Pml

Ya

b

c

a

ARRAA

KMAPCD

1

pb[acpmidifma

0d

Journal of Chromatography A, 1216 (2009) 5525–5532

Contents lists available at ScienceDirect

Journal of Chromatography A

journa l homepage: www.e lsev ier .com/ locate /chroma

reparation of polymeric monoliths by copolymerization of acrylateonomers with amine functionalities for anion-exchange capillary

iquid chromatography of proteins

un Lia, Binghe Gub, H. Dennis Tolleyc, Milton L. Leea,∗

Department of Chemistry and Biochemistry, Brigham Young University, Provo, UT 84602, USAAnalytical Sciences, The Dow Chemical Company, Midland, Michigan 48667, USADepartment of Statistics, Brigham Young University, Provo, UT 84602, USA

r t i c l e i n f o

rticle history:eceived 3 March 2009eceived in revised form 11 May 2009ccepted 15 May 2009vailable online 21 May 2009

eywords:onoliths

nion-exchange chromatographyroteinsapillary liquid chromatography

a b s t r a c t

Two novel polymeric monoliths for anion-exchange capillary liquid chromatography of proteins were pre-pared in a single step by a simple photoinitiated copolymerization of 2-(diethylamino)ethyl methacrylateand polyethylene glycol diacrylate (PEGDA), or copolymerization of 2-(acryloyloxy)ethyl trimethylam-monium chloride and PEGDA, in the presence of selected porogens. The resulting monoliths containedfunctionalities of diethylaminoethyl (DEAE) as a weak anion-exchanger and quaternary amine as a stronganion-exchanger, respectively. An alternative weak anion-exchange monolith with DEAE functionali-ties was also synthesized by chemical modification after photoinitiated copolymerization of glycidylmethacrylate (GMA) and PEGDA. Important physical and chromatographic properties of the synthe-sized monoliths were characterized. The dynamic binding capacities of the three monoliths (24 mg/mL,56 mg/mL and 32 mg/mL of column volume, respectively) were comparable or superior to values that

ynamic binding capacity have been reported for various other monoliths. Chromatographic performance was also similar to thatprovided by a modified poly(GMA-ethylene glycol dimethacrylate) monolith. Separation of standard pro-teins was achieved under gradient elution conditions using these monolithic columns. Peak capacities of34, 58 and 36 proteins were obtained with analysis times of 20–30 min. This work represents a success-ful attempt to prepare functionalized monoliths via direct copolymerization of monomers with desired

to eto im

functionalities. ComparedPEGDA crosslinker helped

. Introduction

Anion-exchange chromatography is often used for analyzingroteins because it is typically performed under neutral or slightlyasic pH conditions where proteins remain in their native states1–3]. The technique is based on the reversible binding of neg-tively charged functional groups on biomolecules to positivelyharged groups in the stationary phase and their subsequent dis-lacement by either increasing the ionic strength or pH of theobile phase. Commonly used anion-exchangers are functional-

zed with amines [4–6]. Typically, weak anion-exchangers containiethylaminoethyl (DEAE) or polyethyleneimine (PEI) functional-

ties, while strong anion-exchangers are quaternary amines. Asor stationary phase support materials, silica and organic poly-

ers have replaced polysaccharides (such as cellulose, agarosend dextran) because of their excellent mechanical stabilities [7,8].

∗ Corresponding author. Fax: +1 801 422 0157.E-mail address: milton lee@byu.edu (M.L. Lee).

021-9673/$ – see front matter © 2009 Elsevier B.V. All rights reserved.oi:10.1016/j.chroma.2009.05.037

arlier publications, additional surface modifications were avoided and theprove the biocompatibility of the monolithic backbone.

© 2009 Elsevier B.V. All rights reserved.

Additionally, polymers, such as highly cross-linked polystyrenedivinylbenzene (PSDVB), are popular because they have a wider pHstability range than silica, while possessing comparable mechanicalstability if highly cross-linked [8,9].

Monolithic stationary phases have grown in interest becauseof their characteristic enhanced mass transfer and simple prepa-ration. Polymeric monoliths, in particular, offer the advantagethat by choosing the right functional monomer, monoliths forvarious chromatographic applications can be designed, such asreversed-phase and affinity chromatography [10,11]. DEAE, whichis used to prepare weak anion-exchangers (WAX), was intro-duced into a monolithic stationary phase after thermally initiatedpolymerization of a rod of poly(glycidyl methacrylate-co-ethyleneglycol dimethacrylate) (GMA-co-EDMA) by reaction of the epoxygroups on the rod with diethylamine (DEA) [12,13]. Similarly, a

PEI modified weak anion-exchanger was also prepared based on apoly(GMA-co-EDMA) monolith [14]. In another report, WAX func-tionalities were introduced into poly(GMA) chains grafted onto aporous-fiber membrane by radiation-induced graft polymerizationand subsequent functionalization [15]. Based on poly(GMA-co-

5 gr. A 1

DwscamocstmapaA([c

iim((hrcsgtfaa

pmadhac

526 Y. Li et al. / J. Chromato

VB) monoliths, DEAE or quaternary ammonium functionalitiesere introduced to provide WAX or strong anion-exchange (SAX)

tationary phases for the separation of oligonucleotides, or oligonu-leotides and nucleosides, respectively [16,17]. Acetonitrile wasdded to the mobile phase to suppress the hydrophobic interactionainly from divinylbenzene. By this surface modification approach,

nly a single functionality could be introduced. Additionally, whenonducting the in situ reaction, in order to achieve maximumurface coverage and high reproducibility, special care had to beaken with respect to compatibility of the solvent adsorbed on the

atrix surface, reaction temperature, and reaction time. In anotherpproach, an SAX stationary phase was prepared by coating aorous poly(butyl methacrylate-co-ethylene dimethacrylate-co-2-crylamido-2-methyl-1-propanesulfonic acid) (BMA-co-EDMA-co-MPS) monolith by electrostatic attachment of quaternary amine

QA)-functionalized latex particles for the separation of saccharides18]. However, pore size and flow permeability were decreased afteroating with a layer of latex particles.

Surface grafting is another effective approach for introduc-ng ionizable groups onto the surface of monoliths, which alsonvolves two main steps, matrix formation and functionality attach-

ent. For example, anion-exchange polymer chains of poly[2-dimethylamino)ethyl methacrylate] (pDMAEMA) and poly[(2-methacryloyloxy)ethyl trimethylammonium chloride] (pMETAC)ave been grafted onto macroporous polyacrylamide gels by free-adical polymerization. These monoliths achieved protein-bindingapacities of up to 6–12 mg/mL, but no separation was demon-trated [19]. The ionizable monomers AMPS and METAC were alsorafted onto macroporous polymer monoliths for fabricating elec-roosmotic pumps [20]. Surface grafting can introduce multipleunctionalities onto a single monolith. However, this approach usu-lly suffers from low grafting efficiency, swelling of grafted polymernd, therefore, reduced permeability of the final monolith [21].

Direct copolymerization of functional monomers undoubtedlyrovides the simplest approach for preparation of functionalizedonoliths. Using this technique, Gu et al. designed and synthesizedseries of strong cation-exchange (SCX) monoliths by incorporating

ifferent sulfonic acid-functionalized monomers with decreasingydrophobicity [22,23]. High-performance separations of peptidesnd proteins were achieved using these columns. A poly(AMPS-o-PEGDA) monolith exhibited a dynamic binding capacity (DBC)

Fig. 1. Reaction scheme and monomers involved in

216 (2009) 5525–5532

that exceeded 150 �equiv/mL of peptides or 300 mg/mL of proteins,which were much higher than for an AMPS grafted poly(GMA-co-EDMA) monolith [24]. However, the preparation of such monolithsusing direct copolymerization of functional monomers is notalways successful because of difficulties experienced in finding aporogen system that dissolves all of the monomers while enablingthe formation of monoliths with desired pore structure. Despiteadvances in the theoretical understanding of polymerization pro-cesses [25,26], the selection of porogens still remains largelyempirical.

Minimal non-specific interaction between analytes and station-ary phase is very desirable in order to keep proteins in their nativestates and to allow reversible adsorption [27,28]. Polyethylene gly-col (PEG) is a hydrophilic, biocompatible polymer. Surface-boundPEG can resist the non-specific adsorption of proteins and otherbiological species [29], which has been achieved by incorporat-ing PEG monomers into polymer networks by graft polymerization[30], or by attaching PEG chains directly to polymer surfaces byvarious coupling reactions [31]. Recently, reaction of PEG function-alized monomers, such as polyethylene glycol methyl ether acrylate(PEGMEA) and polyethylene glycol diacrylate (PEGDA) yieldedmonoliths that demonstrated negligible non-specific adsorption ofproteins [32]. Krenkova et al. reported the grafting of poly(GMA-co-EDMA) monolith with poly(ethylene glycol) methacrylate beforeattaching reactive azlactone functionalities for enzyme immobi-lization in order to minimize non-specific adsorption of proteinsand peptides [33].

Based on previous work in our laboratory [22,23], thedesign of successful functionalized monoliths has followed twomain strategies: (1) incorporation of monomers that simul-taneously provide functionalities with high selectivity, and(2) application of a biocompatible crosslinker (PEGDA) thatforms monoliths with decreased non-specific adsorption. In thispaper, two functional monomers containing anion-exchange func-tionalities, 2-(diethylamino)ethyl methacrylate (DEAEMA) and2-(acryloyloxy)ethyl trimethylammonium chloride (AETAC) wereutilized, as shown in Fig. 1. Two novel anion-exchange poly-

meric monoliths were prepared by direct copolymerization ofDEAEMA and PEGDA, and copolymerization of AETAC and PEGDA.The resulting monoliths contained DEAE (a WAX group) and QA(a SAX group), respectively. Both demonstrated comparable chro-

synthesizing the anion-exchange monoliths.

gr. A 1216 (2009) 5525–5532 5527

mfbscif

2

2

amGw(tsctpfi

bsch[

2

hgt[bppTfrircwrcPwmluTcep

mio

e

siti

onan

dp

hysi

calp

rop

erti

esof

mon

olit

hs

syn

thes

ized

inth

isst

ud

y.

eage

nt

com

pos

itio

nPh

ysic

alp

rop

erti

es

MPA

)M

onom

er(g

)PE

GD

A(g

)C

yclo

hex

anol

(g)

Met

han

ol(g

)Et

hyle

ther

(g)

Hex

anes

(g)

UV

tim

e(m

in)

DB

C(m

g/m

L)R

ecov

ery

(%)

Mon

omer

conv

sn(w

t%)

Perm

eabi

lity

(×10

−14m

2)Po

rosi

ty(%

)

.01a

0.4

80.

32a

1.10

0.10

a/

//a

3689

–1.

163

.008

0.4

80.

321.

100.

10/

/10

3210

283

1.9

69.0

080.

48

0.32

//

/0.

4010

2498

48

6.2

82.0

008

0.40

0.4

8/

0.90

0.90

/5

5610

010

01.

059

o-ED

MA

)w

asp

rep

ared

asa

refe

ren

cem

ater

ial;

0.01

gA

IBN

was

use

das

init

iato

r,0.

10g

dod

ecan

olw

asu

sed

asco

-por

ogen

,an

d22

hh

eat

exp

osu

rew

asap

pli

ed[1

2,13

].

Y. Li et al. / J. Chromato

atographic properties and did not require additional surfaceunctionalization. An alternative WAX monolith was also obtainedy copolymerization of GMA and PEGDA (replacing EDMA) andubsequent functionalization. These anion-exchange monolithicolumns were prepared in a capillary format using 75 �m i.d. cap-llaries by photoinitiated polymerization and then were evaluatedor LC separation of standard proteins.

. Experimental

.1. Chemicals

2,2-Dimethoxy-2-phenyl-acetophenone (DMPA, 99%), 2,2′-zobisisobutyronitrile (AIBN, 98%), 3-(trimethoxysilyl)propylethacrylate (TMSPMA, 98%), PEGDA (Mn ∼258), EDMA (98%),MA (>97%), DEAEMA (99%), and AETAC (80 wt. % solution inater) were purchased from Sigma–Aldrich (Milwaukee, WI). DEA

99+%) was obtained from Acros Organics (Morris Plains, NJ). All ofhe monomers were used without further purification. Porogenicolvents, cyclohexanol, methanol, ethylether, and hexanes, andhemicals for use in mobile phases were HPLC grade. Buffer solu-ions were prepared with deionized water from a Millipore waterurifier (Billerica, MA) and filtered through a 0.22-�m membranelter.

Proteins, myoglobin (MYO) from equine skeletal muscle, conal-umin (CON) from chicken egg white, ovalbumin (OVA), bovineerum albumin (BSA), and soybean trypsin inhibitor (STI) were pur-hased from Sigma–Aldrich. The total protein from lysate of theost Escherichia coli DH5� was obtained as described previously34].

.2. Polymer monolith preparation

First, a fused-silica capillary with UV-transparent fluorinatedydrocarbon polymer (75 �m i.d., 375 �m o.d., Polymicro Technolo-ies, Phoenix, AZ) was silanized using TMSPMA in order to anchorhe polymer monolith on the capillary wall as previously described23]. In this study, PEGDA was chosen as the only crosslinkerecause of its effectiveness in reducing non-specific adsorption ofroteins on the monolith. Then the polymer precursor was pre-ared by mixing monomer, crosslinker, porogens, and initiator (seeable 1 for reagent compositions). The mixture was ultrasonicatedor ∼20 s considering the volatility of ethyl ether. Subsequently, theeaction mixture was introduced into the silanized capillary by cap-llary action, and irradiated using a UV curing system as previouslyeported [22]. After polymerization was complete, the monolithicolumn was connected to an HPLC pump and extensively flushedith methanol to remove the porogenic solvents and unreacted

eagents. As mentioned in the introduction, DEAE functionalitiesould be introduced after polymerization of the rod of poly(GMA-EGDA) by reaction of the epoxy groups with DEA. Typically, DEAas pumped through the column to completely “wet” the poly-er first, and then the column was sealed at both ends by a closed

oop filled with DEA, and heated at 55 ◦C for 12 h in an oven. Resid-al epoxy groups were blocked by washing the column with 1.0 Mris buffer, pH 8.0. After preparation, both ends of the monolithicolumn were placed in water to keep the monolith wet. For ref-rence, a DEA modified poly(GMA-co-EDMA) monolith was alsorepared.

Monolithic columns were characterized by scanning electron

icroscopy (SEM, FEI Philips XL30 ESEM FEG, Hillsboro, OR). SEM

mages of the monolith inside the capillary provided informationn the pore structure, i.e., approximate globule and pore sizes.

Monomer conversion is an important measurement used tovaluate polymerization efficiency. To measure monomer conver- Ta

ble

1R

eage

nt

com

po

Mon

olit

hR D (g

GM

Aa

0G

MA

0D

EAEM

A0

AET

AC

0

aPo

ly(G

MA

-c

5 gr. A 1

scemast

2

pmtHT(laFtBdatcg

p(rov

2

c

F

528 Y. Li et al. / J. Chromato

ion, a bulk monolith was prepared under the same polymerizationonditions as used to prepare the same monolith in a capillaryxcept that the bulk monolith was prepared in a mold [23]. Theonolith formed was Soxhlet extracted with methanol overnight,

nd then dried under vacuum at ∼60 ◦C for 5 h. Monomer conver-ion was calculated by comparing the weight of polymer monolitho the weight of the total monomers.

.3. Capillary liquid chromatography (CLC)

Chromatographic separations were carried out using a systemreviously described [22]. It consisted of two syringe pumps (ISCOodel 100 DM, ISCO, Lincoln, NE), a static mixer, a Valco splitting

ee, and a Valco sample injector with an internal 60-nL loop (Valco,ouston, TX). The mobile phases in the two pumps were 10 mMris buffer, pH 7.6 (buffer A), and sodium chloride plus buffer Abuffer B), respectively. The splitter had a 33-cm fused-silica capil-ary (30 �m i.d., 375 �m o.d.) connected to its third outlet to provide∼1:1000 split ratio between the analytical column and waste.

low rates were measured at the exits of the column and the split-er capillary using empty Hamilton syringes (Hamilton, Reno, NV).oth pumps were operated in the constant flow mode to deliver theesired gradient program, and they were set to deliver solvents atconstant flow rate in the range of 70–200 nL/min. Standard pro-

eins dissolved in buffer A were used for testing these monolithicolumns. On-line UV detection at 214 nm was used. The chromato-raphic conditions are given in the figure captions.

To investigate the permeability of these monolithic columns,ressure drop measurements were made at room temperature∼23 ◦C) using pure water as the permeating fluid at flow ratesanging from 50 nL/min to 200 nL/min. Each column pressure wasbtained by subtracting the system pressure from the measuredalue on the read-out display.

.4. DBC and mass recovery measurements

DBC is one of the most important properties of an ion-exchangeolumn, which determines the resolution, column loadability, and

ig. 2. SEM photographs of (A) DEA modified poly(GMA-co-EDMA); (B) DEA modified po

216 (2009) 5525–5532

gradient elution strength. The DBC was examined via open-loopfrontal analysis [35]. Specifically, the monolithic column was firstequilibrated with buffer A, and then a solution of 3.0 mg/mL BSAin buffer A that had been preloaded in a sample loop (∼3 m long,320 �m i.d.) was pumped through the monolithic column (∼10 cmlong, 75 �m i.d.) at a flow rate of 150 nL/min. One end of the sampleloop was connected to the pump and the other end was connectedto the column using a True ZDV Union (Upchurch Scientific, OakHarbor, WA). The volume of protein solution needed to saturatethe column was measured by a breakthrough curve which could berecorded directly by monitoring the UV absorbance. Protein massrecovery was determined by a direct comparison of the protein peakareas obtained with and without the analytical column as describedpreviously [36].

2.5. Safety considerations

The GMA, DEAEMA and AETAC monomers, and the PEGDA andEDMA crosslinkers may act as sensitizers. Additionally, DEA is asevere eye and respiratory tract irritant. Relative MSDS informationabout these reagents should be consulted and special care shouldbe taken for handling these reagents. Caution must be taken (suchas wearing glasses and gloves) during polymerization steps, sincehigh power UV radiation can cause severe burns to skin and eyes.

3. Results and discussion

3.1. Polymer monolith preparation

Various monomers (DEAEMA, AETAC and GMA) containinganion-exchange groups or reactive groups were selected for copoly-merization with the biocompatible crosslinker, PEGDA. Usually,

good solvents serve as microporogens to provide high surface area,while poor solvents act like macroporogens to provide good bulkflow properties. Whether a solvent is good or poor mostly dependson its relative polarity compared to the polymer. Polymerizationconditions optimized for one system cannot be transferred directly

ly(GMA-co-PEGDA); (C) poly(DEAEMA-co-PEGDA); (D) poly(AETAC-co-PEGDA).

gr. A 1

ti

wp(mimfmcims(E

smpHpetftr

cflmasgparomws

csc(

Fp32

Y. Li et al. / J. Chromato

o another since different monomers with different polarities arenvolved.

The reference monolith poly(GMA-co-EDMA) reported by Svecas chosen as a starting point [12,13]. This monolith was pre-ared in a mixture of cyclohexanol (microporogen) and dodecanolmacroporogen) by thermally initiated polymerization. The ratio of

onomer to crosslinker was fixed at 3:2 (Table 1). When replac-ng EDMA with PEGDA as crosslinker, methanol was selected as the

acroporogen instead of dodecanol since it was proven efficientor the formation of macroporous throughpores in a poly(PEGDA)

onolith. As a result, the new poly(GMA-co-PEGDA) monolithould be formed in 10 min using photoinitiated polymerizationnstead of thermally initiated polymerization. An SEM image of the

odified poly(GMA-co-PEGDA) monolith inside a capillary (Fig. 2B)howed a similar porous structure and somewhat larger globule∼2 �m) and pore sizes (∼3 �m), compared to the poly(GMA-co-DMA) monolith (Fig. 2A).

Ethyl ether was found to be another effective porogen forynthesizing monoliths from PEG-based monomers. With opti-ized combination of methanol and ethyl ether as porogens, a

oly(AETAC-co-PEGDA) monolith was formed as shown in Fig. 2D.owever, in this polymerization system, methanol was a micro-orogen, while ethyl ether was a macroporogen. Absence ofthylether or excess methanol led to a glassy, transparent struc-ure. A favorable mass ratio (i.e., 2:3) of AETAC to PEGDA was foundor this monolithic column. This polymerization reaction was so fasthat only 0.0008 g (∼0.03 wt. %) DMPA was enough for initiating theeaction.

Unfortunately, neither methanol nor ethyl ether, nor otherombinations of common solvents were successful in producingow-through poly(DEAEMA-co-PEGDA) monoliths. A flow-throughonolith was finally formed using a single porogen, hexanes. There

re two possible reasons for this result. One is that most polarolvents cannot be used because they are such good microporo-ens that poor flow-through monoliths are produced even in theresence of a macroporogen. The second reason is that unre-cted monomers can serve in part as microporogens since they areelatively good solvents for the polymer compared to the macrop-rogen. The application of a single macroporogen was also in agree-ent with the SEM image of poly(DEAEMA-co-PEGDA) (see Fig. 2C),hich demonstrated large globule (∼3.6 �m) and pore (∼5 �m)

izes and, consequently, much higher permeability (Table 1).

From these images, it can be immediately seen that poly(GMA-

o-PEGDA) and poly(DEAEMA-co-PEGDA) have structures that areimilar to poly(GMA-co-EDMA) reported earlier, i.e., clusters ofongregated irregular micro-globules with voids between thempores). However, the morphology of the poly(AETAC-co-PEGDA)

ig. 3. Anion-exchange chromatography of standard proteins. Conditions: stationary phasoly(DEAEMA-co-PEGDA), and (C) 75 �m i.d. × 17 cm poly(AETAC-co-PEGDA); mobile ph0 min linear gradient from 0 M to 0.5 M NaCl (17 mM/min) in buffer A, and (C) 30 min14 nm; Peaks, 1, MYO; 2, OVA; 3, STI; 4, BSA; 5, CON.

216 (2009) 5525–5532 5529

monolith was very different from the others, but quite similar tothe SCX poly(vinylsulfonic acid-co-PEGDA) monolith synthesizedpreviously by us [23], i.e., a labyrinth of tortuous interconnectedcavities of different sizes formed by aggregated chains of particles.

3.2. Biocompatibility

In two previous papers, clean fluorescence images obtained afterflushing PEGDA-based monoliths with fluorescein isothiocyanate(FITC)-labeled BSA were demonstrated to show negligible non-specific adsorption of proteins [32]. Furthermore, the hydropho-bicities of the monoliths were compared under reversed-phasechromatography conditions, yielding retention factors of almostzero for the poly(PEGDA) monolith and 4.4 for the poly(EDMA)monolith [23]. To further evaluate non-specific protein adsorption,protein mass recoveries were examined using STI. A recovery ofmore than 98% was obtained using monoliths incorporating PEGDAas crosslinker, compared to a recovery of 89% when using modifiedpoly(GMA-co-EDMA) (Table 1). These tests indicate that polymermonoliths that incorporate PEGDA as crosslinker are more biocom-patible.

3.3. Chromatographic performance

The anion-exchange monoliths reported in this study demon-strated efficient separations of proteins of varying molecularweights and pI values. The modified poly(GMA-co-PEGDA) mono-lith was able to separate OVA, BSA, and STI, which differ by as littleas ∼0.1 pI value (Fig. 3A). In isoelectric focusing (IEF), conalbuminseparates into 3 major components with pI values of 6.2, 6.6 and 7.2,representing different metalloforms (diferric, monoferric and ion-free, respectively). Similarly, the WAX poly(DEAEMA-co-PEGDA)monolith showed similar separation of conalbumin (Fig. 3B), whichindicates its potential for separating protein isoforms or variants.Compared to other monoliths, faster separation of standard pro-teins was achieved using the poly(DEAEMA-co-PEGDA) monolithdue to a much lower back pressure (data not shown). The SAXpoly(AETAC-co-PEGDA) monolith also demonstrated good sepa-ration of standard proteins, MYO, OVA and STI (Fig. 3C). Bothpoly(GMA-co-PEGDA) and poly(AETAC-co-PEGDA) afforded higherretention and resolution compared to poly(DEAEMA-co-PEGDA)even under stronger elution strength, which was presumably

due to their higher binding capacities for proteins. Furthermore,poly(AETAC-co-PEGDA) provided better protein separations due toa higher peak capacity of 58 compared to peak capacities of 34and 36 for poly(DEAEMA-co-PEGDA) and poly(GMA-co-PEGDA),respectively.

e, (A) 75 �m i.d. × 13 cm DEA modified poly(GMA-co-PEGDA), (B) 75 �m i.d. × 16 cmase, (A) 20 min linear gradient from 0 M to 1.0 M NaCl (50 mM/min) in buffer A, (B)linear gradient from 0 M to 1.0 M NaCl (33 mM/min) in buffer A; detection, UV at

5530 Y. Li et al. / J. Chromatogr. A 1216 (2009) 5525–5532

Table 2Comparison of DBCs of various anion-exchangers.

WAX media Data source DBC (mg/mL) SAX media Data source DBC (mg/mL)

DEAE-GMA-PEGDA This work 32 AETAC-PEGDA This work 56DEAEMA-PEGDA This work 24 BIA CIM-QA [38] 30DEAE-GMA-EDMA [12] 35 Bio-Rad UNO-Q [38] 40BIA CIM-DEAE [37] 20–30 TSK-Gel Q-5PW-HR [40] 42PEI-GMA-EDMA [14] 14 Q Sepharose FF [40] 54DMAEMA-pAAm cryogels [19] 6–12 Q HyperD 20 [40] 85T 6D 76

3

pa3tttairp1orlDl

mcftFo5buc2tsiD

csahaotiit

3

ntr

Fig. 4. Breakthrough curves obtained by fontal analysis. Conditions: 75 �m

3000 psi. Calculated permeabilities according to Darcy’s Law arelisted in Table 1 [42]. The permeability of the poly(GMA-co-EDMA)monolith was 1.1 × 10−14 m2, which is in good agreement with acommercial CIM DEAE disk monolithic column from BIA Separa-

Table 3Reproducibilities of the new monolithic columnsa.

Reproducibilities Run-to-run Column-to-column

(RSD% of retention time)

oyopearl DEAE 650M, MacroPrep DEAE [39] 20,1EAE Sepharose FF, DEAE Ceramic HyperD 20 [39] 59,

.4. Dynamic binding capacity (DBC)

Using frontal analysis, the DBC values for the two WAXoly(GMA-co-PEGDA) and poly(DEAEMA-co-PEGDA) monoliths,nd one SAX poly(AETAC-co-PEGDA) monolith were measured as2 mg/mL, 24 mg/mL and 56 mg/mL of column volume, respec-ively. These data are comparable or superior to the DBC valueshat have been reported for various monoliths (Table 2). The DBC ofhe modified poly(GMA-co-EDMA) monolith developed by Svec etl. was 35 mg/mL [12]. Hahn et al. reported BSA binding capacitiesn the range of 20–30 mg/mL on CIM-DEAE disks from BIA Sepa-ations, which were presumably fabricated from ∼1 �m diameterarticles [37]. Our results are also much higher than the DBC of4 mg/mL reported for a PEI modified weak ion-exchanger basedn a poly(GMA-co-EDMA) monolith [14], or 6–12 mg/mL that waseported for pDMAEMA-grafted pAAm cryogels [19]. For SAX mono-iths, Iberer et al. reported a BSA DBC of ∼30 mg/mL for a BIA CIM-QAisk and ∼40 mg/mL for a Bio-Rad UNO-Q monolith [38], which are

ower than the value of 56 mg/mL for poly(AETAC-PEGDA).The DBC values of 24 mg/mL and 32 mg/mL for the WAX

onoliths in this work are superior to those measured for someommercial anion-exchange resins (e.g., 20 mg/mL and 16 mg/mLor Toyopearl DEAE 650 M and MacroPrep DEAE Support, respec-ively), but lower than 59 mg/mL and 76 mg/mL for DEAE SepharoseF and DEAE Ceramic HyperD 20, respectively [39]. The DBC valuef 56 mg/mL for the SAX monolith is superior to 42 mg/mL and4 mg/mL for TSK-Gel Q-5PW-HR and Q Sepharose FF, respectively,ut lower than 85 mg/mL for Q HyperD 20 [40]. The lower val-es could be due to less solid phase in the column. For particulatehromatographic media, the total mass of solid is approximately–3 times that of a highly porous monolith [41]. The good DBCs ofhese monoliths are attributed, in part, to their high-through poretructures that afford good pore accessibility by proteins. A highncorporation of functional monomer could also contribute to highBC.

The sharpness of the breakthrough curve is indicative of the effi-iency of mass transfer within the pores of the separation media. Ashown in Fig. 4, all three new anion-exchange monolithic columnschieved very steep breakthrough slopes (b, c, and d) and, hence,ighly efficient binding of proteins that provides efficient peaksnd high resolution. By careful observation, the breakthrough curvef the modified poly(GMA-co-PEGDA) monolith is slightly steeperhan that of the modified poly(GMA-co-EDMA) monolith, whichndicates faster kinetic adsorption of proteins on the former. Thiss attributed to decreased non-specific interaction of proteins withhe monolith backbone when using PEGDA as crosslinker.

.5. Column characterization

Run-to-run and column-to-column reproducibilities of the threeew monoliths were investigated (Table 3). For three consecu-ive runs, the percent relative standard deviations (RSD%) of theetention times of OVA and STI were less than 1.4%, 1.2%, and 2.0%

i.d. × 10 cm stationary phase, (a) DEA modified poly(GMA-co-EDMA), (b) DEAmodified poly(GMA-co-PEGDA), (c) poly(DEAEMA-co-PEGDA), and (d) poly(AETAC-co-PEGDA) monolithic columns, loading phase, 3.0 mg/mL BSA in buffer A; flow rate,0.15 �L/min; detection, UV at 214 nm.

for the modified poly(GMA-co-PEGDA), poly(DEAEMA-co-PEGDA)and poly(AETAC-co-PEGDA) monoliths, respectively, which indi-cates that reproducible separations can be achieved using thesemonolithic capillary columns. Column-to-column reproducibilitymeasurements gave retention time RSD% values (n = 3) ranging from1.7% to 3.8%. Additionally, these monolithic capillary columns werestable during continuous use for at least two months with constantback pressure. Of course, this is dependent on good attachment ofthe monolith to the capillary wall and keeping the monolith filledwith buffer solution.

Permeability represents the ease or difficulty of mobile phaseflow-through a chromatographic column. Pure water was pumpedthrough the monoliths to measure their permeabilities since onlyaqueous buffers were used as mobile phases in this work. Good lin-earity between pressure drop and flow rate clearly demonstratedthat the monoliths were mechanically stable at pressures up to

OVA STR OVA STI

Poly(GMA-co-PEGDA) 0.9 1.4 2.0 2.9Poly(DEAEMA-co-PEGDA) 0.3 1.2 1.7 2.3Poly(AETAC-co-PEGDA) 2.0 0.9 3.8 3.0

a For HPLC conditions, see Fig. 3.

Y. Li et al. / J. Chromatogr. A 1216 (2009) 5525–5532 5531

Ff2

tpipPsiipw8mpmpg

tmilD

pmurlwltpi

3

tmsOc

Fig. 6. Strong anion-exchange (SAX) chromatography of total protein lysate of E. coliDH5�. Conditions: 75 �m i.d. × 18 cm poly(AETAC-co-PEGDA) monolithic column;

ig. 5. Back pressure dependency on linear velocity for the AETAC and DEAEMA

unctionalized monoliths. Conditions: 15 cm × 75 �m i.d. columns; mobile phase,0 mM Tris buffer with pH 7.0, 8.0 and 9.0.

ions [43]. The poly(GMA-co-PEGDA) monolith, and especially theoly(DEAEMA-co-PEGDA) monolith, gave even higher permeabil-

ties due to larger globule and pore sizes. A ∼5 fold increase inermeability makes fast separations using the poly(DEAEMA-co-EGDA) monolith possible. In addition, since these three monolithshare similar pore structures, their permeabilities did increase withncrease in globule and pore sizes (Fig. 2). Furthermore, as shownn Fig. 5, a linear dependency of flow rate on back pressure foroly(AETAC-co-PEGDA) and poly(DEAEMA-co-PEGDA) monolithsas also achieved using different pH aqueous Tris buffers (pH = 7.0,.0 and 9.0), which reveals minimal swelling of these functionalonoliths in aqueous buffers within this pH range. Increasing the

H from 7.0 to 9.0 increased the permeability of these functionalonoliths. This observation is consistent with reported data and is

robably a consequence of the lower association of anion-exchangeroups [44].

The monomer conversions of these monoliths were rela-ively high, except for poly(DEAEMA-co-PEGDA) which gave 48%

onomer conversion (Table 1). This is likely due to poor copolymer-zation resulting from unfavorable monomer reactivity ratio. Such aow monomer conversion may be responsible for its relatively lowBC (24 mg/mL), and probably its high permeability.

Using inverse size-exclusion chromatography (ISEC), totalorosities were estimated from retention volumes of an unretainedolecular tracer (uracil, MW = 112 g/mol) and the geometrical vol-

mes of the columns [45]. The porosity does not seem to be directlyelated to permeability. A high porosity combined with relativelyarge volumetric through-pore fraction yields monolithic columns

ith high permeability. Unfortunately, the PEGDA-based mono-iths usually crack or deform during drying, which likely changeshe original pore structure. Because of this, mercury intrusionorosimetry was abandoned for investigating the pore properties

n our work.

.6. Application

SAX chromatography of the total protein fraction from a lysate of

he host E. coli DH5� was achieved using a poly(AETAC-co-PEGDA)

onolith (Fig. 6). Although these peaks were not identified, theeparation power of the monolithic column was demonstrated.f course, only acidic proteins could be separated using thisolumn.

[

[[

mobile phase, buffer A was 20 mM Tris (pH 7.6) and buffer B was buffer A plus 0.5 MNaCl, linear AB gradient from 2% to 98% B in 30 min, and held for 5 min; flow rate,100 nL/min; analyte concentration, 1.0 mg/mL; detection, UV at 214 nm; the baselinedrift was mostly due to the difference in UV absorbances of buffers A and B.

4. Conclusions

Two porous monolithic columns were prepared by directcopolymerization of functional monomers with a biocompati-ble crosslinker, PEGDA, in 75 �m i.d. capillaries in the presenceof suitable porogens. They were then evaluated as stationaryphases in anion-exchange capillary LC of proteins. The resultingpoly(DEAEMA-co-PEGDA) and poly(AETAC-co-PEGDA) monolithscontain DEAE and QA functionalities that are required for WAXand SAX chromatography, respectively. Even without additionalsurface functionalization, these monoliths produced comparablechromatographic efficiencies and resolution for standard proteinscompared to most anion-exchange monoliths reported previously.Compared to the more common poly(GMA-co-EDMA) monolith,the poly(GMA-co-PEGDA) and poly(DEAEMA-co-PEGDA) mono-liths were synthesized faster by photoinitiated polymerization(10 min), provided comparable DBC, demonstrated faster kineticadsorption of proteins, and provided higher permeability. The SAXpoly(AETAC-co-PEGDA) monolith provided partial separation of acomplex protein mixture, mostly due to its superior binding capac-ity and separation efficiency.

Acknowledgment

This work was funded by Berkeley HeartLab (Burlingame, CA).

References

[1] M.P. Henry, in: R.W.A. Oliver (Ed.), HPLC of Macromolecules: A PracticalApproach, IRL Press at Oxford University Press, Oxford, 1989, p. p91.

[2] P. Stanton, in: M.-I. Aguilar (Ed.), HPLC of Peptides and Proteins: Methods andProtocols, Humana Press, Totowa, NJ, 2004, p. p23.

[3] S.H. Chang, R. Noel, F.E. Regnier, Anal. Chem. 48 (1976) 1839.[4] J. Horvath, E. Boschetti, L. Guerrier, N. Cooke, J. Chromatogr. A 679 (1994) 11.[5] E. Boschetti, J. Chromatogr. A 658 (1994) 207.[6] D. Josic, J. Reusch, J. Chromatogr. 590 (1992) 59.[7] A.J. Alpert, F.E. Regnier, J. Chromatogr. 185 (1979) 375.[8] M.A. Rounds, W.D. Rounds, F.E. Regnier, J. Chromatogr. 397 (1987) 25.

[9] N.D. Marlin, N.W. Smith, Anal. Bioanal. Chem. 382 (2005) 493.10] A. Premstaller, H. Oberacher, C.G. Huber, Anal. Chem. 72 (2000) 4386.

[11] J.M. Armenta, B. Gu, P.H. Humble, C.D. Thulin, M.L. Lee, J. Chromatogr. A 1097(2005) 171.

12] F. Svec, J.M.J. Fréchet, Biotechnol. Bioeng. 48 (1995) 476.13] D. Sykora, F. Svec, J.M.J. Fréchet, J. Chromatogr. A 852 (1999) 297.

5 gr. A 1

[[[[[[[

[

[[[[[[[[

[[[[[[[[

[[

[

532 Y. Li et al. / J. Chromato

14] M. Wang, J. Xu, X. Zhou, T. Tan, J. Chromatogr. A 1147 (2007) 24.15] I. Koguma, K. Sugita, K. Saito, T. Sugo, Biotechnol. Prog. 16 (2000) 456.16] C.P. Bisjak, R. Bakry, C.W. Huck, G.K. Bonn, Chromatographia 62 (2005) S31.17] W. Wieder, C.P. Bisjak, R. Bakry, C.W. Huck, G.K. Bonn, J. Sep. Sci. 29 (2006) 2478.18] E.F. Hilder, F. Svec, J.M.J. Fréchet, J. Chromatogr. A 1053 (2004) 101.19] I.N. Savina, I.Y. Galaev, B. Mattiasson, J. Mol. Recognit. 19 (2006) 313.20] J.A. Tripp, F. Svec, J.M.J. Fréchet, S. Zeng, J.C. Mikkelsen, J.G. Santiago, Sens.

Actuators B 99 (2004) 66.21] V. Frankovic, A. Podgornik, N.L. Krajnc, F. Smrekar, P. Krajnc, A. Strancar, J. Chro-

matogr. A 1207 (2008) 84.22] B. Gu, Z. Chen, C.D. Thulin, M.L. Lee, Anal. Chem. 78 (2006) 3509.23] B. Gu, Y. Li, M.L. Lee, Anal. Chem. 79 (2007) 5848.24] C. Viklund, F. Svec, J.M.J. Fréchet, K. Irgum, Biotechnol. Prog. 13 (1997) 597.25] O. Okay, Progr. Polym. Sci. 25 (2000) 711.

26] J. Courtois, E. Byström, K. Irgum, Polymer 47 (2006) 2603.27] C. Calmon, React. Polym. 1 (1982) 3.28] J.D. Andrade, V. Hlady, Adv. Polym. Sci. 79 (1986) 1.29] J.M. Harris, S. Zalipsky, Poly(ethylene glycol): Chemistry and Biological Applica-

tions, in: ACS Symposium Series 680, American Chemical Society, Washington,D.C, 1997.

[

[[

[

216 (2009) 5525–5532

30] Y.H. Su, W.R. Gomboltz, Polym. Prep. 28 (1987) 292.31] S.W. Lee, P.E. Laibinis, Biomaterials 19 (1998) 1669.32] B. Gu, J.M. Armenta, M.L. Lee, J. Chromatogr. A 1079 (2005) 382.33] J. Krenkova, N.A. Lacher, F. Svec, Anal. Chem. 81 (2009) 2004.34] P.S. Lee, K.H. Lee, Biotechnol. Bioeng. 84 (2003) 801.35] J. Jacobson, J. Frenz, C. Horváth, J. Chromatogr. 316 (1984) 53.36] Y. Yang, K. Harrison, J. Kindsvater, J. Chromatogr. A 723 (1996) 1.37] R. Hahn, M. Panzer, E. Hansen, J. Mollerup, A. Jungbauer, Sep. Sci. Technol. 37

(2002) 1545.38] G. Iberer, R. Hahn, A. Jungbauer, LC-GC 17 (1999) 998.39] A. Staby, R.H. Jensen, M. Bensch, J. Hubbuch, D.L. Dünweber, J. Krarup, J. Nielsen,

M. Lund, S. Kidal, T.B. Hansen, I.H. Jensen, J. Chromatogr. A 1164 (2007) 82.40] A. Staby, I.H. Jensen, I. Mollerup, J. Chromatogr. A 897 (2000) 99.

[41] F.C. Leinweber, U. Tallarek, J. Chromatogr. A 1006 (2003) 207.

42] H. Darcy, in: R.A. Freeze, W. Back (Eds.), Physical Hydrology, Hutchinson Ross,

Stroudsburg, PA, 1983.43] I. Mihelic, D. Nemec, A. Podgornik, T. Koloini, J. Chromatogr. A 1065 (2005) 59.44] K. Kontturi, S. Mafé, J.A. Manzanares, B.L. Svarfvar, P. Viinikka, Macromolecules

29 (1996) 5740.45] M. Al-Bokari, D. Cherrak, G. Guiochon, J. Chromatogr. A 975 (2002) 275.