics & Materials Inc. (Newtown, CT, U.S.A.). Proteoliposomes were created by sonication with 20% amplitude and a total sonication time of 10 min. The pulse time and the resting time were both set to 5 s in order to reduce heating by the probe. To further reduce heating of the sample the sonication took place in an ice-bath. Debris was removed by centrifugation (2000g, 5 min). Proteoliposome Preparation from CHO Cells. Adherent Chinese hamster ovary (CHO) cells with a tetracyclineregulated expression system (T-REx) were cultivated in medium (DMEM/F12 with glutamine) supplemented with 10% fetal bovine serum, zeocin (350 μg/mL), and blasticidin (5 μg/mL) in T175 or T500 culture flasks (Nunc). Then, 18−24 h before harvesting, the cells were incubated in medium (DMEM/F12 with glutamine) supplemented with 10% fetal bovine serum and doxycycline (1 μg/mL). Cells were harvested using Accutase and frozen in medium (DMEM/F12 with glutamine) supplemented with 10% fetal bovine serum and 10% DMSO at −80 °C until use. Harvested cells were thawed and washed three times with phosphate-buffered saline by centrifugation (600g, 3 min, 4 °C). The supernatant was removed, and the pellet was suspended in 10 volumes (1 volume = volume of the pellet) of lysis buffer (10 mM NaHCO3, 1 mM EDTA, 1 mM PMSF, pH 7.4) in order to swell the cells. After 10 min, the cells were homogenized using a hand-operated Dounce with a glass type B pestle (20 strokes). All procedures were performed at 0−4 °C. Lysed cell samples were centrifuged (600g, 3 min, 4 °C) to remove cell debris and nuclei. The supernatant was centrifuged (10 000g, 10 min, 4 °C) to remove mitochondria. The supernatant was aliquoted in 1.5 mL tubes, each supplied with 20 μL of 100 mM PMSF, snap-frozen in liquid N2, and stored at −80 °C until use. Membrane fragments from lysed cells were pelleted using an ultracentrifuge (100 000g, 60 min, 4 °C) and washed two times with washing buffer (300 mM NaCl, 10 mM Tris, pH 7.4) by centrifugation (100 000g, 60 min, 4 °C). For suspension of pellets, wide-bore pipette tips were used to avoid clogging. Membranes were resuspended in washing buffer in a coneshaped glass vial and vesiculated by tip-sonication three times (interspaced with 2 min of resting) using a Vibra Cell (model 501) from Sonics & Materials Inc. (Newtown, CT, U.S.A.). Proteoliposomes were created by sonication with 7% amplitude and a total sonication time of 1 min. The pulse time and the resting time were both set to 0.5 s in order to reduce heating by the probe. To further reduce heating of the sample the sonication took place in an ice-bath. Debris was removed by centrifugation (2000g, 3 min). The supernatant was snap-frozen in liquid N2 and stored at −80 °C until use. The total protein concentration in the proteoliposome samples was determined using a Pierce BCA protein assay kit. In the proteoliposome

Δm = ( −C Δf )/n

(1)

where C is the sensitivity of the used quartz (C = 17.7 ng/ cm2·Hz).19 All experiments were performed in batch mode at 20 °C. Digestion Protocols for Protein Detection. CHO proteoliposome sample (50 μL) was injected into each channel of a flow cell and was incubated at room temperature for ∼1 h. Unbound material was washed out with 1 mL of washing buffer (300 mM NaCl, 10 mM Tris, pH 7.4), followed by 1 mL of ambic (20 mM NH4HCO3, pH 8). Digestion was initiated by injecting 100 μL of trypsin (5 μg/mL in ambic), and the chip was incubated in room temperature for a certain time period (cf. below). Peptides were then eluted by injecting 100 μL of ambic into the inlet and collecting the eluate at the outlet (see Figure 1 for an overview of the workflow) and were digested overnight in an Eppendorf tube at room temperature and thereafter frozen at −20 °C. Sequential and single digestions were performed at room temperature, where a single flow cell lane was required for each type of digest. The single digestions were executed as follows: one channel in a flow cell was loaded with proteoliposome sample and washed. A trypsin solution (100 μL, 5 μg/mL) was 5584



Article

by searching the reversed database. A false discovery rate of 0.01 on peptide level and two unique peptides were required for protein identification. Detected proteins were classified as membrane-associated, transmembrane, lipid anchored, peripheral, or nonmembraneassociated (see the Supporting Information for definitions). It was possible for a single protein to have multiple classifications. The classes transmembrane, lipid anchored, and peripheral are regarded as subclasses of the membrane-associated class. Full lists of search terms for protein classification are found in the Supporting Information. For determination of statistical significance, Student’s unpaired t test was applied. Data is presented as mean plus/ minus one standard deviation.

injected in the channel, and digestion was allowed to proceed for or 120 min (Figure 1). Samples were run in six replication series and eluted with 100 μL of ambic, incubated overnight, and thereafter frozen (cf. above). Sequential digestions were executed as follows: one channel in a flow cell was loaded with proteoliposome sample and washed. A trypsin solution (100 μL, 5 μg/mL) was injected, and the immediate eluate was collected as a 0 min sample. The flow cell was further incubated for 5 min. Ambic (100 μL) was injected, and the eluate was collected as a 5 min sample. The channel was washed with 1 mL of ambic in order to flush the flow cell from residual peptides. Again, the channel was injected with 100 μL of a 5 μg/mL trypsin solution and incubated for 10 min. Ambic (100 μL) was injected, and the eluate was collected as a 15 min sample. From here, injection of trypsin, elution, and wash were carried out as above, with increasing incubation times, yielding a set of samples with cumulative incubation times of 0, 5, 15, 30, 60, and 120 min, respectively (Figure 1). These digestions were also run in six replication series, incubated overnight, and thereafter frozen (cf. above). Liquid Chromatography with Tandem Mass Spectrometry. Peptide samples from digestions in flow cells of CHO proteoliposomes were analyzed by the Proteomics Core Facility at the University of Gothenburg, Göteborg, Sweden. Samples were vacuum-centrifuged to dryness and reconstituted in 18 μL of 0.1% formic acid in water. Each sample was centrifuged for 15 min at 13 000g, and 17 μL was transferred to a sample vial. A 3 μL aliquot of this sample was diluted 25 times and transferred to an autosampler vial of the liquid chromatography with tandem mass spectrometry (LC−MS/ MS) system. The separation was performed on an EasyLC system (Thermo Scientific), and the tryptic peptides were separated on a 200 mm × 0.05 mm i.d. fused-silica column packed inhouse with 3 μm ReproSil-Pur C18 AQ particles (Dr. Maisch, GmbH, Ammerbuch, Germany). A 2 μL aliquot of the sample was injected, and the peptides were first trapped on a precolumn (45 × 0.075 mm i.d.) packed with 3 μm C18bonded particles. An 80 min gradient of 5−35% acetonitrile in 0.2% formic acid followed by a 10 min gradient of 35−80% acetonitrile was then applied with a flow rate of 200 nL/min. Mass analyses were performed in a Linear Trap Quadrupole− Orbitrap Velos mass spectrometer (Thermo Scientific) equipped with an in-house modified nanospray ion source. The instrument was operated in the data-dependent mode to automatically switch between MS and MS/MS acquisition. MS spectra were acquired in the Orbitrap Velos, whereas MS/MS spectra were acquired in the LTQ-Velos. For each MS scan, the 10 most intense, doubly, triply, or quadruply charged ions were sequentially fragmented in the linear trap by collision-induced dissociation (CID). Fragmented target ions were excluded for MS/MS selection for 120 s after two repeated MS/MS. Thermo Proteome Discoverer v. 1.3 (Thermo Scientific) was used to validate MS/MS based peptide and protein identifications. All tandem mass spectra were searched by MASCOT (Matrix Science, London, U.K.) against UniProtKB release 2012_04, (Chinese hamster [Cricetulus griseus]; SwissProt 231 sequences, TrEMBL 24 188 sequences). MASCOT was set to search the database assuming trypsin as digestion enzyme, allowing for one missed cleavage. The fragment ion mass tolerance was set to 0.5 Da, and the parent ion tolerance was set to 5 ppm. Oxidation of methionine was specified as a variable modification. The false discovery rate was determined

■

RESULTS AND DISCUSSION The workflow for utilizing the flow cell consists of four steps (Figure 1), where the two first steps consist of membrane extraction and proteoliposome preparation from biological samples, the third step deploys immobilized proteoliposomes on to the flow cell channel surface, and finally the fourth step allows for various application-dependent protocols for performing chemistry on proteoliposomes. In this study, we specifically investigated the physical properties of the flow cell. This included a hydrodynamic characterization as well as studies on the rate and extent of immobilization of proteoliposomes. Furthermore, sequential digestion on a complex proteomic sample was performed and demonstrated to identify an increased number of membraneassociated proteins in LC−MS/MS analysis compared with a single-digest protocol. It is believed that multiple rounds of protease exposure, and wash-out, result in an altered membrane protein structure that makes the protease capable to access cleavage sites of the protein previously hidden, as well as the inherent peptide fractionation which creates less complex sample matrixes amenable for MS detection. Fluid Dynamics of the Flow Cell. In order to estimate the forces acting on the immobilized proteoliposomes during flow cell operation, the flow profile in the flow cell channel needs to be derived. The Reynolds number is defined as Re = ρvLC/μ, where ρ is the density of the fluid, v is the flow velocity, LC is the characteristic length (for a wide duct, the characteristic length equals twice the height of the duct, LC = 2h = 100 μm), and μ is the viscosity of the fluid. Assuming that the fluid has water-like properties (ρ = 1 kg/m3, μ = 1 mPa·s), and a volume flow QF of 20 μL/s during regular use (which corresponds to an average velocity v ̅ of 0.08 m/s), we yield a Re of 0.008; thus, the flow in the channel is laminar, far below the limit for turbulent flow. The width w = 5 mm of the channel is 2 orders of magnitude larger than the height h = 50 μm of the channel; therefore, we approximate the flow in the channel with a two-dimensional parabolic flow. The flow profile can then for a given volume flow Q be expressed as 6Q vx(z) = 3 (h − z)z (2) hw where vx(z) is the flow velocity as a function of height and x and z are the length and height coordinates, respectively (Figure 1; a full derivation of eq 2 is given in the Supporting Information). With the above given height and width of the flow cell channel, a volume flow QF = 20 μL/s yields a maximum velocity of 0.12 m/s at midheight of the channel. 5585



Article

Immobilization of Proteoliposomes on Flow Cell Channel Surfaces. For the estimation of the magnitude of forces acting on the surface-adhered proteoliposomes, two cases were considered: a complete coverage of proteoliposomes on the surface, or a single proteoliposome adhered to the surface. The shear stress in a Newtonian fluid in laminar flow is directly proportional to the velocity gradient and is for a flow profile with velocity variations only in the z-direction defined as τ (z ) = μ

dv dz

Hence, the magnitude of the force acting on a proteoliposome during operation in the flow cell is in the range of 0.1−1 pN. The large surface-to-volume ratio of the liquid film provides fast transport rates of proteoliposomes to the stationary surface; as measured by QCM-D on a flow cell equivalent gold surface, 95% of the material was attached to the surface within 5 min, and 1 h after the addition of proteoliposomes, they had not detached from the surface in the flow of the QCM-D (Figure 2). Immobilization studies with injection and elution of RBC proteoliposomes indicated that the maximal amount of material that can be bound to the surfaces of the flow cell in its present design is on the order of 1 μg/cm2 (Table 1). A concentration

(3)

where τ is the shear stress and μ is the viscosity of the liquid. By combining eqs 2 and 3, we can estimate the shear force on an area similar to that exposed to the flow for a proteoliposome at full surface coverage. Assuming the proteoliposomes to have a radius r = 50 nm and an exposed area to the flow on the order of 10−14 m2 at a height z = 100 nm, gives that, for a typical volume flow QF = 20 μL/s, the force per proteoliposome resulting from shear stress is on the order of 0.1 pN. For a single spherical proteoliposome on a surface, the drag force FD is derived from Stokes’ law as FD ≈ 1.7(6πμrv )̃

Table 1. Quantitation of RBC Proteoliposomes Immobilized in the Flow Cell Channel, with Varying Deposition Time and Proteoliposome Concentrationa concn (μg/mL)

deposition time (h)

material bound (μg/cm2)

bound/injected material (mass/mass %)

1000

1 3 16 1 3 16 1 3 16

1.05 1.11 0.97 1.17 1.06 0.91 0.78 0.65 0.61

31 33 29 64 64 54 90 76 73

500

(4)

where r is the radius of the proteoliposome, μ is the viscosity of the liquid, and ṽ is the velocity of the shear flow at the particle center for undisturbed shear flow.20 As the height of the channel h compared to the height of the proteoliposomes hp differs by 2 orders of magnitude (hp ≪ h), the velocity gradient is considered to be uniform over the proteoliposomes. The drag force experienced by the proteoliposome is according to eq 4 proportional to the volume flow Q and is evaluated to be 1 pN for a typical volume flow QF. For comparison, the force required to pull a lipid tube from a vesicle membrane is on the order of 10 pN.21 The coverage of the RBC proteoliposome sample was shown to be evenly distributed with tapping-mode AFM on gold equivalent to the flow cell surfaces (Figure 2), as well as

250

Vesicle suspensions of varying concentrations were added to the flow cell. The flow cell was eluted with buffer after varying immobilization times, and the eluate with nonbound vesicles was collected. The absorbance at 214 nm was measured for the collected eluates and correlated to a standard curve performed with varying concentrations of vesicles. a

of membrane preparation for immobilization of at least 500 μg/ mL was chosen for the remaining experiments, as this minimum concentration allows for saturation of the flow cell with proteoliposomes. When using 250 μg/mL as immobilization concentration, the surface binds less membrane material and may thus not be as covered as at the conditions used in this work. In addition to RBCs, proteoliposomes derived from bacteria, mammalian cell lines, and other human cell types have been investigated and found to immobilize equally well as the RBC preparations.22−26 Sequential and Single Tryptic Digestion Protocols. Tryptic digestions were performed with the sequential and single-digestions protocols, for 120 min of total digestion time for both protocols. Performing tryptic digestions of immobilized CHO proteoliposomes with the sequential protocol yielded ∼65% more unique membrane-associated protein, and ∼43% more unique nonmembrane-associated protein (p < 0.001, n = 6), compared to the single-digestion protocol (Table 2). Protein identifications in the subclasses of membraneassociated proteins were also significantly increased by 63− 120% with the sequential protocol compared to single digestions (p < 0.001, n = 6 for transmembrane and peripheral protein, and p < 0.01, n = 6 for lipid anchored protein). The increase of identified membrane-associated proteins was larger and statistically significant compared to the increase of nonmembrane-associated proteins (p < 0.001, n = 6), showing increased identification of membrane-associated protein to be

Figure 2. Functionalization of gold surfaces equivalent to the flow cell surfaces with RBC proteoliposomes. (A and B) Tapping mode AFM image (area 2 × 2 μm2) showing the surface (A) before and (B) after addition of proteoliposomes. In panel B, evenly distributed, immobilized proteoliposomes are detected. (C) QCM-D trace showing adsorption of the RBC proteoliposomes over time on a gold surface equivalent to the flow cell surfaces; the arrow denotes addition of proteoliposomes to the assay; mass gain is observed to be stable 1 h after addition of proteoliposomes.

experiments where the immobilized proteoliposomes were stained with fluorescent probes and visualized with a fluorescence microscope (data not shown). An unsaturated surface area coverage means that the fluid is exposed to a slightly rugged surface rather than a surface with solitary proteoliposomes placed at a large distance from each other. 5586



5587

6.6 5.9 3.7 3.1 3.4 2.3 1.4 0.75 1.0 0.52 ± ± ± ± ± ± ± ± ± ± 175.3 115.3 61.5 27.3 13.3 5.3 2.0 0.83 0.7 0.33 9.3 12 9.9 2.7 2.3 2.3 0.75 0.82 0.98 0.52 ± ± ± ± ± ± ± ± ± ± 223.2 168 88.7 47.3 27.5 12.7 4.17 0.33 1.17 0.67 1.6 2.6 2.1 1.5 0.41 ± ± ± ± ± 10.7 12.0 6.5 1.8 0.17 2.3 3.1 3.7 2.0 0.55 ± ± ± ± ± 0.17 ± 0.41

17.2 19.5 11.7 3.3 0.50 1.2 ± 1.2 0.67 ± 0.82

a

A single, 120 min digestion is compared with a cumulative incubation time of 120 min with trypsin, n = 6.

0.41 0.52 0.52 0.41 ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 0−10 10−20 20−30 30−40 40−50 50−60 60−70 70−80 80−90 90−100

peripheral

sequential single

lipid-anchored

sequential single

transmembrane

sequential single

membrane-associated

sequential

CONCLUSIONS We have here presented a microfluidic flow cell, which enables flexible chemistry to be performed on proteoliposomes directly derived from cultured cells. The flow conditions used do not rupture or remove the immobilized vesicles. Illustrated here, several chemical treatments on membrane protein samples are enabled by the immobilization approach. We show that several sequential proteolytic treatments of the same immobilized sample may give an increase in the number of identified membrane-associated proteins, compared to a single digestion. The sequential chemical reaction mode is naturally amenable to the use of different proteases to further optimize digestions of membrane protein samples.

sequence coverage (%)

■

no. of identified proteins

Table 3. Single Protein Sequence Coverage for Sequential and Single Tryptic Digestions of Immobilized CHO Proteoliposomesa

favored over nonmembrane-associated proteins with sequential digestion. The amino acid sequence coverage of proteins was compared between sequential and single tryptic digestion of immobilized CHO proteoliposomes. The distribution resulting from binning the sequence coverage on a single-protein level was shown to be similar for the two protocols for all classes of protein (Table 3). The only difference observed in the tabulated histogram was in amplitudes for the binned data, reflecting the increased count of identified proteins with sequential digestion in Table 2. Hence, the benefit of the sequential digestion technique presented here is based on the improved amount of identified membrane-associated protein. The generation of many consecutive samples eluted from a single flow cell lane is simple and quick and does not, compared with a test tube digestion, require any intermediate centrifugation or resuspension steps. Such washing steps of the small amount of proteoliposomes used herein, without the aid of immobilization, would result in loss of sample between digestions. The sequential digestion protocol allows a single injection of biological sample to generate more peptide samples for LC− MS/MS analysis compared with the single-digestion protocol, as the immobilization of proteoliposomes on the flow cell surface allows per se for many digestions with washes in between. With sequential digestion, certain peptides of a protein, which initially are more readily accessible for digestion than others, could be cleaved off and eluted at an early stage, which would contribute to the fractionation of peptides from the digested proteoliposome sample. We believe that this step of peptide separation provided by multiple digestions on a stationary phase of proteoliposomes, in combination with the generation of many peptide samples, improves the detection of peptides for bottom-up protein identification.

0.83 2.67 0.67 0.17

0.094 0.11 1.8 0.20 0.061

2.4 1.8 1.7 1.0 0.41 0.52

± ± ± ± ±

± ± ± ± ± ±

1.651 1.63 2.2 1.67 1.427

23.5 11.7 5.8 1.3 0.17 0.33

2.8 2.2 1.7 3.1 12

3.1 2.6 3.0 1.5 0.55 0.52

± ± ± ± ±

± ± ± ± ± ±

74.0 42.8 2.0 31.2 402

38.5 18.8 9.7 1.8 0.50 0.67

5.2 3.0 0.52 3.5 18

0.98 3.0 3.2 1.9 0.52 0.52

± ± ± ± ±

ratio sequential/ single

34.17 23.7 12.3 3.2 0.33 0.33

122.2 70.0 4.33 52.2 574

single

single

membrane-associated transmembrane lipid-anchored peripheral nonmembraneassociated

sequential

3.8 3.8 5.8 3.4 0.89 0.52

protein classification

55.7 38.3 21.3 5.2 1.00 0.67

nonmembrane-associated

no. of identified proteins

sequential

Table 2. Proteins Identified with Sequential and Single Tryptic Digestions of Immobilized CHO Proteoliposomes, Sorted by Protein Classification, n = 6

single

Article



■

Article

(15) Goshe, M. B.; Blonder, J.; Smith, R. D. J. Proteome Res. 2003, 2, 153−161. (16) Weekes, M. P.; Antrobus, R.; Lill, J. R.; Duncan, L. M.; Hör, S.; Lehner, P. J. J. Biomol. Tech. 2010, 21, 108−115. (17) Fontana, A.; de Laureto, P. P.; Spolaore, B.; Frare, E.; Picotti, P.; Zambonin, M. Acta Biochim. Pol. 2004, 51, 299−321. (18) Hubbard, S. J. Biochim. Biophys. Acta 1998, 1382, 191−206. (19) Sauerbrey, G. Z. Phys. A: Hadrons Nucl. 1959, 155, 206−222. (20) O’Neill, M. E. Chem. Eng. Sci. 1968, 23, 1293−1298. (21) Derényi, I.; Koster, G.; van Duijn, M.; Czövek, A.; Dogterom, M.; Prost, J. In Controlled Nanoscale Motion; Linke, H., Månsson, A., Eds.; Lecture Notes in Physics; Springer: Berlin/Heidelberg, Germany, 2007; Vol. 711; pp 141−159. (22) Bauer, B.; Davidson, M.; Orwar, O. Angew. Chem., Int. Ed. 2009, 48, 1656−1659. (23) Chooneea, D.; Karlsson, R.; Encheva, V.; Arnold, C.; Appleton, H.; Shah, H. BMC Microbiol. 2010, 10, 44. (24) Hansson, S. F.; Henriksson, Å.; Johansson, L.; Korsgren, O.; Eriksson, J. W.; Tornqvist, H.; Davidsson, P. Clin. Proteomics 2010, 6, 195−207. (25) Padliya, N. D.; Bhatia, M. B.; Hofgärtner, W. T.; Hariri, R. J. Anal. Methods 2010, 2, 539−545. (26) Karlsson, R.; Davidson, M.; Svensson-Stadler, L.; Karlsson, A.; Olesen, K.; Carlsohn, E.; Moore, E. R. B. J. Proteome Res. 2012, 11, 2710−2720.

ASSOCIATED CONTENT

S Supporting Information *

A description of the punched-out tape used for flow cell assembly, a derivation of the equation describing the flow profile in the flow cell channel, protein classification search terms, and a summary of identified proteins with sequential and single digestions. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Present Address ¶

Department of Pathology, Stanford University School of Medicine, Stanford, CA 94305−5324.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the Knut and Alice Wallenberg Foundation, the European Research Council (ERC) through an Advanced ERC Grant, and the Swedish Foundation for Strategic Research (SSF). We thank Prof. Johan Bergenholz at University of Gothenburg for help with dynamic light scattering measurements, Prof. Michael Zäch at Chalmers University of Technology for help with AFM measurements, and the staff at the Proteomic Core Facility at University of Gothenburg for valuable discussions and comments on the manuscript.

■

REFERENCES

(1) Cutler, P. Proteomics 2003, 3, 3−18. (2) Structural Genomics Consortium, Architecture et Fonction des Macromolécules Biologiques, Berkeley Structural Genomics Center, China Structural Genomics Consortium, Integrated Center for Structure and Function Innovation, Israel Structural Proteomics Center, Joint Center for Structural Genomics, Midwest Center for Structural Genomics, New York Structural GenomiX Research Center for Structural Genomics, Northeast Structural Genomics Consortium, Oxford Protein Production Facility, Protein Sample Production Facility, Max Delbrück Center for Molecular Medicine, RIKEN Structural Genomics/Proteomics Initiative, and SPINE2-Complexes. Nat. Methods 2008, 5, 135−146. (3) Bill, R. M.; Henderson, P. J. F.; Iwata, S.; Kunji, E. R. S.; Michel, H.; Neutze, R.; Newstead, S.; Poolman, B.; Tate, C. G.; Vogel, H. Nat. Biotechnol. 2011, 29, 335−340. (4) Mann, M.; Wilm, M. Anal. Chem. 1994, 66, 4390−4399. (5) Han, X.; Jin, M.; Breuker, K.; McLafferty, F. W. Science 2006, 314, 109−112. (6) Dowell, J. A.; Frost, D. C.; Zhang, J.; Li, L. Anal. Chem. 2008, 80, 6715−6723. (7) Tang, J.; Gao, M.; Deng, C.; Zhang, X. J. Chromatogr., B: Anal. Technol. Biomed. Life Sci. 2008, 866, 123−132. (8) Yates, J. R.; Ruse, C. I.; Nakorchevsky, A. Annu. Rev. Biomed. Eng. 2009, 11, 49−79. (9) Aebersold, R.; Goodlett, D. R. Chem. Rev. 2001, 101, 269−295. (10) Aebersold, R.; Mann, M. Nature 2003, 422, 198−207. (11) Washburn, M. P.; Wolters, D.; Yates, J. R. Nat. Biotechnol. 2001, 19, 242−247. (12) Steen, H.; Mann, M. Nat. Rev. Mol. Cell Biol. 2004, 5, 699−711. (13) Han, D. K.; Eng, J.; Zhou, H.; Aebersold, R. Nat. Biotechnol. 2001, 19, 946−951. (14) Blonder, J.; Goshe, M. B.; Moore, R. J.; Pasa-Tolic, L.; Masselon, C. D.; Lipton, M. S.; Smith, R. D. J. Proteome Res. 2002, 1, 351−360. 5588


Microfluidic Flow Cell for Sequential Digestion of Immobilized

Recommend Documents