Thiol-ene monolithic pepsin microreactor with a 3D ... - ACS Publications

This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Article pubs.acs.org/ac

Thiol-ene Monolithic Pepsin Microreactor with a 3D-Printed Interface for Efficient UPLC-MS Peptide Mapping Analyses Alexander Jönsson,† Rasmus R. Svejdal,† Nanna Bøgelund, Tam T. T. N. Nguyen, Henrik Flindt, Jörg P. Kutter, Kasper D. Rand,* and Josiane P. Lafleur*,‡ Department of Pharmacy, Copenhagen University, Universitetsparken 2, Copenhagen E DK-2100, Denmark S Supporting Information *

ABSTRACT: To improve the sample handling, and reduce cost and preparation time, of peptide mapping LC-MS workflows in protein analytical research, we here investigate the possibility of replacing conventional enzymatic digestion methods with a polymer microfluidic chip based enzyme reactor. Off-stoichiometric thiol-ene is utilized as both bulk material and as a monolithic stationary phase for immobilization of the proteolytic enzyme pepsin. The digestion efficiency of the, thiol-ene based, immobilized enzyme reactor (IMER) is compared to that of a conventional, agarose packed bed, pepsin IMER column commonly used in LC-MS based protein analyses. The chip IMER is found to rival the conventional column in terms of digestion efficiency at comparable residence time and, using a 3Dprinted interface, be directly interfaceable with LC-MS. n the field of protein analytical chemistry, one of the most important, and often unavoidable, steps is the cleavage of proteins into peptides.1 Protein cleavage, or digestion, is most commonly achieved through the use of a proteolytic enzyme (protease). The protein sample and protease are mixed directly and incubated over a considerable amount of time, often overnight,2 in order to produce peptides of significant abundance to allow identification by liquid chromatography (LC) and mass spectrometry (MS). This digestion procedure can lead to a contamination of the sample due to autodigested fragments from the enzyme used. To avoid contamination, the enzyme can be immobilized on a solid support.3 Immobilized enzyme reactors (IMERs) typically consist of a small diameter, microbore, column and can be packed with an enzyme functionalized solid support, e.g., beads, that act as a stationary phase. The IMER is most commonly coupled to a pump system that controls the flow of sample across the enclosed solid support.4 IMERs not only provide online digestion without sample contamination and minimal autodigestion of the proteolytic enzyme but also significantly reduce the sample digestion time.5 The proteolytic IMERs have become a common feature of various so-called bottom-up LC-MS proteomic workflows,6−8 that benefit greatly from fast online digestion, including hydrogen/deuterium exchange mass spectrometry (HDX-MS).9−11 Challenges, such as handling minute sample amounts and optimizing the digestion reaction, however still need to be addressed. It has been shown that miniaturization of the IMER allows for smaller sample volumes and improvements in proteolytic efficiency.12−14 The matter of miniaturization can

I

© 2017 American Chemical Society

be tackled in a number of different ways depending on the intended application.15 Microfluidic chip based IMERs are one such solution that has the benefit of versatility in terms of both material16 and design.12,17 The conventional use of photolithography and wet etching techniques for making microfluidic devices out of glass is expensive and labor-intensive and requires several steps to achieve biofunctionalization.18−20 Polymeric materials based on thiol-ene (TE) photochemistry alleviate these issues, making them an attractive alternative for the manufacture of microfluidic chips.21−27 One of the main advantages of TE resins prepared in-house is that variable stoichiometric ratios can be used to tailor the presence of either sulfhydryl (“thiol”) or allyl (“ene”) functional groups on the channel walls or solid support surface. These “off-stoichiometric” TEs (OSTEs)28,29 are a relatively new addition to the field of microfluidics and make it possible to use the wellknown photoinitiated thiol-ene “click-chemistry” reaction to further modify the generated surface through covalent attachment of biomolecules.30−32 We have recently demonstrated the immobilization of enzymes on ene-excess OSTE via ascorbic acid linkage.32 Conversely, thiol-excess OSTE has also been shown to provide an excellent support for enzyme immobilization through linkage with maleic acid.33 Manual packing of columns with stationary phase materials, for the generation of miniaturized on-chip IMERs, is not easily adapted to microfluidic dimensions, often requiring retaining Received: December 24, 2016 Accepted: March 21, 2017 Published: March 21, 2017 4573

DOI: 10.1021/acs.analchem.6b05103 Anal. Chem. 2017, 89, 4573−4580

Analytical Chemistry

■

frits, which are challenging to make, and often result in dead volumes or nonuniform packing.34 Porous monoliths are an attractive alternative to these packed beds and can be readily manufactured from polymeric resins.35,36 While the use of polymeric monoliths is well established,37 a number of challenges still remain. Shrinkage upon polymerization, in combination with poor adhesion to the channel walls, can lead to the monolith detaching, creating large dead volumes and poor reproducibility. To minimize these risks, it is preferable if the monolith can be covalently bound to the channel itself.38 TE exhibits low shrinkage upon polymerization39 and can adhere to itself, making it a suitable material for in situ generation of monolithic solid supports inside TE microchannels, providing uniformly distributed and well anchored monoliths.32 Trypsin and pepsin are two proteases used in bottom-up protein analysis workflows.10,40 Trypsin provides a specific cleavage at the carboxyl side of Lys/Arg residues and is mainly compatible with a sample pH between 7 and 9, whereas pepsin provides a largely unspecific cleavage and is compatible with sample pH < 3. The use of these particular proteases in IMERs is in part due to the ease of obtaining both in large quantities. Immobilization of a protease onto a packed column can require as much as 60 mg of enzyme, complicating the use of many, rare and/or expensive, proteases.41−44 Because of this, immobilization techniques and devices compatible with low amounts of enzyme are continuously sought.41,45 Microfluidic IMERs are inherently suitable for use with low abundance proteases due to their small internal volumes. A challenge of chip based IMERs, and microfluidic chips in general, is the “chip-to-world” interface, i.e., the connection to external systems such as pumps and detectors.46,47 Interfacing of the widely used poly(dimethylsiloxane) (PDMS) chips48 often relies on PDMS’s inherent sealing capabilities when simply inserting connecting tubes directly into the chip. This does not provide high resistance to pressure and is prone to leakage, which limits the use of PDMS in high pressure environments such as a direct interface with UPLC-MS.49 Other, more universal solutions include the use of glued-on sample ports50 or simply gluing the tubing directly to the chip.51,52 These permanent modifications run the risk of clogging the inlet channels during attachment and often introduce permanent dead volumes.47 Special commercially available holders53 could potentially alleviate the problems associated with both glued on and direct insert inlets, but their high price, combined with the limitations they put on the device design, restricts their use. The increased availability of 3D-printers allows for the fabrication of customized interfaces to match individual chip designs.54,55 An example of this is the 3D-printed, click-on, chip interconnect developed by Paydar et al.,56 which can withstand pressures up to 400 kPa (58 psi). As previously mentioned, OSTE, chip based and monolithic, IMERs featuring galactose oxidase and PNGase F have recently been demonstrated.32 In this contribution, we improve the design of these microfluidic reactors to facilitate ease of production and develop a chip based pepsin IMER for use in proteomics type applications. Replica molding together with a versatile immobilization procedure provides an inexpensive and disposable alternative to commonly used packed-bed pepsin columns. We demonstrate the potential of direct integration of the microreactor in a typical proteomics workflow by direct coupling to UPLC-MS through a custom-designed 3D-printed interface with sufficient pressure tolerance.

Article

EXPERIMENTAL SECTION

Materials and Reagents. Pentaerythritol-tetrakis(3-mercaptopropionate) (“tetrathiol”), triallyl-1,3,5-triazine-2,4,6(1H,3H,5H)-trione (“triallyl”), 2-(boc amino)ethanethiol, Lascorbic acid (ASA), pepsin from porcine gastric mucosa (lyophilized powder ≥2500 U/mg), hemoglobin (human, lyophilized powder), formic acid, trifluoroacetic acid (TFA) (CHROMASOLV ≥ 99.0%), hydrochloric acid (HCl) (reagent grade, 37%), and methanol (99.8%) were all obtained from Sigma-Aldrich (Brøndby, Denmark). Trichloroacetic acid (TCA) (EMSURE ACS, Reag.) was obtained from Merck (Darmstadt, Germany). Immobilized pepsin (on agarose resin) was obtained from Thermo Fischer Scientific (Waltham, MA, USA). The Sylgard 184−poly(dimethylsiloxane) (PDMS) elastomer kit was obtained from Dow Corning (Midland, MI, USA). Acetonitrile (ACN) (LiChrosolv) was obtained from VWR (Søborg, Denmark). Lucirin TPO-L was obtained from BASF (Ludvigshafen, Germany). Fabrication of Microfluidic Chips. The two-step replica molding process used to fabricate the TE microfluidic chips has been described in detail elsewhere.32 Briefly, master designs (channel dimensions: 10 mm long × 300 μm deep × 500 μm wide with a tapering to 200 μm over 1 mm at the outlet; channel volume: 1.60 μL) were drawn with computer-aideddesign software (Autodesk Inventor Professional 2015, San Rafael, CA, USA) and patterned onto a poly(methyl methacrylate) (PMMA) plate by high precision milling (Minitech 3, Minitech Machinery Corp., Norcross, GA, USA). PDMS molds were prepared from these PMMA masters and cured at 80 °C for 2 h. A stoichiometric mixture of tetrathiol and triallyl monomers was poured into the PDMS molds and cured under a UV flood light (15 s each side, 160 mW cm−2 at 365 nm, Dymax EC 5000 Series UV curing flood lamp, Dymax Corp, Torrington, CT, USA). Complementary pieces were manually pressed into conformal contact immediately after demolding and further exposed to the UV flood light for 1 min. Assembled chips were placed in an oven at 60 °C overnight under a weight and allowed to cool before use. Monolith Preparation. An emulsion of OSTE (1:1.4 ratio tetrathiol/triallyl) in methanol (80% w/w) was magnetically stirred at 1500 rpm in a sealed container for 5 min. Photoinitiator (10% v/v Lucirin TPO-L in ethanol) was added to the emulsion to a final concentration of 0.05% v/v and stirred for 1 min. The emulsion was directly injected into the TE microfluidic chip with the use of a pipet and immediately cured under collimated UV light (30 s, 20.5 mW cm−2 at 365 nm, LS-100-3C2 near UV light source, Bachur & Associates, Santa Clara, CA, USA). This resulted in a monolithic structure filling the entire channel (Figure S-1). The monolith was rinsed thoroughly with distilled deionized water (DDW) using a syringe pump (5 min at 30 μL/min), and the chip was sealed for storage to prevent drying. Enzyme Immobilization. A technique for immobilization of enzymes on OSTE has previously been described by Lafleur et al.;32 the same method was adapted here for the immobilization of pepsin. To introduce an amino functionality at the surface of the allyl-excess monolith, the chip was filled with 2-(boc amino)ethanethiol containing 5% v/v photoinitiator (lucirin TPO-L) and exposed to collimated UV light (30 s, 20.5 mW cm−2 at 365 nm). Unreacted products were washed out with water (5 min, 10 μL/min). Deprotection with 4574


Article


For online digestion at 13 μL/min, the chip-to-world interface was connected in-line with the G2-Si UPLC-MS system and 100 μL of 0.04 μM hemoglobin was loaded. A schematic of the setup can be found in Figure S-3. The reference column IMER was also mounted in-line with the G2Si UPLC-MS system. Trapping of peptides on the UPLC system was done on a Waters 2.1 mm × 5 mm VanGuard 1.7 μm BEH C18 precolumn. Elution was done at a flow of 40 μL/ min, using a gradient of 8% to 40% mobile phase B (ACN with 0.23% formic acid). Chromatographic separation was performed using a 1 mm × 100 mm Waters ACQUITY UPLC 1.7 μm BEH C18 reverse-phase column. The mass spectrometers were operated in positive ion mode. Data-independent MS/MS was performed using the Waters MSe function from 50 to 2000 m/z. Data Processing. Raw files generated by the mass spectrometer were loaded into ProteinLynx Global Server (PLGS) for peptide mapping based on MSe (MS/MS) data and the known sequence of the protein analyzed. The assigned peptides were filtered using DynamX 3.0 software (Waters, Milford, MA, USA) with the following criteria: minimum 2 product ions, 0.2 minimum product ions per amino acid, 10 ppm mass error on precursor ion, and a given peptide must be identified in a minimum of two out of three separate LC-MS/ MS experiments. Determination of Device Void Volume. Water containing food dye was pumped through the devices at a flow of 10 μL/min. For the chip IMER, the permeation time from inlet to outlet was measured visually through the chip. The flow was measured using a FLOWELL flow meter (Fluigent, Villejuif, France). The column IMER void volume was determined between marked points on the tubing before and after the column with the internal volume of the tubing subtracted. Comparison of the void volumes of devices with and without monolithic stationary phase yields the packing density of the finished device. Determination of Solid Phase Surface Area. For the chip based reactor, the total surface area can be estimated from the monolith volume using a density of 1.218 mg/μL and a surface area of 21 cm2/mg.32 The agarose gel in the packed column can be approximated as a single continuous fiber57 with a radius of 5.9 nm and a fiber ratio of 0.2 m3fiber/m3gel.58 Enzyme Activity. A modified version of the Anson protocol59 was used to determine the activity of the IMERs. As reference, a digestion was done using free pepsin. Briefly, 200 μL of 20 mg/mL hemoglobin (in water acidified to pH 2.0 with HCl) was mixed with 40 μL of 0.05 mg/mL pepsin (in 10 mM HCl) and incubated at 37 °C for 10 min. 400 μL of 50 mg/mL TCA was added to the digest, and after mixing, the solution was incubated again at 37 °C for 5 min. The solution was then filtered through a 0.45 μm syringe filter, and absorbance was measured using a Nanodrop 2000 UV/vis spectrometer at 280 nm. Activity was determined using eq 1, where 62.5 is a constant derived from the unit definition, t is the reaction time in minutes, mpepsin is the amount of added pepsin in mg, and Vtot is the total filtrate volume (0.64 mL).

hydrochloric acid overnight (4 M, 12 h at 100 μL/h) was performed to expose the NH2-functionalities. After washing, the monolith was conditioned with L-ascorbic acid (∼7% w/v ASA in 66% methanol) and sealed for 30 min. The incubation attaches ASA as a cross-linker to the NH2-functionalized monolith. After washing with DDW water (5 min, 30 μL/min), the channel was filled with enzyme solution (pepsin, 10 mg/mL in DDW), sealed, and incubated overnight (24 h, 4 °C). The immobilized amount was determined by measuring the absorbance of the enzyme solution, before loading and after incubation, using a Nanodrop 2000 UV/vis spectrometer (Thermo Fischer Scientific, Waltham, MA, USA) at 280 nm and correlating with a standard curve. Design and Fabrication of the 3D-Printed Fluidic Interface. Fluidic access to the chip was achieved through a 3D-printed interface. The interface was modeled with the help of computer-aided-design software (Autodesk Fusion 360, San Rafael, CA, USA) and fabricated on an “Ultimaker 2” 3Dprinter (Ultimaker, Geldermalsen, The Netherlands) from polylactic acid (PLA, Innofil3D, Emmen, The Netherlands) at 80 μm layer height, 1.2 mm shell thickness, and 50 % infill. Standard #10-32 coned tight fittings, or 1/4 in.-28 flat bottom fittings depending on the design (both from IDEX Health & Science LLC, Oak Harbor, WA, USA), together with 1/16 in. Teflon tubing can be used directly with the interface, facilitating integration with conventional HPLC and UPLC systems. Cylindrical PDMS O-rings (height 3 mm, OD 4 mm, ID 1.5 mm, ending in a torus with r = 0.35 mm and R = 1.35 mm) were fabricated in-house using the method previously described for the fabrication of PDMS chip molds. Interface back plates were cut from PMMA, using a desktop laser (Epilog Laser Mini model 8000 (30 Watt), Epilog Corporation, Golden, CO, USA), to fit the different designs. A schematic representation of the fluidic connections can be found in Figure S-2. Pressure Testing. The TE chips were tested for pressure tolerance by connection to a LC-pump system (Binary Solvent Manager, ACQUITY, Waters, Milford, MA, USA) operated with an empty (no monolith) chip and a constriction after the fluidic interface. Flow speed was increased from 0 to 2000 μL/ min, and the system pressure was monitored using the ACQUITY UPLC console software, until deformation could be observed. Device Operation and Comparison. All hemoglobin sample solutions were made in mobile phase A (0.23% formic acid (pH 2.5)) and run at room temperature with reference digestion performed on each sample using a column IMER (pepsin-immobilized packed agarose bead column (2 mm × 20 mm)) at 200 μL/min. For digestion at 2, 4, and 10 μL/min, a 0.5 μM solution of hemoglobin was pumped through the chip IMER at room temperature using a syringe pump. Twenty-five μL of the digest was injected into a G2 UPLC-MS system (Synapt G2 mass spectrometer with nanoACQUITY UPLC and HDX manager, Waters, Milford, MA, USA). For digestion at 13 μL/min, 20 μL of 0.03 mM hemoglobin was injected using a rheodyne injection valve (Model 7725i), with a 20 μL stainless steel injection loop. The LC-pump was run with mobile phase A, and a total volume of 120 μL was collected (6× dilution to 5 μM). The sample was further diluted to 0.04 μM, and 100 μL of the diluted sample was injected into the G2-Si UPLC-MS system (Synapt G2-Si mass spectrometer with nanoACQUITY UPLC and HDX manager, Waters, Milford, MA, USA).

activity(U/mg) =

Test Blank (A 280 − A 280 ) × 62.5 × Vtot t × mpepsin

(1)

The activity of the reactors was determined using a variation of the same method. The hemoglobin solution was digested onchip and on-column at 37 °C at flow rates of 13 and 200 μL/ 4575


Article


Figure 1. Interface design and testing. (A) Model of the “brick”-style interface with two-point clamping. (B) Model of the “jewelbox”-style interface with evenly distributed clamping. (C) Cross-section of the 1/4 in.-28 inlet. (D) Cross-section of the #10-32 inlet. (E) Pressure tolerance of the different interface designs. Asterisks indicate leakage.

min, respectively. The digest was then mixed with the previously described TCA and pepsin solutions and incubated at 37 °C for 5 min before filtering and absorbance measurement. Eq 2 shows how flow rate can be translated into a time equivalent, where Vsubstrate is the volume of hemoglobin solution used (0.2 mL) and Q is the flow rate in mL/min. t=

Vsubstrate Vtot = Q 3.2 × Q

(2)

SDS-PAGE. SDS-PAGE was carried out using a MiniPROTEAN Tetra Cell system (Bio-Rad, Hercules, CA, USA) and ready-made Bio-Rad Mini-PROTEAN TGX 7.5% polyacrylamide gels. 50 μL of hemoglobin solution (2 mg/mL) was digested on-chip and on-column at flow rates of 13 and 200 μL/min, respectively. 150 μL was collected to ensure complete elution of the sample (3× dilution). Thirty μL of the digested sample was mixed with 10 μL of Bio-Rad 4× Laemmli Sample Buffer and 0.5 μL of 6 M NaOH to ensure alkalinity of the solution. The resulting solution was denatured by heating at 95 °C for 5 min, and 20 μL of sample mixture containing 0.33 mg/ mL (0.11 mmol) of hemoglobin was loaded for electrophoresis. The gel was run at 160 V, 400 mA with Mark 12 unstained standard (Thermo Fisher Scientific, Waltham, MA). The bands were visualized using PhastGel Blue R-350 staining.

Figure 2. Effect of internal pressure on channel structural integrity. (A) After exposure to 200 psi. (B) After exposure to the maximum tolerated pressure of the interface (∼680 psi). Pressure was applied through flow from left to right.

in pressure tolerance was observed between the two designs. Due to the two-point clamping of the “brick” style interface, the chip-to-interface seal was less reproducible, as is apparent from the spread in breakout pressure for the replicates. Depending on the buffer composition used, clear oxidation could be seen on the metal bolts used to clamp the interface together due to leakage during demounting of the chip IMER. The “jewelbox” design provides an even pressure over the surface of the mounted chip due to its circumscribed threads and also is more convenient to handle manually than the “brick” style design, which required application of a uniform pressure at both bolts. The lack of metal parts also makes it more resistant overall to acidic buffers. Use of #10-32 coned fittings (Figure 1D) eliminated the flat surface at the bottom of the threaded port, and therefore, it printed with a smoother finish, increasing the sealing capabilities of the threaded port. Using coned fittings together with the “jewelbox” design allowed for a pressure tolerance of 684.3 ± 0.6 psi (Figure 1E). The TE chips investigated here were made from monomers mixed in stoichiometric ratios. These chips were found to have limited mechanical strength at high pressure (above 500 psi). Clear deformation as well as delamination could be observed before reaching the maximum pressure tolerance of the interface as seen in Figure 2. Under normal operating parameters with flows in the range of 2−50 μL/min, the pressure does not exceed 500 psi, which can be handled by these microreactors without problems. For operations requiring higher pressure, microfluidic chips where an excess of “ene” monomers is used will result in harder less deformable chips that can resist higher pressures.60

■

RESULTS AND DISCUSSION Pressure Testing of Chips in 3D-Printed Fluidic Interfaces. The pressure tolerance of the fluidic interface was tested using an LC-pump. Two basic interface designs were tested, a simple “brick” style with two-point bolt clamping (Figure 1A) as well as a hexagonal “jewelbox” design with integrated threads and evenly distributed clamping (Figure 1B). 1/4 in.-28 flat bottom fittings were tested for both designs (Figure 2C). When printing the respective interfaces, the orientation was very important. To ensure proper sealing, holes were printed orthogonally to the printed layers. The top part of the print inevitably has a smoother surface finish than the bottom part, which is why important structures, such as O-ring and chip alignment pockets, were printed facing up. Thus, the threaded ports accepting fittings were printed facing downward, and as such, the surface finish at the bottom of the port was somewhat compromised in the case of 1/4 in.-28 flat bottom fittings, the results of which can be seen in the lower pressure tolerance of these designs relative to the design incorporating #10-32 fittings (Figure 1D,E). Overall, only a small difference 4576


Article


described by León et al.61 Peptide identification using this method is dependent on the tandem mass spectrometry data and the software used to process this data (PLGS). The method describes the use of a single filtering criteria in that an identified peptide must be present in 2 out of 3 replicate runs. However, pepsin does not exhibit the same narrow specificity as trypsin and thus produces vastly higher peptide diversity upon digestion. Because of this, direct comparison of the set of peptic peptides derived from MS/MS identification between different conditions can result in dissimilar peptide coverage. To avoid introduction of false positives and/or negatives and to monitor a maximum number of peptides, a filtering approach via the DynamX software was used to generate an unbiased master peptide list62 (MPL) containing identified peptides from all samples. The MPL is then used as reference for all comparisons between digestions. An in-depth description of our workflow for unbiased comparison of pepsin digestions across different devices and/or conditions can be found in Figure S-4. Digestion Efficiency. As for the enzyme activity determination, flow rates of 200 and 13 μL/min, respectively, for the column and chip IMER digestion, were used. SDS-PAGE (Figure 3) of the digested samples revealed complete digestion

Comparison of the Physical Properties of the Chip Based IMER vs a Conventional Column Based IMER. Flow-through testing of empty chips showed a channel volume of 1.6 ± 0.1 μL [n = 5] which is in agreement with the theoretical volume of the design (1.60 μL). Flow-through testing of an in situ polymerized monolith on chip [n = 3] and a packed conventional column [n = 2] gave packing densities of 60 ± 3% and 84 ± 3%, respectively. Results from the tests can be found in Table 1. Table 1. Comparison of the Physical Dimensions and Properties of the Chip Based and Column Based IMER characteristic

chipa

columna

empty channel volume (μL) void volume of reactor (μL) packing density (%) bead diameter (μm) total surface area (cm2)

1.6 ± 0.1 0.66 ± 0.01 60 ± 3 1.4 ± 0.4c 25 ± 3c

62.8b 10 ± 2 84 ± 3 125 (mean)b 36000 ± 1000d

a

All calculated values are given at the 95% confidence level. bValues from manufacturer. cAssuming characteristics similar to previous publication by Lafleur et al.32 dAssuming 67.8 m2/mLgel.

Enzyme Immobilization and Activity. Pepsin immobilization on chip was carried out overnight, and UV-absorbance measurements on the collected pepsin solution were used to determine immobilized amounts. Enzyme activities of digested samples for the column and chip were estimated from UVabsorbance measurements and compared to standardized batchwise digestion. At a standard flow rate of 200 μL/min, the sample residence time in the column IMER was 3 s. For an unbiased comparison, samples were introduced to the chip IMER at a flow of 13 μL/min ensuring an identical sample residence time as for the column IMER during digestion. Results from both immobilization and activity determination can be seen in Table 2. The chip and column show approximately identical activity for the immobilized enzyme of about 520 U/mg. This is considerably less than what can be seen for free enzyme, which is to be expected due to the flow-through nature of the IMERs. Interestingly, the coverage when considering total surface area is about 70 times higher for the chip than for the column, resulting in a packing material specific activity (U/mL) that is 2.7× higher. This can likely be explained by the availability of the immobilized enzyme. The in situ polymerized monolith in the chip based IMER consists of effectively nonporous beads, suffering much less from diffusion limitations than the highly porous agarose gel. Workflow for MS Comparison of Pepsin Digestion. An unbiased approach for data processing and protein identification, of samples digested by trypsin, with LC-MSe has been

Figure 3. SDS-PAGE analysis of digested hemoglobin samples. M denotes the molecular weight ladder, and P denotes an undigested reference sample. Lanes 1−8 show loaded samples from four sequential digestions on the same chip IMER (lanes 1−4) or column IMER (lanes 5−8). Reference and digested samples contained intact hemoglobin at a concentration of 0.03 mM (0.6 nmol total in each digestion).

at 3 s residence time, of the α- and β-subunits of hemoglobin, for both the column and chip IMER, as can be seen from a lack of bands in the region of 14−20 kDa. The high efficiency of both the chip IMER and column IMER results in a high number of lower molecular weight peptides that are not resolved on the polyacrylamide gel, indicated by the low

Table 2. Amount and Activity of Immobilized Pepsin in the Chip and Column Based IMERs characteristic

batcha

chipa

columna

amount of pepsin (μg) surface coverage (ng/cm2) activity (U) enzyme specific activity (U/mg) support specific activity (U/cm3)

1.4 N/A N/A 4000 ± 100 N/A

6.6 ± 1.0 260 ± 70 3.4 ± 1.0 520 ± 110 3500 ± 1400

130 ± 5 3.7b 66 ± 20 520 ± 170 1300 ± 400c

a

All calculated values are given at the 95% confidence level with [n = 2]. b2.5 mgpepsin/mggel according to manufacturer. c6125 U/cm3 according to manufacturer. 4577


Article

Analytical Chemistry amount of visible bands on the gel. Over sequential repetitive runs, both the column and chip IMER displayed a reduction in digestion efficiency. Accordingly, an increase in the abundance of detectable peptides is apparent on the gel. Furthermore, in the case of the column IMER, the decrease in efficiency can also be noted as the band representing undigested α- and βsubunits of hemoglobin (approximately 15 kDa each) becomes more pronounced. This is likely a result of the unusually high amounts of substrate used during the SDS-PAGE experiments. The high concentration of substrate initially causes significant unspecific binding which then saturates over consecutive runs. Mass spectrometry was used to monitor the abundance of produced peptides following digestion and provides a more diverse and sensitive measure of digestion efficiency. Additionally, LC-MS represents the primary application area for both column and chip IMERs. Comparative digestions were performed at a range of flow rates from 2 to 10 μL/min, corresponding to residence times of 4−20 s, and the resulting digests were analyzed by LC-MS. By comparing the number of unique peptides identified, as well as the average length of these peptides as a function of residence time, the performance of the chip IMER was compared to the column IMER. In the measured range, a linear relationship can be seen between residence time and both the number of peptides and average peptide length (Figure 4). Thus, we can approximate that, to obtain a digestion identical to the column IMER operated with a residence time of 3 s, the chip IMER requires a residence time of 2.6 and 3.6 s, as assessed in terms of either the number of

peptides produced or average peptide length, respectively. The results therefore indicate that the two devices have very similar total digestion efficiency. Importantly, no decrease in activity over consecutive runs could be observed. A comparison of relative peak intensity for a number of representative peptides can be found in Figure S-5 which further underline the similar performance of the chip and column IMERs at comparable residence times. Peptide Mapping. The chip IMER was found to produce 107 ± 5 unique identifiable peptides on average compared to 93 ± 10 for the column IMER, under comparable conditions, a nonsignificant difference (p ≤ 0.05). The sequence coverage achieved after digestion is, in addition to the number of produced peptides, dependent on the length of said peptides. Average lengths of 21.0 ± 0.2 and 20.9 ± 0.2 amino acids, respectively, for the chip and column IMER were recorded, again with a nonsignificant difference (p ≤ 0.05). The assessment of number of peptides and sequence coverage demonstrates that chip IMERs achieve comparable performance to the classical packed column IMERs. Figure 5 shows the sequence coverage of the microreactor and packed column for hemoglobin chains α and β following digestion. Both devices, column and chip IMER, show complete coverage (99.3%, due to cotranslational cleavage of the initiator methionine63). Performance of Chip IMER Online. Utilizing the 3Dprinted interface, a successful coupling of the pepsin immobilized chip IMER directly to a UPLC-MS system was achieved. The UPLC-MS system is conventionally operated with the packed pepsin column IMER at a flow of 200 μL/min and an analysis time of 15 min. Utilizing the chip IMER at a comparable residence time of 3 s would translate into a flow rate of 13 μL/min, with a total analysis time of ∼60 min. The drastically increased analysis time would severely limit the usability of the microreactor in common proteomics workflows. To mitigate some of the time loss, the chip IMER was tested at a flow rate of 50 μL/min, corresponding to 0.8 s of residence time and a total analysis time of 18 min including subsequent trapping, desalting, and elution. Under these circumstances, the chip IMER still achieves a similar performance in relation to the number of identified peptides (93 ± 10) and average peptide length (20.9 ± 0.4). At these conditions, digestion is performed at a system pressure of ∼380 psi whereas the separation pressure is ∼4800 psi. Thus, using only 8 pmol of hemoglobin at a concentration of 0.04 μM, complete sequence coverage is still achieved for both the α- and β-chain (Figure S-6), demonstrating that the chip IMER can be used successfully for peptide mapping, even at such short residence times and in an online UPLC-MS setup.

■

CONCLUSIONS We have here demonstrated that a thiol-ene based pepsin microreactor (chip IMER) can achieve, under similar conditions, a digestion performance comparable to that of a classical pepsin column IMER, in which a column is packed with pepsin immobilized on agarose beads. Chip-immobilized pepsin show virtually identical apparent activity to the commercially available agarose packing material, but with a 70× higher surface coverage, the result of which is an almost 3× higher activity per volume of solid phase at the tested conditions. The vastly smaller size (void volume of 0.66 μL compared to 10 μL for the column) is made possible due to the monolithic thiol-ene stationary phase, which can be polymerized in situ and allows for easy immobilization of the

Figure 4. Relationship between residence time and (A) number of unique peptides produced or (B) average peptide length of the hemoglobin β-chain. Linear regressions show the necessary residence time for the chip IMER to reach comparable efficiency to the column IMER. The extrapolated residence time denotes when equal efficiency can be assumed. 4578


Article


Figure 5. Mapping of peptic peptides produced upon offline digestion of hemoglobin, at comparable conditions, using a chip IMER (shown in green) and conventional column IMER (shown in blue), respectively. The sequence coverage is shown for (A) the hemoglobin α-chain and (B) the hemoglobin β-chain.

■

enzyme. The small size of the chip IMER makes it possible to, in the future, include it in integrated lab-on-a-chip solutions, resulting in reduced sample consumption compared to LC based systems. Of further advantage, the design allows for a very low consumption of enzyme during the immobilization step, which should facilitate the generation of chip IMERs containing rare proteases that are difficult to obtain in high quantities but possess analytically useful alternative cleavage specificities. The introduction of a novel custom-designed 3Dprinted chip-interface provides fluidic access to the chip IMER and allows online coupling to high-pressure LC-MS setups through cheap and easily adaptable means. The pressure tolerance of ∼700 psi is sufficient for moderate pressure HPLC and UHPLC applications where a decoupled digestion is employed, as the IMER will be subjected to a much lower system pressure (∼380 psi) than is used during separation (∼4800 psi). Improvements in ongoing work could allow even higher pressures to further enhance the applicability of the chip based IMER in protein analysis and proteomics workflows. While the introduction of chip IMERs into an LC setup has been shown, the increase in analysis time limits their usability in high-throughput applications. Further work will focus on integrating consecutive unit operations, such as peptide trapping and ESI-interface, on the chip to provide low-volume, high speed, analysis of protein samples, as well as integrating the chip IMER into a HDX-MS conceptual setup. Overall, we believe that the thiol-ene based microfluidic reactor has great potential not only as a small, inexpensive, and low volume alternative to the currently available commercial columns but also as a platform for further development into a fully contained micro total analysis system for proteomics.

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.6b05103. Schematic of the fluidic interface, SEM micrograph of the thiol-ene monolith, comparison of our peptide identification workflow compared to the commonly used method, list of selected peptides, relationship between residence time and mean peak intensity, schematic of instrumental setup, and online-chip peptide map (PDF)

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: josiane.lafl[email protected]. *E-mail: [email protected]. Phone: +45 35 33 62 75. ORCID

Kasper D. Rand: 0000-0002-6337-5489 Josiane P. Lafleur: 0000-0003-4459-8801 Present Address ‡

J.P.L.: Institute of Applied Synthetic Chemistry, Faculty of Technical Chemistry, Getreidemarkt 9, 1060 Vienna, Austria. Author Contributions †

A.J. and R.R.S. contributed equally.

Notes

The authors declare no competing financial interest. 4579


Article


■

(32) Lafleur, J. P.; Senkbeil, S.; Novotny, J.; Nys, G.; Bøgelund, N.; Rand, K. D.; Foret, F.; Kutter, J. P. Lab Chip 2015, 15, 2162−2172. (33) Ç akmakçi, E.; Yuce-Dursun, B.; Demir, S. React. Funct. Polym. 2017, 111, 38−43. (34) Ríos, A.; Escarpa, A.; González, M. C.; Crevillén, A. G. TrAC, Trends Anal. Chem. 2006, 25, 467−479. (35) Svec, F.; Huber, C. G. Anal. Chem. 2006, 78, 2100−2107. (36) Sproß, J.; Sinz, A. Anal. Bioanal. Chem. 2009, 395, 1583. (37) Svec, F. J. Chromatogr. A 2010, 1217, 902−924. (38) Stachowiak, T. B.; Rohr, T.; Hilder, E. F.; Peterson, D. S.; Yi, M.; Svec, F.; Fréchet, J. M. J. Electrophoresis 2003, 24, 3689−3693. (39) Hoyle, C. E.; Bowman, C. N. Angew. Chem., Int. Ed. 2010, 49, 1540−1573. (40) Jensen, O. N. Nat. Rev. Mol. Cell Biol. 2006, 7, 391−403. (41) Rey, M.; Man, P.; Brandolin, G.; Forest, E.; Pelosi, L. Rapid Commun. Mass Spectrom. 2009, 23, 3431−3438. (42) Kadek, A.; Tretyachenko, V.; Mrazek, H.; Ivanova, L.; Halada, P.; Rey, M.; Schriemer, D. C.; Man, P. Protein Expression Purif. 2014, 95, 121−128. (43) Yang, M.; Hoeppner, M.; Rey, M.; Kadek, A.; Man, P.; Schriemer, D. C. Anal. Chem. 2015, 87, 6681−6687. (44) Athauda, S. B. P.; Matsumoto, K.; Rajapakshe, S.; Kuribayashi, M.; Kojima, M.; Kubomura-Yoshida, N.; Iwamatsu, A.; Shibata, C.; Inoue, H.; Takahashi, K. Biochem. J. 2004, 381, 295−306. (45) López-Ferrer, D.; Petritis, K.; Robinson, E. W.; Hixson, K. K.; Tian, Z.; Lee, J. H.; Lee, S.-W.; Tolić, N.; Weitz, K. K.; Belov, M. E.; Smith, R. D.; Paša-Tolić, L. Mol. Cell. Proteomics 2011, 10, M110.001479. (46) Fredrickson, C. K.; Fan, Z. H. Lab Chip 2004, 4, 526−533. (47) Sabourin, D.; Snakenborg, D.; Dufva, M. J. Micromech. Microeng. 2009, 19, 035021. (48) Kunstmann-Olsen, C.; Hanczyc, M. M.; Hoyland, J.; Rasmussen, S.; Rubahn, H.-G. Sens. Actuators, B 2016, 229, 7−13. (49) Saarela, V.; Franssila, S.; Tuomikoski, S.; Marttila, S.; Ö stman, P.; Sikanen, T.; Kotiaho, T.; Kostiainen, R. Sens. Actuators, B 2006, 114, 552−557. (50) Poulos, A. S.; Nania, M.; Lapham, P.; Miller, R. M.; Smith, A. J.; Tantawy, H.; Caragay, J.; Gummel, J.; Ces, O.; Robles, E. S. J.; Cabral, J. T. Langmuir 2016, 32, 5852−5861. (51) Tarn, M. D.; Hirota, N.; Iles, A.; Pamme, N. Sci. Technol. Adv. Mater. 2009, 10, 014611. (52) Kim, M.-K.; Kim, M. A.; Jenjob, R.; Lee, D.-H.; Yang, S.-G. Mater. Sci. Eng., C 2016, 62, 391−397. (53) Bokharaei, M.; Schneider, T.; Dutz, S.; Stone, R. C.; Mefford, O. T.; Häfeli, U. O. Microfluid. Nanofluid. 2016, 20, 1−14. (54) Pearce, J. M. Science 2012, 337, 1303−1304. (55) Gross, B. C.; Erkal, J. L.; Lockwood, S. Y.; Chen, C.; Spence, D. M. Anal. Chem. 2014, 86, 3240−3253. (56) Paydar, O. H.; Paredes, C. N.; Hwang, Y.; Paz, J.; Shah, N. B.; Candler, R. N. Sens. Actuators, A 2014, 205, 199−203. (57) Ma, G.; Su, Z.-G. Microspheres and Microcapsules in Biotechnology: Design, Preparation and Applications; Pan Standford Publishing Pte. Ltd.: Singapore, 2013. (58) Axelsson, I. J. Chromatogr. A 1978, 152, 21−32. (59) Anson, M. L. J. Gen. Physiol. 1938, 22, 79−89. (60) Sikanen, T. M.; Lafleur, J. P.; Moilanen, M.-E.; Zhuang, G.; Jensen, T. G.; Kutter, J. P. J. Micromech. Microeng. 2013, 23, 037002. (61) León, I. R.; Schwämmle, V.; Jensen, O. N.; Sprenger, R. R. Mol. Cell. Proteomics 2013, 12, 2992−3005. (62) Wales, T. E.; Eggertson, M. J.; Engen, J. R. Methods Mol. Biol. 2013, 1007, 263−288. (63) Gevaert, K.; Goethals, M.; Martens, L.; Van Damme, J.; Staes, A.; Thomas, G. R.; Vandekerckhove, J. Nat. Biotechnol. 2003, 21, 566− 569.

ACKNOWLEDGMENTS This project was funded by the Danish Council for Independent Research − Technology and Production Sciences (Grant DFF-4005-00341 and Sapere Aude Grant DFF-418400537A).

■

REFERENCES

(1) Wang, R.; Chait, B. T. Curr. Opin. Biotechnol. 1994, 5, 77−84. (2) Manabe, T.; Jin, Y.; Tani, O. Electrophoresis 2007, 28, 843−863. (3) Yin, Y.; Xiao, Y.; Lin, G.; Xiao, Q.; Lin, Z.; Cai, Z. J. Mater. Chem. B 2015, 3, 2295−2300. (4) Girelli, A. M.; Mattei, E. J. Chromatogr. B: Anal. Technol. Biomed. Life Sci. 2005, 819, 3−16. (5) Calleri, E.; Temporini, C.; Perani, E.; Stella, C.; Rudaz, S.; Lubda, D.; Mellerio, G.; Veuthey, J.-L.; Caccialanza, G.; Massolini, G. J. Chromatogr. A 2004, 1045, 99−109. (6) López-Ferrer, D.; Petritis, K.; Lourette, N. M.; Clowers, B.; Hixson, K. K.; Heibeck, T.; Prior, D. C.; Paša-Tolić, L.; Camp, D. G.; Belov, M. E.; Smith, R. D. Anal. Chem. 2008, 80, 8930−8936. (7) Ma, J.; Zhang, L.; Liang, Z.; Shan, Y.; Zhang, Y. TrAC, Trends Anal. Chem. 2011, 30, 691−702. (8) Geiser, L.; Eeltink, S.; Svec, F.; Fréchet, J. M. J. J. Chromatogr. A 2008, 1188, 88−96. (9) Hamuro, Y.; Wong, L.; Shaffer, J.; Kim, J. S.; Stranz, D. D.; Jennings, P. A.; Woods, V. L., Jr.; Adams, J. A. J. Mol. Biol. 2002, 323, 871−881. (10) Leurs, U.; Mistarz, U. H.; Rand, K. D. Eur. J. Pharm. Biopharm. 2015, 93, 95−109. (11) Wales, T. E.; Engen, J. R. Mass Spectrom. Rev. 2006, 25, 158− 170. (12) Slovakova, M.; Minc, N.; Bilkova, Z.; Smadja, C.; Faigle, W.; Fütterer, C.; Taverna, M.; Viovy, J.-L. Lab Chip 2005, 5, 935−942. (13) Peterson, D. S.; Rohr, T.; Svec, F.; Fréchet, J. M. J. Anal. Chem. 2002, 74, 4081−4088. (14) Liuni, P.; Rob, T.; Wilson, D. J. Rapid Commun. Mass Spectrom. 2010, 24, 315−320. (15) Křenková, J.; Foret, F. Electrophoresis 2004, 25, 3550−3563. (16) Ren, K.; Zhou, J.; Wu, H. Acc. Chem. Res. 2013, 46, 2396−2406. (17) Laurell, T.; Nilsson, J.; Marko-Varga, G. TrAC, Trends Anal. Chem. 2001, 20, 225−231. (18) Jang, L.-S.; Liu, H.-J. Biomed. Microdevices 2009, 11, 331−338. (19) Kim, J.; Park, H.; Jung, D.; Kim, S. Anal. Biochem. 2003, 313, 41−45. (20) Qin, M.; Hou, S.; Wang, L.; Feng, X.; Wang, R.; Yang, Y.; Wang, C.; Yu, L.; Shao, B.; Qiao, M. Colloids Surf., B 2007, 60, 243−249. (21) Cramer, N. B.; Davies, T.; O’Brien, A. K.; Bowman, C. N. Macromolecules 2003, 36, 4631−4636. (22) Harrison, C.; Cabral, J. T.; Stafford, C. M.; Karim, A.; Amis, E. J. J. Micromech. Microeng. 2004, 14, 153. (23) Good, B. T.; Reddy, S.; Davis, R. H.; Bowman, C. N. Sens. Actuators, B 2007, 120, 473−480. (24) Scott, T. F.; Schneider, A. D.; Cook, W. D.; Bowman, C. N. Science 2005, 308, 1615−1617. (25) Ashley, J. F.; Cramer, N. B.; Davis, R. H.; Bowman, C. N. Lab Chip 2011, 11, 2772−2778. (26) Wägli, P.; Homsy, A.; de Rooij, N. F. Sens. Actuators, B 2011, 156, 994−1001. (27) Levaché, B.; Azioune, A.; Bourrel, M.; Studer, V.; Bartolo, D. Lab Chip 2012, 12, 3028−3031. (28) Carlborg, C. F.; Haraldsson, T.; Ö berg, K.; Malkoch, M.; van der Wijngaart, W. Lab Chip 2011, 11, 3136−3147. (29) Saharil, F.; Carlborg, C. F.; Haraldsson, T.; van der Wijngaart, W. Lab Chip 2012, 12, 3032−3035. (30) Lafleur, J. P.; Kwapiszewski, R.; Jensen, T. G.; Kutter, J. P. Analyst 2013, 138, 845−849. (31) Feidenhans’l, N. A.; Lafleur, J. P.; Jensen, T. G.; Kutter, J. P. Electrophoresis 2014, 35, 282−288. 4580
