design and optimization of porous polymer enzymatic digestors for proteomics

Wei LinCameron D. Skinner

Department of Chemistry andBiochemistry, ConcordiaUniversity, Montr�al, Qu�bec,Canada

Original Paper

Design and optimization of porous polymerenzymatic digestors for proteomics

Effective protein characterization and identification are demanding and time-con-suming operations in proteomics because of long-protein purification/separationprocedures, and even longer enzymatic digestions. In this work, polymer-basedmonolithic enzyme reactors were fabricated in fused-silica capillaries, and perform-ance was characterized through protein digestion and identification by MALDI-MSand ESI-MS. Reactors were prepared by fabricating a porous methacrylate basemonolith followed by photografting with glycidyl methacrylate, and immobiliza-tion of the enzyme(s) with carbonyldiimidazole. Trypsin and Staphylococcus aureus V-8protease (Glu-C) were used to produce three types of reactors: trypsin-based, Glu-C-based, and trypsin combined with Glu-C. Protein digestions, performed by perfusingprotein solutions through the reactor under pressure, were evaluated based on thepeptide map generated when directly coupled to an ESI mass spectrometer. Excel-lent digestion was observed over flow rates from 0.2 to 1 lL/min, which correspondsto reactor residence times of 0.24–1.4 min. As a proof of principle, chromatographicseparation of model proteins followed by the digestion of specific fractions usingthese proteolytic enzyme reactors and ESI-MS is demonstrated.

Keywords: Dual-functional reactors / Immobilized Glu-C / Immobilized trypsin / Monolithic col-umn / Top-down approach /

Received: March 31, 2009; revised: May 16, 2009; accepted: May 16, 2009

DOI 10.1002/jssc.200900221

1 Introduction

The proteome is the entire set of proteins expressed bythe genome. In contrast to the genome, the proteome ishighly dynamic since the type of proteins expressed,their abundance, extent of modification, at any giventime and location, are dependent on the physiologicalstate of the organism or signals from the environment.Proteomics, the study of the proteome, has receivedremarkable attention due to its critical role in advancingknowledge for the life sciences, the identification of dis-ease processes, and in understanding drug actions [1, 2].

In general, there are two approaches used in proteo-mic surveys, top-down and bottom-up. The top-downapproach commences with the separation of a complex

mixture of proteins harvested from the system understudy followed by proteolytic digestion of specific frac-tions and identification of the components in those frac-tions based on the peptides generated [3]. Bottom-up pro-teomics relies on peptide mapping, in which all the pro-teins in the sample are digested first, followed by separa-tion of the peptides and identification of these peptides[4]. Both techniques usually rely on MS for peptide identi-fication. No matter which proteomic approach isemployed, separation, enzymatic digestion, and peptideidentification are required.

Conventionally, digestion is achieved by incubation ofthe protein(s) and protease in a suitable buffer solution [5,6]. To minimize proteolytic autolysis, low protease to sub-strate ratios are used, typically 1:20–1:100 w/w that neces-sitate extended incubation times of 3–24 h and lower thedigestion efficiency. An additional limitation of in-solu-tion digestion is that it often includes extensive manualsample handling steps. A new concept of protein diges-tion using immobilized proteases in microreactors hasdrawn significant attention during the last few decades[7–10]. This method has a number of advantages over thein-solution digestion method, such as minimization, oreven elimination, of protease autolysis, larger enzyme tosubstrate ratios, shortened digestion time, and high

Correspondence: Dr. Cameron D. Skinner, Department ofChemistry and Biochemistry, Concordia University, 7141 Sher-brooke Ouest, Montr�al, Qu�bec, H4B 1R6, CanadaE-mail: [email protected]: +1-514-848-2868

Abbreviation: ACTH, adrenocorticotropic hormone fragment1 – 10 human; AMPS, 2-acrylamido-2-methyl-1-propanesulfonicacid; BAPNA, benzoyl-L-arginine 4-nitroanilide hydrochloride;Glu-C, Staphylococcus aureus V-8 protease; GMA, glycidyl methacry-late; PEG, poly(ethylene oxide)

i 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

2642 W. Lin and C. D. Skinner J. Sep. Sci. 2009, 32, 2642 – 2652

J. Sep. Sci. 2009, 32, 2642 –2652 Other Techniques 2643

digestion efficiency [8]. More importantly, these micro-reactors can be easily coupled with mass spectrometers toperform automated online protein characterization [7].

A variety of proteases, such as trypsin [11–15], chymo-trypsin [16], pepsin [17, 18], elastase [19], and papain [20],have been immobilized onto both the walls and supportpackings in capillaries and microfluidic chips. Enzymeswere attached to the support primarily through threemethods, covalent binding [21], physical adsorption [22–24], and sol–gel encapsulation [25, 26]. Generally, the sup-port that anchors the proteases can be divided into twogroups, silica-based [27–29] and polymer monolith-based[30–36]. Silica-based supports (either the capillary wall ora sol–gel monolith) are not utilized as extensively as theirorganic counterparts mainly because of the limited chem-istries available for silica, potential non-specific interac-tions with silanol groups, and low stability at extreme pHconditions. On the other hand, organic monoliths wereutilized as the support for immobilized enzymes [7] verysoon after they were introduced as HPLC stationaryphases in 1989 [37]. The key reasons for their popularityseem to be the absence of interparticulate voids, fast masstransfer kinetics, ease of fabrication, and facile modifica-tion with a wide variety of chemistries [38, 39].

Several research groups have directed their worktoward the development of bottom-up online proteincharacterization systems consisting of a high-throughputproteolytic reactor followed by LC or CE separations withMS detection [21, 40–44]. However, less attention hasbeen focused on developing top-down systems [45]. Thisapproach is worth exploring, since it not only facilitatesthe association of generated peptides with a particularprotein, but it can also be adjusted dynamically for pro-tein abundance which would aid in the identification oflow abundance proteins and de novo protein sequencing.

This paper describes the preparation and characteriza-tion of methacrylate monolithic enzyme reactors infused-silica capillaries. Two types of proteolytic reactors,produced by covalently immobilizing trypsin and S. aur-eus V-8 protease (Glu-C) onto glycidyl methacrylate (GMA)grafted monolithic support will be shown. The reactorswill be evaluated by quantifying the immobilizedenzymes, evaluating the effects of contact time, reactortemperature, and buffer pH. Finally, a monolithic col-umn for protein separation and one monolithic trypticreactor for protein digestion were serially linked for aproof of principle demonstration of top-down automatedonline protein identification.

2 Experimental

2.1 Materials and reagents

Unless otherwise stated, all reagents were purchasedfrom Sigma–Aldrich (Oakville, ON, Canada). Glu-C was

purchased from Roche (Laval, QC, Canada). Amino termi-nated poly(ethylene oxide) (PEG) was obtained from Poly-mer Source (Montr�al, QC, Canada, MW 5000). The BCAkit was from Pierce (Rockford, IL, USA). All solvents wereHPLC grade from Fisher (Ottawa, ON, Canada) and mix-tures are volume percent unless otherwise noted. Waterwas obtained from a Barnstead EASYpurem II UV Ultra-pure water system (Dubuque, IA, USA) at 18.2 MX cm.

2.2 Preparation of monolithic capillary columns

Capillary columns were silanized according to themethod described by Ngola et al. [46]. Fused-silica capilla-ries (UV transparent 100 lm id, 365 lm od or 535 lm id,665 lm od with 5 cm of the polyimide removed) fromPolymicro Technologies (Phoenix, AZ, USA) were treatedsuccessively with ethanol, 1 M sodium hydroxide, andwater, followed by a silanization mixture composed ofwater, glacial acetic acid, and 3-(trimethoxysilyl)propylmethacrylate (50:30:20). After 12 h, the silanized capilla-ries were consecutively washed with methanol, water,and then dried with nitrogen gas.

The monomer mixture was composed of butyl acrylate(BAC), 1,3-butanediol diacrylate (BDDA), 3-(trimethoxysi-lyl)propyl methacrylate (an adhesion promoter), 2-acryl-amido-2-methyl-1-propanesulfonic acid (AMPS)(69.2:30:0.3:0.5) and the photo-initiator, benzoin methylether (BME, 0.5 wt%), was added just prior to exposure.The porogenic solvent consisted of ACN, ethanol, and5 mM phosphate buffer, pH 7.0 (60:20:20). The capillarieswere filled with polymerization mixture (1:2 monomermixture/porogenic solvent) and 10 cm was exposed to UVlight for 25 min from a high-pressure mercury lamp.Unreacted monomers were purged with ACN at 6.9 barfor 1 h.

The photografting polymerization mixture, 199 lLGMA and 15 mg BME in 1 mL ACN, was flushed throughthe capillary and the monolith was exposed to UV lightfor 10 min. The capillaries were rotated at the fifthminute in the photografting process based on the sugges-tion from Hilder et al. [47]. Unreacted monomers wereflushed out with ACN.

2.3 Immobilization of proteases

The process for the immobilization of trypsin wasadapted from the methods of Hearn et al. [15] and Ben-cina et al. [14]. Briefly, GMA grafted monolithic columnswere incubated with 0.5 M sulfuric acid overnight at508C and successively rinsed with 90 mM 1,19-carbonyl-diimidazole (CID), ACN, and 1 mg/mL protease in 0.1 Mbicarbonate buffer, pH 8.0, containing 10 mM calciumchloride as an inhibitor using pressure generated flow at6.9 bar, producing l0.4 lL/min for 1 h. The trypsin-basedand Glu-C-based reactors were stored in 10 mM Tris-HCl



(pH 8.0), 10 mM calcium chloride, and sodium azide(0.02%) at 48C prior to use.

2.4 Immobilization of trypsin and Glu-C in onecolumn

Capillary columns containing two 5 cm long sections ofmonolith (without GMA) separated by 15 cm of opencapillary were fabricated as described in Section 2.2. Onemonolith was GMA modified and trypsin was immobi-lized by flushing the enzyme through the entire capil-lary using the same procedure described in Section 2.3.Pressure was applied from the end to be photografted for� of the capillary dead time (l100 s) to fill the second sec-tion of monolith with GMA grafting solution andexposed to UV for 10 min, while the rest of the capillarywas covered with aluminum foil. Immobilization of Glu-C was carried-out by flushing with 1 mg/mL Glu-C and10 mM calcium chloride in 0.1 M bicarbonate buffer,pH 8.0, at 48C for 2 days, flushed and stored as above. Theoverall column schematic is shown in Fig. 1.

2.5 Digestion with dual enzyme reactor

Adrenocorticotropic hormone fragment 1–10 human(ACTH, 1 mg/mL in 50 mM ammonium bicarbonate,pH 7.8) digestion was conducted under constant pressure

after each immobilization step. The digests were col-lected and analyzed by CE and ESI-MS.

2.6 Characterization of proteolytic reactors

Quantification of immobilized enzyme: the BCA reagentsolution was pumped through 5 cm long, 535 lm idmonolithic columns with, and without, immobilizedproteases by pressure (6.9 bar) for 1 h yielding about600 lL. A 200 lL aliquot of the effluent was analyzedusing the BCA assay according to the instructions fromthe manufacturer.

Buffer pH and temperature effect: cytochrome c(0.5 mg/mL) solutions were prepared in 50 mM ammo-nium bicarbonate and flushed through tryptic and Glu-Creactors under constant pressure (6.9 bar, produc-ing L 0.4 lL/min). For the pH study (pH 6.5–9.0) thedigestor was held at room temperature (238C) but for thetemperature study (23–538C, pH = 8.0), the reactor wasimmersed in a temperature controlled water bath dur-ing digestion. The digests were collected and analyzed byMALDI-MS.

Flow rate effect: Online cytochrome c tryptic diges-tions were carried out with the set-up shown in Fig. 2. AHarvard Apparatus (Saint-Laurent, QC, Canada) syringepump varied the flow rates through the digestor from0.1 to 2 lL/min. After allowing the sample to passthrough the digestor for 15 min, a 5 lL fixed volume of


Figure 1. Scheme for the preparation ofenzyme reactors with two proteases. (A)Empty Teflon coated capillary (100 lmid/365 lm od); (B) monolith column fabri-cated as described in Section 2.2; (C)one section of monolith was photograftedwith GMA by masking the other sectionduring exposure; (D) trypsin was immobi-lized onto the GMA grafted monolith; (E)the second section of monolith was pho-tografted with GMA; (F) Glu-C wasimmobilized onto the second GMAgrafted monolith.


peptide solution was retained in the C18 trapping col-umn (NanoEaseTM, 0.18 mm id623.5 mm, Waters) andeluted for ESI-MS using a 1 lL/min, 0–95% linear gra-dient of ACN in 0.1% formic acid over 45 min.

2.7 Tryptic digestion

Cytochrome c, apo-myoglobin, a lactalbumin and BSAsolution phase tryptic digestions were carried out at aprotein monomer to trypsin ratio of 20:1 w/w at roomtemperature for 15 h in 50 mM ammonium bicarbonate,pH 8.0. a Lactalbumin and BSA were reduced with100 mM DTT and alkylated using 50 mM N-ethylmale-imide for 1 h prior to both in-solution and online diges-tions. Excess DTT and N-ethylmaleimide were removedby centrifugal ultrafiltration (ultrafree 0.5 centrifugal fil-ters; Millipore, Billerica, MA, USA). The protein digestswere diluted 5–30-fold into 50% ACN containing 0.1%formic acid, and introduced into a Waters Micromass Q-ToF2 mass spectrometer by direct infusion.

These four proteins were also digested by the trypticreactor and compared to the in-solution method withthe experimental set-up shown in Fig. 2.

2.8 Online protein separation and tryptic digestion

A schematic representation of combined online proteinseparation and digestion at 238C is shown in Fig. 3. Onehalf microliter of 166 lM cytochrome c and 100 lM apo-myoglobin mixture was injected into a tandem 15 cmmonolithic and 10 cm digestion column (both 100 lmid). Separation and digestion were carried out under iso-cratic conditions using 20% ACN in 50 mM ammoniumbicarbonate, pH 8.0 at 0.5 lL/min. The peptides elutedfrom the tryptic reactor were diluted with 0.5 lL/min

ACN and 0.1% formic acid through the Tee and analyzedby ESI-MS.

2.9 Monolith surface modification by PEG

Five monolithic reactors with varying amounts of trypsinand amino terminated PEG were fabricated as describedin Section 2.3. During immobilization, 2 mg/mL mix-tures of trypsin and PEG (0, 25, 50, 75, and 100 wt%) werereacted with the surface.

A peptide mixture (Gly–Tyr, Val–Tyr–Val, methionineenkephalin acetate, leucine enkephalin, angiotensin IIacetate, total 10 pmol) was injected into the trypsin-PEGreactors. The mobile phase was 20% ACN in 50 mMammonium bicarbonate, pH 8.0 at 0.5 lL/min, and waspost-column mixed with 0.5 lL/min ACN and 0.1% for-mic acid through the Tee and analyzed by ESI-MS. Cyto-chrome c (14 pmol) was used to study the digestion effi-ciency of these reactors with the experimental set-updescribed above.

2.10 Instrumentation

Samples were prepared for imaging by applying pieces ofcolumn to sticky carbon foils on standard aluminumspecimen stubs and coated with a 20 nm thick gold layerby an Edwards sputter coating unit (Crawley, West Sus-sex, UK). Microscopic analysis of all samples was carriedout in a Hitachi S-4300 SE/N SEM (Hitachi High Technol-ogy, Pleasanton, CA, USA) operated at 5–10 kV, 100 pAprobe current, and 08 tilt angle.

CE experiments were performed using a P/ACE MDQCE System equipped with a diode array detector (DAD)and Karat 32 software (Beckman Coulter, Fullerton, CA,USA).


Figure 2. Schematic representation ofthe system used for online tryptic diges-tion. Solvent A was 0.1% formic acid; sol-vent B was 0.1% formic acid in ACN. Thereactor temperature was 238C.

Figure 3. Representation of the systemfor online protein separation and diges-tion. Solvent A was 50 mM ammoniumbicarbonate, pH 8.0; solvent B was ACN;and solvent C was 50% ACN and 0.1%formic acid. The temperature was 238C.


Protein sequencing was performed on a Spec E MALDI-TOF mass spectrometer from Micromass (Waters Micro-mass) equipped with a 337 nm N2 laser. Calibration wascarried out using angiotension I (1296.5 Da), Glu-fibrinopeptide (1570.6 Da), and ACTH (fragment 18–39,2465.7 Da). Digested samples were desalted using C18 Zip-Tip pipette tips (Millipore, Bedford, MA, USA), and resus-pended in 1.5 lL ACN (60%) and TFA (0.1%). The treatedsample was then mixed with an equal volume of a solu-tion of 50 mM a-cyano-4-hydroxy cinnamic acid in 50:50ACN/ethanol. The room temperature dried-dropletmethod described by Karas and Hillenkamp [48] was usedfor the duplicate analysis of the samples (1 lL) depositedon a MALDI PrepTarget stainless steel plate. Data wererecorded in reflectron positive ion mode.

ESI-MS analysis was performed on a Waters MicromassQ-ToF2 mass spectrometer equipped with a Z-spray ionsource and a ternary gradient Micromass CapLC system.Mass calibration was carried out using human [Glu1]-fibrinopeptide B, and yielded a mass accuracy ofl30 ppm in positive ion mode. The instrumental param-eters are listed in the figure legends. Data analysis wasperformed using MassLynx 4.0 software (Waters Micro-mass), and peptide fingerprint searches were carried outusing MASCOT (Matrix Science).

3 Results and discussion

The morphology of monolithic stationary phases is ofcrucial importance, when these types of stationary phaseare applied to separations [49] and enzymatic digestors[21]. Bed homogeneity largely determines the separationefficiency (plate height) that can be attained for a givenmonolith and is one of the chief advantages of in situ poly-merized monoliths over similarly dimensioned packedcapillary columns [50]. The high surface area to volumeratio, and high degree of interconnected channels pro-duced in photopolymerized monoliths provides rapidmass transfer for both large and small analyte moleculesand is advantageous for both chromatographic and pro-teolytic digestor applications. The methacrylate-basedpolymer used here has demonstrated its suitability forhigh efficiency CEC separations [46, 51] which rely onEOF generated by the AMPS. Although AMPS was notdirectly necessary for the digestor experiments, it wasmaintained in the polymer recipe so that the morphol-ogy of the base polymer would not be altered comparedto previous reports [46, 51] and would be available to pro-vide EOF in electro-driven applications, even after photo-grafting [52].

The key aspects to producing a successfully modifiedmonolith are the ability to incorporate the desired func-tional groups onto the surface without altering the basicmorphology of the monolith. Figure 4 shows SEM micro-

graphs of the cross-section of the photopolymerizedmonolith fabricated in a 100 lm id capillary. Theseimages show that the monolith is homogeneous acrossthe capillary and that the silanization step allows themonolith to be anchored to the capillary wall. Althoughit is not easily visible at the 9006 magnification in Fig. 4,it is possible to observe that the wall is covered in a skinof polymer and the majority of the wall surface is popu-lated with polymer nodules/agglomerated microglo-bules which should reduce wall-bed bandbroadeningand silica –analyte interactions. Similar images wereobtained for the GMA photografted polymer (not shown)and no visible changes in morphology were observed.The average microglobule size determined from SEMimages was l1 lm and is similar to those observed forthe same type of monolith prepared by Ngola et al. [46].Validation of the morphology, by SEM imaging, is criticalsince slight variations in experimental factors such asirradiation source, light intensity, and capillary temper-ature can dramatically affect the size of the microglo-bules and the pore size distribution [53].

There are several routes to preparing enzymatic reac-tors but photografting has demonstrated its capacity toproduce well-defined monoliths with a wide variety ofsurface chemistries [54]. This technique permits con-trolled introduction of specific surface functionalitiesonto the monolith that may be difficult to achieve whenthese same functionalities are incorporated into themonomer mixture. By taking advantage of this uniquefeature of photografting, we easily immobilized bothtrypsin and Glu-C onto the monolith but any number ofother molecules could have been immobilized (e.g., anti-bodies, aptamers, chiral selection moieties, etc.).

Our interests were in developing a versatile enzymaticdigestor and immobilization strategy so trypsin and Glu-C were chosen as good initial candidates for proteolyticenzymes due to their popularity in proteomic research.In this work, quantification of the protease immobilized,using the BCA protein assay, was made on larger 535 lmid reactors because the smaller 100 lm id reactors did


Figure 4. Electron micrographs of the cross-section of amonolith in a 100 lm id capillary. Polymerization mixture:69.2% BAC, 30% BDDA, 0.3% 3-(trimethoxysilyl)propylmethacrylate, and 0.5% AMPS. Magnification: 9006 (left)and 30006 (right).


not produce sufficient product (Cu+) for quantification.The results showed that 340 l 140 lg of trypsin wereimmobilized per 1 g of support; and the Glu-C columnhad 320 l 140 lg/g support (Table 1). This is in sharp con-trast to preliminary results where we obtained 5.8 mgtrypsin/g of monolith using the absorbance-basedmethod. The three reports that used the absorbance-based method found 7.75 mg trypsin/mL polymer [40],2.9–7.1 mg papain/g support [20], and 4.9 mg trypsin/gcontrolled pore glass [55]. The disparity between theabsorbance and BCA results cannot be reconciled but theBCA assay directly probes the protein immobilized onthe monolith whereas the absorbance-based assay meas-ures the difference in the immobilization protein con-centration and assumes that all of that protein is immo-bilized, a potentially tenuous assumption.

One of the most important parameters that affects pro-teolytic digestion speed and efficiency is the enzyme tosubstrate ratio. With in-solution digestion this ratio istrivial to control and was fixed at the frequently usedvalue of 1:20. Assuming that the results from the BCAmeasurements are correct the protease to protein ratiowas 3:1 for the reactors. This relatively high local concen-tration of the protease ensured that the reactor enzy-matic digestions were faster than the in-solution diges-tion while producing comparable sequence coverages.

It is well known that enzyme activity in solution islargely influenced by several parameters, such as buffercomposition, pH, digestion time, digestion temperature,and so on. It has been reported that the optimal bufferpH and digestion temperature for trypsin are pH 8.0 and378C [56] while for Glu-C a pH 7.8 and 378C [57] are opti-mal. Compared to their counterparts in solution, immo-bilized enzymes might have different values for thoseparameters due to the influence of the immobilizationprocess (restraint on conformational change and rota-tion of enzymes, steric hindrance of active sites, etc.) andunique properties of the stationary support such as theelectrical double layer established by the AMPS. Thus, weinvestigated the digestion performance of both immobi-lized trypsin and Glu-C reactors in 50 mM ammoniumbicarbonate, pH 6.5–9.0, at 238C and from 23 to 538C but

at a fixed buffer pH of 8.0. Both experiments were per-formed with a constant flow rate of 0.5 lL/min usingcytochrome c as a test substrate. Table 2 shows that theoptimal pH for immobilized trypsin was from 8.0 to 8.5,which agrees with that reported for in solution digestionand with those observed from trypsin immobilized onporous monoliths [45, 58] and on porous glass [55]. TheGlu-C reactor showed a wider optimal pH range thantrypsin with relatively good digestion efficiency betweenpH 7.0 and 8.0 that compares well with the reported in-solution efficiency at pH 7.8. A relatively flat tempera-ture profile for both immobilized trypsin and Glu-C wasobserved with uniform sequence coverages obtainedbetween 33 and 438C for trypsin and 37–508C for Glu-C.However, the selective cleavage at glutamic acid of theGlu-C digestor led to a lower than expected sequence cov-erage that was explained by the realization that thedigestor should produce a peptide (4470 Da) that repre-sents 40% of the cytochrome c sequence, but the MS scanparameters only allowed peptides up to 3000 m/z to bedetected. These two studies suggested that the basic prop-erties of these two proteases were not altered by theimmobilization process and are in agreement with tryp-sin immobilized on other supports using different chem-istries [59–61].

Continuous online protein digestion using proteolyticreactors is dependent on flow rate, since it determinesthe contact time between the substrate and the proteasewithin the reactor. Therefore, investigation of the effectof flow rate on protein digestion in tryptic reactors usingcytochrome c as a model protein was carried out (Fig. 5)with the experimental set-up shown in Fig. 2. The calcu-lated contact time at each flow rate, listed in Table 3, isbased on a monolith total porosity of 0.6 determinedfrom conductivity measurements following the methodproposed by Tavarna [62]. At the lowest flow rate meas-ured, 0.1 lL/min (Fig. 5A), all the peptide ions identifiedwere free of missed cleavages except for ions at m/z 562and 1598. The low flow rate allowed the longest contacttime, therefore, resulting in a more complete digestion


Table 2. Sequence coverage (percent) of cytochrome c as afunction of pH and temperature

Buffer pH Temperature

pH Trypsin Glu-C Tempera-ture (8C)

Trypsin Glu-C

6.5 N/A 8 23 60 177.0 68 26 33 60 267.5 69 28 37 76 268.0 81 28 43 77 288.5 82 20 50 69 289.0 67 N/A

N/A, not applicable.

Table 1. Quantification of enzyme immobilized on 535 lm idcolumns and normalized to a 100 lm id column

Trypsin Glu-C

Enzyme mass (lg/535 lmid column)

10.2 l 4.25 9.45 l 4.25

Enzyme mass (lg/g polymer)(on 535 lm id column)

339 l 142 315 l 142

Expected enzyme mass(lg/100 lm id column)

1.02 l 0.43 0.950 l 0.43

Enzyme/protein (g/g)(at 0.5 mg/mL protein)

3:1 3:1


and fewer missed cleavages. As the flow rate wasincreased (0.3–1 lL/min), more peptide ions wereobserved and consequently sequence coverage increasedfrom 60 up to 87%. However, the number of peptideswith missed cleavages also increased from 2 to 7 suggest-ing that there was insufficient contact time to allow com-plete digestion. For the digestion at the highest flow rate

(2 lL/min, Fig. 5E), intact protein ions and large peptideions with missed cleavages were dominant in the spec-trum. At these high flow rates, inefficiency of diffusivemass transport of the substrate to the enzyme signifi-cantly limited the enzymatic reaction. Incomplete diges-tion is undesirable since the presence of the intact pro-tein ions not only competed with peptide ions for ioniza-tion, but also complicated analysis of the MS spectrum.As a result, both the peptide signal intensity and the S/Ndecreased resulting in a modest sequence coverage of62%. No MS/MS sequencing was required for the positiveidentification of the peptides since the mass error wasaround 30 ppm. The optimal flow rate for cytochrome cdigestion was 0.5–1 lL/min, but the optimal flow ratefor digestion of other proteins may be different. Four pro-teins, with molecular masses from 12 to 69 kDa and pIsfrom 4.5 to 10.0, were digested by the tryptic reactor andcompared to the in-solution method to demonstrate thatthe reactors were capable of digesting a wide variety ofproteins. The experimental set-up for online digestion isshown in Fig. 2 with the digestion results presented inFig. 6. The salient points from this experiment are thatthe digestor produces sequence coverages from 43 to85% with less than two and a half minutes of contact andcan produce peptides not observed with in-solutiondigestion. The larger proteins did not produce as largesequence coverages as the smaller proteins and wouldsuggest that a compromise flow rate would have to befound in the case where multiple proteins were to bedigested. Higher BSA sequence coverage from immobi-lized trypsin, with more than 4–8 min digestion time,has been reported in the literature [63, 64], but in ourexperiment, the digestion time was 1.4 min. It suggeststhat in general large and/or resistant-to-digest substratesrequire longer contact times in order to obtain highsequence coverage. In all cases, protease autolysis prod-ucts were absent in the digestates. This compares very


Figure 5. Online tryptic digestionof 0.5 (lg/lL cyctochrome c in50 mM ammonium bicarbonate,pH 8.0, at various flow rates. (A)0.1 lL/min, (B) 0.3 lL/min, (C)0.5 lL/min, (D) 1.0 lL/min, and(E) 2.0 lL/min. For MS analysis,the instrumental parameterswere as follows: source blocktemperature, 808C; capillary volt-age, 3.5 kV; cone voltage, 35 kV;collision voltage, 10 V (no colli-sion gas); TOF, –9.1 kV; MCP,1.8 kV. RA is the relative abun-dance of the ions, and TIC is totalion count. Identified peptides arelabeled with W, and the intact pro-tein envelope is labeled with C.

Table 3. Overview of online cytochrome c digestion withimmobilized trypsin at different flow rates

Flow rate (lL/min) 0.1 0.3 0.5 1 2

Contact time (s) 282 94.2 56.4 28.2 14.4Peptide m/z MCa) MCa) MCa) MCa) MCa)

261 0 0 0332 1 1405 0434 0562 1 1604 0 0 0634 0 0 0 0 0678 0 0 0 0 0779 0 0 0 0 0806 1 1964 0 0 0

1168 0 0 0 0 01350 1 1 1 11470 0 0 0 01478 2 2 21495 0 01598 1 1 1 1 11606 3 31623 1 11633 1 1 1 11712 1Sequencecoverage

60% 73% 87% 82% 62%

a) MC is the number of missed cleavages in the detectedpeptide. An empty cell means that the peptide was notdetected at that flow rate. Note, not all of the expectedtryptic peptides are listed.


favorably with the in-solution digestion which requiredl15 h and several autolysis peptides were observed.

There are several enzyme immobilization monolithtechniques, including direct immobilization with GMA[58]; however significant non-specific sorption of proteinand digestion products has been observed [65]. We inves-tigated the potential for peptide adsorption on themonolith by injecting a mixture of peptides with varyinghydrophobicities [66] and charges onto monoliths thatwere immobilized with varying amounts of trypsin(100 fi 0%) and aminated PEG (0 fi 100%). Interestingly,all peptides co-eluted (with 20% ACN) for all of the col-umns. The retention factor decreased steadily from 5 fora purely trypsin column through to 3 for a pure “PEG”column. This data suggest that there is some mode ofretention but hydrophobic, size-exclusion, or ion-exchange mechanisms would not explain the data. Atthis time we do not have an adequate explanation for theretention mechanism at work. With these same columns,cytochrome c sequence coverage was 72 l 10% and didnot show any systematic variation for the columns withtrypsin.

Given the versatility of the monolith and photograft-ing strategy two novel tandem columns were investi-gated. First we demonstrate a protein separation directlyfollowed by a proteolytic digestion of the separated pro-teins with online MS detection. In order to perform theonline protein separation and digestion in tandem, theamount of organic composition in the solvent needed tobe a compromise between the high organic fractionsthat provided adequate protein separation and the low

tolerance for organic solvents of the protease. Peterson etal. [59] have demonstrated enhanced stability of immobi-lized enzymes in organic solvents but on a different sub-strate. One convenient method of assessing activity is touse a chromogenic substrate such as benzoyl-L-arginine4-nitroanilide hydrochloride (BAPNA) [67] but we foundthat adsorption of BAPNA on the stationary phase pre-vented its use. Therefore, we conducted the BAPNA activ-ity assay of trypsin in solution, and made the assumptionthat the data should indicate a concentration of ACNthat would not be detrimental to the immobilizedenzymes in light of Peterson's findings. Figure 7 showsthat trypsin maintained its activity up to 20% ACN butthe activity rapidly decreased after 40% ACN. Therefore, asolvent system consisting of 50 mM ammonium bicar-bonate, pH 8.0 and 20% ACN was used to perform anonline isocratic protein separation followed by digestion


Figure 6. Pie chart representa-tions for the digestion of fourproteins comparing thesequence coverages for both in-solution and online with thedigestor. For example in chart Acytochrome c was digested.Both methods had 74% of thesequence coverage in commonbut the online digestor had anadditional 13% coverage pro-ducing a total of 87%. Bothmethods failed to produce pep-tides over 13% of the sequence(no coverage); (B) apo-myoglo-bin; (C) a lactalbumin; (D) BSA.SC, sequence coverage.

Figure 7. In-solution trypsin activity at 238C of 0.1 mg/mLtrypsin in 50 mM ammonium bicarbonate, pH 8.0 in ACNwith 2 mM BAPNA as a substrate.


(Fig. 3). The total ion current chromatogram and corre-sponding MS spectra are presented in Fig. 8 where sixpeptides from apo-myoglobin were identified in the twopeaks beginning at 18 min through to 21 min, yielding asequence coverage of 40%. No peptides from cytochromec were found in this region. Twelve peptides from cyto-chrome c were observed in the last two peaks from 21 to26 min and yielded a sequence coverage of 73% but therewas one peptide from apo-myoglobin (2150.3 Da).

The data demonstrate that cytochrome c and apo-myo-globin were base-line separated by the monolithic col-umn, and that the separation was maintained in thedigestor even as the protein was degraded into peptides.One problem that defies explanation at this time is whythe digested material eluted as two separate peaks, butthere is a gap of open capillary between the end of theseparation monolith and the start of the reactor mono-lith of about 4 cm which may result in unexpected peakdistortions. It is noteworthy that the sequence coveragewas significantly lower for apo-myoglobin compared tothe previous experiments where 85% was found, but themass of protein in this experiment was about one orderof magnitude smaller than when the protein was (contin-uously) infused through the reactor and the missing pep-tides produced some of the lowest ion intensities relativeto the peptides detected in both experiments. We dis-counted the possible loss of peptides in the reactor dueto adsorption based on the coelution of injected peptidesas explained above. The most important point of thisexperiment is that separation and identification of thesetwo proteins via their peptides was completed within30 min. With further efforts to improve the peak effi-ciency, e.g., by using these monoliths in CEC mode andincreasing the mobile phase elutropic strength withadditives, higher quality protein separations are possible

[51]. Alternatively, a solid-phase extractor combined witha enzymatic microreactor [59] would be useful.

A second type of tandem column was investigatedwhere both trypsin and Glu-C were sequentially immobi-lized by repeating the photografting and immobilizationsteps in separate areas of the column. Multiple, sequen-tial, enzyme reactors have been known in the literaturesince the 1960s [68] and have recently been demon-strated in microfluidic devices [69], however, we believethat this is the first report of multiple proteolytic reac-tors being developed for proteomic applications. ACTH,which has only one cleavage site for trypsin and one for


Figure 8. Total ion current chroma-togram (top) and MS spectra (bot-tom) obtained from online cyto-chrome c and myoglobin separationfollowed by tryptic digestion. TheMS spectrum (bottom left) corre-sponds to the two peaks from 18 to20 min, and the MS spectrum (bot-tom right) represents the peak at21 min in the chromatogram. Myo-globin and cytochrome c peptidesare labeled with M and C, respec-tively. For MS analysis, the experi-mental conditions are given in Fig-ure 5. RA is the relative abundanceof the ions.

Figure 9. CE separation of ACTH and digest. From top: (1)intact ACTH, (2) ACTH digested by trypsin reactor, (3)ACTH digested by trypsin + Glu C reactor. A separation volt-age of 15 kV was applied to the 60 cm (50 cm to window)75 lm id capillary with 50 mM phosphate, pH 2.5 as BGE.Samples were injected using a 5 s, 34.5 mbar hydrodynamicinjection and separated at 288C. T1 and T2 are tryptic pepti-des; G1 and G2 are Glu-C peptides.


Glu-C, was digested after each immobilization step. Thedigests were then analyzed by both CE and MS. The CEseparation of ACTH and digest is shown in Fig. 9 wherethe two tryptic peptides were observed when only trypsinhad been immobilized. After immobilization of Glu-Cboth the tryptic and Glu-C peptides were observed. Theidentity of these peptides was confirmed by MS analysis(data not shown). This simple example demonstratesthat selective and sequential immobilization of enzymesthrough selective photografting is possible and allowsthe reactor to be tailored to specific applications. Forexample, the length of each reactor zone can be easilyadjusted to yield the optimal contact time for a givenflow rate. Other combinations are also easily achievedsuch as deglycosylation enzymes (PNGase F) followed byproteolytic digestion to simplify glycoprotein analysis.

4 Concluding remarks

Monolithic supports are suitable substrates for the syn-thesis of high performance analytical devices. With thesite-specific UV photografting technique, monoliths canbe tailored to accommodate multiple functionalitiesincluding single or multiple enzyme reactors. We havefabricated polymer-based monolithic tryptic and Glu-Creactors in fused-silica capillaries, and characterizedthem through protein digestion and identification byMALDI and ESI-MS. We have also demonstrated anapproach to immobilize both trypsin and Glu-C onto adual-functional enzyme reactor. Importantly, high diges-tion efficiency was obtained with short contact times.Furthermore, a monolithic separation and tryptic reac-tor showed that protein separation and digestion can beintegrated. With further development, high throughputseparations and digestion of a protein mixture is possiblefor “top-down” proteomics applications. To accomplishthis will require reduction of peptide retention in thedigestor and increasing the protein separation effi-ciency, possibly by exploiting electro driven flow usingthe residual AMPS functionalities.

This work was supported by the National Science and EngineeringResearch Council of Canada (NSERC). The authors would like toexpress their gratitude to Alain Tessier from CBAMS, ConcordiaUniversity for his help in MS analysis and Raymond Mineau fromUniversit� du Qu�bec � Montr�al for SEM acquisition.

The authors declared no conflict of interest

5 References[1] Gomase, V. S., Kale, K. V., Tagore, S., Hatture, S. R., Curr. Drug

Metab. 2008, 9, 213 – 220.

[2] Sleno, L., Emili, A., Curr. Opin. Chem. Biol. 2008, 12, 46 – 54.

[3] Kelleher, N. L., Anal. Chem. 2004, 76, 196A – 203A.

[4] Han, X., Aslanian, A., Yates, III, J. R., Curr. Opin. Chem. Biol. 2008,12, 483 – 490.

[5] Allen, G., Sequencing of Proteins and Peptides, Elsevier, Amsterdam1989.

[6] Washburn, M. P., Wolters, D., Yates, J. R., Nat. Biotechnol. 2001, 19,242 – 247.

[7] �vec, F., Electrophoresis 2006, 27, 947 – 961.

[8] Massolini, G., Calleri, E., J. Sep. Sci. 2005, 28, 7 – 21.

[9] Ma, J., Zhang, L., Liang, Z., Zhang, W., Zhang, Y., J. Sep. Sci. 2007,30, 3050 – 3059.

[10] Krenkov�, J., Foret, F., Electrophoresis 2004, 25, 3550 – 3563.

[11] 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.

[12] Palm, A. K., Novotny, M. V., Rapid Commun. Mass Spectrom. 2004,18, 1374 – 1382.

[13] Turkov�, J., Bl�ha, K., Malan�kov�, M., Vancurov�, D., �vec, F.,K�lal, J., Biochim. Biophys. Acta 1978, 524, 162 – 165.

[14] Bencina, K., Podgornik, A., �trancer, A., Bencina, M., J. Sep. Sci.2004, 27, 811 – 818.

[15] Hearn, W., Bethell, G. S., Ayers, J. S., Hancock, W. S., J. Chromatogr.1979, 185, 463 – 470.

[16] Temporini, C., Calleri, E., Camp�se, D., Cabrera, K., F�lix, G., Mas-solini, G., J. Sep. Sci. 2007, 30, 3069 – 3076.

[17] Kato, M., Sakai-Kato, K., Jin, H. M., Kubota, K., Miyano, H.,Toyo'oka, T., Dulay, M. T., Zare, R. N., Anal. Chem. 2004, 76, 1896 –1902.

[18] Stigter, E. C. A., de Jong, G. J., van Bennekom, W. P., Anal. Chim.Acta 2008, 619, 231 – 238.

[19] Lim, Y.-P., Josic, D., Callanan, H., Brown, J., Hixson, D. C., J. Chro-matogr. A 2005, 1065, 39 – 43.

[20] Luo, Q., Mao, X., Kong, L., Huang, X., Zou, H., J. Chromatogr. B2002, 776, 139 – 147.

[21] Geiser, L., Eeltink, S., �vec, F., Fr�chet, J. M. J., J. Chromatogr. A2008, 1188, 88 – 96.

[22] Gao, J., Xu, J., Locascio, L. E., Lee, C. S., Anal. Chem. 2001, 73,2648 – 2655.

[23] Cooper, J. W., Chen, J., Li, Y., Lee, C. S., Anal. Chem. 2003, 75,1067 – 1074.

[24] Amankwa, L. N., Kuhr, W. G., Anal. Chem. 1993, 65, 2693 – 2697.

[25] Sakai-Kato, K., Kato, M., Toyo'oka, T., Anal. Chem. 2002, 74, 2943 –2949.

[26] Sakai-Kato, K., Kato, M., Toyo'oka, T., Anal. Chem. 2003, 75, 388 –393.

[27] Calleri, E., Massolini, G., Lubda, D., Temporini, C., Loiodice, F.,Caccialanza, G., J. Chromatogr. 2004, 1031, 93 – 100.

[28] Kato, M., Inuzuka, K., Sakai-Kato, K., Toyo'oka, T., Anal. Chem.2005, 77, 1813 – 1818.

[29] Dulay, M. T., Baca, Q. J., Zare, R. N., Anal. Chem. 2005, 77, 4604 –4610.

[30] Gaber, A. M., Mersal, U. B., Electrophoresis 2005, 2303 – 2312.

[31] Palm, A. K., Novotny, M. V., Rapid Commun. Mass Spectrom. 2005,19, 1730 – 1738.

[32] Vodopivec, M., Podgornik, A., Berovic, M., �trancar, A., J. Chroma-togr. B 2003, 795, 105 – 113.

[33] Coleman, P. L., Walker, M. M., Milbrath, D. S., Stauffer, D. M., J.Chromatogr. A 1990, 512, 345 – 363.

[34] Drtina, G. J., Haddad, L. C., Rasmussen, J. K., Gaddam, B. N., Wil-liams, M. G., Moeller, S. J., Fitzsimons, R. T., Fansler, D. D., Buhl,T. L., Yang, Y. N., Weller, V. A., Lee, J. M., Beauchamp, T. J., Heil-mann, S. M., React. Funct. Polym. 2005, 64, 13 – 24.

[35] Xie, S., �vec, F., Fr�chet, J. M. J., Biotechnol. Bioeng. 1999, 62, 30 – 35.

[36] Peterson, D. S., Rohr, T., �vec, F., Fr�chet, J. M. J., Anal. Chem.2002, 74, 4081 – 4088.



[37] Hjert�n, S., Liao, J. L., Zhang, R., J. Chromatogr. 1989, 473, 273 –275.

[38] Vlakh, E. G., Tennikova, T. B., J. Sep. Sci. 2007, 30, 2801 – 2813.

[39] Urban, J., Jandera, P., J. Sep. Sci. 2008, 31, 2521 – 2540.

[40] Ye, M., Hu, S., Schoenherr, R. M., Dovichi, N. J., Electrophoresis2004, 25, 1319 – 1326.

[41] Ivanov, A. R., Zang, L., Karger, B. L., Anal. Chem. 2003, 75, 5306 –5316.

[42] Samskog, J., Bylund, D., Jacobsson, S. P., Markides, K. E., J. Chroma-togr. A 2003, 998, 83 – 91.

[43] Li, Y., Cooper, J. W., Lee, C. S., J. Chromatogr. A 2002, 979, 241 – 247.

[44] Temporini, C., Dolcini, L., Abee, A., Calleri, E., Galliano, M., Cac-cialanza, G., Massolini, G., J. Chromatogr. A 2008, 1183, 65 – 75.

[45] Krenkov�, J., Kleparnik, K., Foret, F., J. Chromatogr. A 2007, 1159,110 – 118.

[46] Ngola, S. M., Fintschenko, Y., Choi, W. Y., Shepodd, T. J., Anal.Chem. 2001, 73, 849 – 856.

[47] Hilder, E. F., �vec, F., Fr�chet, J. M. J., Anal. Chem. 2004, 76, 3887 –3892.

[48] Karas, M., Hillenkamp, F., Anal. Chem. 1988, 60, 2301 – 2303.

[49] Eeltink, S., Geiser, L., �vec, F., Fr�chet, J. M. J., J. Sep. Sci. 2007, 30,2814 – 2820.

[50] Palm, A., Novotny, M. V., Anal. Chem. 1997, 69, 4499 – 4507.

[51] Bandilla, D., Skinner, C. D., J. Chromatogr. A 2003, 1004, 167 – 179.

[52] Cabral, J.-L., Development of a New Strategy for the Modification ofMonolithic Stationary Phases Used in Capillary Electrochromatography,Concordia University, Montreal 2007, p. 216.

[53] Eeltink, S., Hilder, E. F., Geiser, L., �vec, F., Fr�chet, J. M. J., Roz-ing, G. P., Schoenmakers, P. J., Kok, W. T., J. Sep. Sci. 2007, 30,407 – 413.

[54] Rohr, T., Hilder, E. F., Donovan, J. J., �vec, F., Fr�chet, J. M. J., Mac-romolecules 2003, 36, 1677 – 1684.

[55] Migneault, I., Dartiguenave, C., Vinh, J., Bertrand, M. J., Wal-dron, K. C., Electrophoresis 2004, 25, 1367 – 1378.

[56] Buck, F. F., Vithayathil, A. J., Bier, M., Nord, F. F., Arch. Biochem.Biophys. 1962, 97, 417 – 424.

[57] Drapeau, G. R., Laszlo, L., Protease from Staphyloccus aureus, Aca-demic Press, New York 1976, pp. 469 – 475.

[58] Krenkov�, J., Bilkov�, Z., Foret, F., J. Sep. Sci. 2005, 28, 1675 – 1684.

[59] Peterson, D. S., Rohr, T., �vec, F., Fr�chet, J. M. J., Anal. Chem.2003, 75, 5328 – 5335.

[60] Ding, L., Li, Y., Jiang, Y., Cao, Z., Huang, J., J. Appl. Polym. Sci. 2002,83, 94 – 102.

[61] Huang, X. L., Catignani, G. L., Swaisgood, H. E., J. Biotechnol. 1997,53, 21 – 27.

[62] Progent, F., Augustin, V., Tran, N. T., Descroix, S., Taverna, M.,Electrophoresis 2006, 27, 757 – 767.

[63] Al-Lawati, H., Watts, P., Welham, K. J., Analyst 2006, 656 – 663.

[64] Krenkov�, J., Lacher, N. A., �vec, F., Anal. Chem. 2009, 81, 2004 –2012.

[65] Stachowiak, T. B., �vec, F., Fr�chet, J. M. J., Chem. Mater. 2006, 18,5950 – 5957.

[66] Hopp, T. P., Woods, K. R., Proc. Natl. Acad. Sci. U. S. A. 1981, 78,3824 – 3828.

[67] Gaertner, H. F., Puigserver, A. J., Enzyme Microb. Technol. 1992, 14,150 – 155.

[68] Brown, H. D., Patel, A. B., Chattopadhyay, S. K., J. Chromatogr.1968, 35, 103 – 105.

[69] Logan, T. C., Clark, D. S., Stachowiak, T. B., �vec, F., Fr�chet, J. M.J., Anal. Chem. 2007, 79, 6592 – 6598.


design and optimization of porous polymer enzymatic digestors for proteomics

Documents