Poly(acrylamide-co-alkylacrylamides) for electrophoretic DNA purification in microchannels

Article abstract of DOI:10.1021/ac049000y

We have created a family of water-soluble block copolymers of acrylamide and N-alkylacrylamides that are designed to selectively remove proteins from DNA via microchannel electrophoresis. It was hypothesized that the inclusion of hydrophobic subunits in the polymer chain, in sufficient concentration, could lead to protein adsorption due to hydrophobic interactions. A series of N-alkylacrylamide comonomers with varying alkyl chain lengths (C4, C6, C8) and also an N,N-dialkyl group (C6-C6) were synthesized via reactions between acryloyl chloride and the respective alkylamines. Copolymers were synthesized using an aqueous "micellar" polymerization technique, which involves dissolving acrylamide in the aqueous phase while hydrophobic monomers are solubilized in sodium dodecyl sulfate micelles. Copolymers comprising up to 4 mol % of a hydrophobic subunit (as verified by ¹H NMR) were synthesized. Polymer molecular weights were determined by tandem gel permeation chromatography-multiangle laser light scattering, and ranged from 1.5 × 10⁶ to 4.3 × 10⁶. Capillary electrophoresis analysis of bovine serum albumin and β-lactoglobulin migration in these matrixes revealed that the octylacrylamide and dihexylacrylamide copolymers show the most significant extent of protein adsorption while butylacrylamides show no noteworthy adsorption trend. All copolymer matrixes studied allowed the passage of a dsDNA digest, and displayed some DNA sieving ability at 0.5% (w/w) in TTE (50 mM Tris, 50 mM TAPS, 2 mM EDTA, pH 8.4) buffer. These matrixes are demonstrated in on-chip experiments to adsorb protein, in a step toward meeting the front-end processing goals of μ-TAS for genetic analysis applications.

Full text of DOI:10.1021/ac049000y

Anal. Chem. 2005, 77, 772-779
Poly(acrylamide-co-alkylacrylamides) for
Electrophoretic DNA Purification in Microchannels
Thomas N. Chiesl, Wei Shi, and Annelise E. Barron*
Department of Chemical and Biological Engineering, Northwestern University, Evanston, Illinois 60208
analytical chemistry community and offers the promise of faster,
highly parallel, automated, integrated devices. These efforts have
been predominantly focused on scaling down current techniques
for biomolecule analysis, such as electrophoresis,3 DNA amplifica-
tion (polymerase chain reaction (PCR)),4-6 and on-line fluorescent
labeling,7 and creating ultra-high-throughput devices that can
analyze samples in a parallel fashion.8,9 Although many achieve-
ments have been made in scaling down these analytical processes,
there remains much work to be done in the development of new
materials designed specifically for use in these microfluidic
devices.10,11 Moreover, to date, with a few notable exceptions,
published reports have dealt with the analysis of simple and
prepurified biological samples, containing only the analytes of
interest, and thus have ignored the integration of preprocessing
sample treatment steps on bioanalysis chips.
While the extraction and isolation of DNA from cellular
material is one of the most commonly used procedures in genetics,
molecular biology, and biochemistry, most of the current macro-
scopic techniques are time-consuming (∼hours) and not amenable
to implementation or integration on a microfluidic platform. The
major hurdles in this regard are centrifugation and precipitation
of DNA. The precipitation and recovery of DNA with ethanol has
yet to be accomplished on a chip device. In fact, the associated
centrifugation steps required, in both this and other laboratory-
scale DNA isolation methods (such as the commonly used Qiagen
kit), typically exceed 5000g. The scaling laws of centrifugation
become unfavorable for small microchip devices, requiring high
spin rates, especially as further device footprint miniaturization
makes microdevices more portable and affordable. Hence, a
simple scaling down of current techniques into the microscale
format (centrifugation and precipitation) may not be a promising
approach to DNA purification from a complex biological milieu.
For microfluidic devices to be useful “in the field”, they must
have the capabilities to perform front-end tasks such as the
We have created a family of water-soluble block copoly-
mers of acrylamide and N-alkylacrylamides that are
designed to selectively remove proteins from DNA via
microchannel electrophoresis. It was hypothesized that
the inclusion of hydrophobic subunits in the polymer
chain, in sufficient concentration, could lead to protein
adsorption due to hydrophobic interactions. A series of
N-alkylacrylamide comonomers with varying alkyl chain
lengths (C4, C6, C8) and also an N,N-dialkyl group (C6-
C6) were synthesized via reactions between acryloyl
chloride and the respective alkylamines. Copolymers were
synthesized using an aqueous “micellar” polymerization
technique, which involves dissolving acrylamide in the
aqueous phase while hydrophobic monomers are solubi-
lized in sodium dodecyl sulfate micelles. Copolymers
comprising up to 4 mol % of a hydrophobic subunit (as
verified by 1H NMR) were synthesized. Polymer molecular
weights were determined by tandem gel permeation
chromatography-multiangle laser light scattering, and
ranged from 1.5 × 106 to 4.3 × 106. Capillary electro-
phoresis analysis of bovine serum albumin and â-lacto-
globulin migration in these matrixes revealed that the
octylacrylamide and dihexylacrylamide copolymers show
the most significant extent of protein adsorption while
butylacrylamides show no noteworthy adsorption trend.
All copolymer matrixes studied allowed the passage of a
dsDNA digest, and displayed some DNA sieving ability
at 0.5% (w/w) in TTE (50 mM Tris, 50 mM TAPS, 2 mM
EDTA, pH 8.4) buffer. These matrixes are demonstrated
in on-chip experiments to adsorb protein, in a step toward
meeting the front-end processing goals of µ-TAS for genetic
analysis applications.
One of the most significant accomplishments of the 20th
century was the sequencing of the human genome.1,2 The
availability of the genome sequence will facilitate undertakings
in gene therapy, forensic analysis, biological weapon detection,
pharmaceutical development, and countless other areas. Over the
past several years the development of “lab-on-a-chip” technologies
for genetic analysis has been at the forefront of research in the
(3) Culbertson, C. T.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 2000, 72,
5814-5819.
(4) Woolley, A. T.; Hadley, D.; Landre, P.; deMello, A. J.; Mathies, R. A.;
Northrup, M. A. Anal. Chem. 1996, 68, 4081-4086.
(5) Khandurina, J.; McKnight, T. E.; Jacobson, S. C.; Waters, L. C.; Foote, R.
S.; Ramsey, J. M. Anal. Chem. 2000, 72, 2995-3000.
(6) Liu, J.; Enzelberger, M.; Quake, S. Electrophoresis 2002, 23, 1531-1536.
(7) Liu, Y.; Foote, R. S.; Jacobson, S. C.; Ramsey, R. S.; Ramsey, J. M. Anal.
Chem. 2000, 72, 4608-4613.
(8) Emrich, C. A.; Tian, H.; Mednitz, I. L.; Mathies, R. A. Anal. Chem. 2002,
74, 5076-5083.
* To whom correspondence should be addressed. Phone: (847) 491-2778.
Fax: (847) 491-3728. E-mail: a-barron@northwestern.edu.
(1) Venter, J. C.; Adams, M. D.; Myers, E. W.; Li, P. W.; Mural, R. J.; Sutton, G.
G.; Smith, H. O.; et al. Science 2001, 291, 1304-1351.
(2) Lander, E. S.; Linton, L. M.; Birren, B.; Nusbaum, C.; Zody, M. C.; Baldwin,
J.; Devon, K.; et al. Nature 2001, 409, 860-921.
(9) Thorsen, T.; Maerkl, S. J.; Quake, S. R. Science 2003, 298, 580.
(10) Mello, A. J. d.; Beard, N. Lab Chip 2003, 3, 11N-19N.
(11) Vreeland, W. N.; Barron, A. E. Curr. Opin. Biotechnol. 2002, 13, 87-94.
772 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005
10.1021/ac049000y CCC: $30.25 © 2005 American Chemical Society
Published on Web 12/16/2004

processing of raw, complex biological samples to isolate the target
analytes of interest, often at low concentrations or of low
abundance.10 Clearly, sample preparation and purification is one
limiting step in the creation of an integrated, microscale DNA
analysis device, which remains to be fully addressed. Only a few
research groups have published demonstrations of on-chip pre-
processing and sample purification methods. Among the ap-
proaches used are filtration,12-17 diffusive laminar liquid extrac-
tion,18-21 specific DNA hybridization,22 and solid-phase extraction
(SPE).23-28
One of the most promising and generally useful preprocessing
techniques demonstrated to date is on-chip SPE with the use of
silica-based beads 1.5-5 µm in diameter, which are typically
coated with an alkylsilane. Loading these particles into micro-
devices so that they form a consistent and uniform bead bed can
be difficult.29 To date, most studies of these systems have
concentrated on the separation of small molecules (such as
fluorescent dyes). The most successful and biologically relevant
demonstration of on-chip sample purification using SPE was
reported by Landers et al.30 Their goal was to extract DNA from
cells, on a microfluidic chip, for subsequent genetic analysis by
electrophoresis. They report packing a reservoir in a microfluidic
device with commercially available SPE beads. Subsequently, they
showed that if they loaded whole blood lysate onto the chip, and
then eluted the adsorbed biomolecules via hydrodynamic flow with
changing buffer conditions, some fractions obtained from the chip
contained enough DNA to be used for subsequent (off-chip) PCR
with successful amplification. In addition to the desired DNA
binding, nonspecific adsorption of proteins and lipids can occur
on the beads. A wash step could be included to remove some of
these proteins; however, it might be difficult to obtain completely
pure DNA (free of proteins and other biomolecules) by this SPE
approach.
The most significant limitation of SPE beds for rigorous on-
chip DNA purification is their limited surface area for adsorption
and thus relatively low load capacity. A relatively large diameter,
nonporous bead has only its outer surface available for molecular
interactions, and this may limit the amount of DNA collected. In
a different approach to biomolecule purification, the surface area
available for extraction has been increased by the introduction of
a monomer solution into microchannels, which is then polymer-
ized in situ to create a well-defined porous structure,31-33 in a
system that showed enrichment factors of 1000 for hydrophobic
peptides and green fluorescent protein (from a dilute peptide/
protein solution31,32). These polymer systems have been shown
to be reusable, but are not replaceable, and have yet to be
demonstrated for purification of DNA from a complex biological
matrix.
In work done by the Mathies group, a very promising on-chip
sample preprocessing technique has been investigated, based on
the use of specific DNA hybridization to selectively capture DNA
strands terminating in a known sequence, for subsequent analysis
by electrophoretic DNA sequencing.22 An anchored, complimen-
tary ssDNA sequence (attached to a polyacrylamide matrix) is
used to capture ssDNA products of a Sanger cycle-sequencing
reaction. When the products of the PCR reaction were electro-
phoresed through the “capture gel”, the DNA amplicons of interest
hybridized while other sample components (salt, primers, poly-
merase) passed through. Electric field and temperature conditions
for optimal DNA capture were identified. This method, while
elegant and nicely applicable for PCR and cycle-sequencing
reaction product purification, may not be able to purify larger,
nonamplified dsDNA targets from raw cell lysate.
The strategy investigated in this work is to purify dsDNA from
proteins and (eventually, in future formulations) from other
cellular debris using a combination of microchip electrophoresis
and hydrophobic adsorption of nongenetic material onto hydro-
phobically modified polyacrylamide (HMPAM) networks. The first
technique, microchannel electrophoresis, enables a kinetic separa-
tion by discriminating molecules on the basis of their charge-to-
hydrodynamic friction ratio, as well as by their polarity of charging
(positive vs negative). DNA, a negatively charged molecule, will
migrate toward the higher potential, while positively charged
molecules will migrate toward the lower potential, and neutral
molecules will remain motionless. Thus, this technique should
intrinsically separate DNA from all neutral or positively charged
biomolecules. Thus, only negatively charged lipids and proteins
must be separated by the second modality, hydrophobic interac-
tion. We have designed copolymers based on a polyacrylamide
backbone to irreversibly adsorb proteins, with (presumably) ample
surface area for adsorption. The inclusion of hydrophobic subunits
in a polymer network could lead to substantial and irreversible
protein and lipid adsorption via polymer interactions with the
protein’s hydrophobic amino acids or lipid alkyl groups, respec-
tively, while (we hypothesized) allowing DNA to freely pass during
electrophoresis. Thus, two modalities of purification/isolation are
(12) Chu, W.-H.; Chin, R.; Huen, T.; Ferrari, M. J. Microelectromech. Syst. 1999,
8, 34-42.
(13) Broyles, B. S.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 2003, 75, 2761-
2767.
(14) Stemme, G.; Kittilsland, G. Appl. Phys. Lett. 1988, 53, 1566.
(15) Kittilsland, G.; Norden, G. S. Sens. Actuators, A 1990, 21, 904-907.
(16) Andersson, H.; Wijngaart, W. v. d.; Stemme, G. Electrophoresis 2001, 22,
249-257.
(17) He, B.; Tan, L.; Regnier, F. Anal. Chem. 1999, 71, 1464-1468.
(18) Weigl, B. H.; Yager, P. Science 1999, 283, 346-347.
(19) Brody, J. P.; Osborn, T. D.; Forster, F. K.; Yager, P. Sens. Actuators, A 1996,
54, 704-708.
(20) Brody, J. P.; Yager, P. Sens. Actuators, A 1997, 58, 13-18.
(21) Tokeshi, M.; Minagawa, T.; Kitamori, T. Anal. Chem. 2000, 72, 1711-
1714.
(22) Paegel, B. M.; Yeung, S. H. I.; Mathies, R. A. Anal. Chem. 2002, 74, 5092-
5098.
(23) Kutter, J. P.; Ramsey, R.S.; Jacobson, S. C.; Ramsey, J. M. J. Microcolumn
Sep. 1998, 10, 313-319.
(24) Xu, Y.; Vaidya, B.; Patel, A. B.; Ford, S. M.; McCarley, R. L.; Soper, S. A.
Anal. Chem. 2003, 75, 2975-2984.
(25) Christel, L. A.; Petersen, K.; McMillan, W.; Northrup, M. A. J. Biomech.
Eng. 1999, 121, 22-27.
(26) Olesch, R. D.; Shultz-Lockyear, L. L.; Ning, Y.; Harrison, D. J. Anal. Chem.
2000, 72, 585-590.
(27) Ekstrom, S.; Malmstrom, J.; Wallman, L.; Lofgren, M.; Nilsson, J.; Laurell,
T.; Marko-Varga, G. Proteomics 2002, 2, 413-421.
(28) Bergkvist, J.; Ekstrom, S.; Wallman, L.; Lofgren, M.; Marko-Varga, G.;
Nilsson, J.; Laurell, T. Proteomics 2002, 2, 422-429.
(29) Oleschuk, R. D.; Shultz-Lockyear, L. L.; Ning, Y. B.; Harrison, D. J. Anal.
Chem. 2000, 72, 585-590.
(31) Yu, C.; Xu, M.; Svec, F.; Frechet, J. M. J. J. Polym. Sci., Part A: Polym.
Chem. 2002, 40, 755-769.
(32) Yu, C.; Davey, M. H.; Svec, F.; Frechet, J. M. J. Anal. Chem. 2001, 73,
5088-5096.
(30) Breadmore, M. C.; Wolfe, K. A.; Arcibal, I. G.; Leung, W. K.; Dickson, D.;
(33) Stachowiak, T. B.; Rohr, T.; Hilder, E. F.; Peterson, D. S.; Yi, M.; Svec, F.;
Frechet, J. M. J. Electrophoresis 2003, 24, 3689-3693.
Giordano, B. C.; Power, M. E.; et al. Anal. Chem. 2003, 75, 1880-1886.
Analytical Chemistry, Vol. 77, No. 3, February 1, 2005 773

Figure 1. Chemical structures of acrylamide and the hydrophobically modified acrylamides studied.
used to create a microchip-based electrokinetic chromatography
“guard column”, or protein/lipid sponge. Success in these
endeavors could lead to the creation of a technology that will
facilitate rapid genetic analyses on integrated microfluidic devices
that can accept and process raw cell lysate.
micelle concentration (cmc). The hydrophobic comonomers,
which are not water-soluble, are dissolved in the interiors of the
SDS micelles. This technique allows the copolymerization of
acrylamide with hydrophobic (and otherwise water-insoluble)
N-alkylacrylamides. The use of SDS microdomains of varying size
also allows direct control over polymer properties such as the
average concentration (mol %) of hydrophobe incorporation and
average hydrophobic block length within the copolymers.35,38
All polymerization reactions were conducted in either a 300
or a 500 mL four-neck flask (Kontes, Vineland, NJ), with 100 mL
of water as the basis for the reaction volume. To water are added
acrylamide (Amresco, Inc., Solon, OH) and SDS (Sigma, St. Louis,
MO). Typically, 3 wt % acrylamide and 2.5-3 g of SDS are initially
dissolved. The mixture is stirred for 30 min, undisturbed, and then
degassed by nitrogen bubbling for an additional 30 min. After
sufficient bubbling time has removed oxygen from the reaction
medium and allowed for hydrophobic monomers to be dissolved,
the initiator 4,4′-azobis(4-cyanovaleric acid) (ACVA) is quickly
added, and allowed to mix into the solution at room temperature.
The vessel is submersed into a water bath held at 50 °C to begin
the polymerization. Free-solution polymerization of acrylamide
occurs until the propagating chain encounters a micelle, at which
point the hydrophobic moieties contained within the micelle are
added to the polymer chain; after this, the acrylamide polymer
continues to propagate in free solution.35 After 4 h of reaction,
the vessel is removed from the water bath for copolymer recovery.
The precipitation of HMPAM with acetone, followed by dialysis
against water, is used to purify the copolymer products. The
polymer solution is first diluted with an additional 100-300 mL
of water, precipitated with acetone, and then physically removed
from solution. The polymer is further rinsed with acetone and
allowed to air-dry. The dried copolymers are redissolved and then
purified by dialysis against deionized distilled water using Spectra/
Por cellulose ester dialysis membranes (Spectrum, Gardena, CA),
having a molecular weight cutoff of 100000, with frequent water
MATERIALS AND METHODS
Monomer Synthesis and Recovery. Hydrophobic N-alkyl-
and N,N-dialkylacrylamide monomers were created by the reaction
between acryloyl chloride and a corresponding alkylamine. The
synthesis procedure used to create these monomers is similar to
that reported by McCormick.34 The amines, including butylamine,
hexylamine, octylamine, and dihexylamine, and acryloyl chloride
(hazardous) were purchased from Sigma-Aldrich (Milwaukee, WI)
and used without further purification. Briefly, the reaction is
carried out in a round-bottom flask chilled to 0 °C. An amine, for
example, hexylamine, is dissolved to 10 mM in 60 mL of
methylene chloride. Triethylamine is included at equal molarity
to neutralize the hydrochloric acid byproduct of the reaction.
Acryloyl chloride is dissolved in 60 mL of methylene chloride (to
a 10 mM concentration) and added to the reaction vessel using
an addition funnel under equal pressure. The alkylacrylamide
product is then extracted by three consecutive additions of
aqueous solution: 100 mM NaOH, 100 mM HCl, and finally water.
Following the final extraction, the mixture is placed in a rotary
evaporator to remove the methylene chloride. A schematic
illustration of the series of N-alkylacrylamide comonomers created
with varying alkyl chain lengths (C4, C6, C8) as well as an N,N-
dialkylacrylamide (C6-C6) is presented in Figure 1.
Polymer Synthesis and Recovery. The polymers synthesized
for this research were created by a technique called “micellar”
polymerization,35-40 which has many similarities to emulsification
polymerization. The reaction medium is primarily water, with
microdomains of sodium dodecyl sulfate (SDS) above its critical
(34) McCormick, C. L.; Nonaka, T.; Johnson, C. B. Polymer 1988, 29, 731-739.
(35) Biggs, S.; Hill, A.; Selb, J.; Candau, F. J. Phys. Chem. 1992, 96, 5-1511.
(36) Candau, F.; Regaldo, E. J.; Selb, J. Macromol. Symp. 2000, 150, 241-249.
(37) Kujawa, P.; Rosiak, J. M.; Selb, J.; Candau, F. Macromol. Chem. Phys. 2001,
202, 1384-1397.
(38) Hill, A.; Candau, F.; Selb, J. Macromolecules 1993, 26, 4521-4532.
(39) Regaldo, E. J.; Selb, J.; Candau, F. Macromolecules 2000, 33, 8720-8730.
(40) Volpert, E.; Selb, J.; Candau, F. Macromolecules 1996, 29, 1452-1463.
774 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

NaH2PO4, and 200 ppm NaN3. The tandem GPC-MALLS data
were processed using ASTRA software from Wyatt Technology.
ASTRA was used to calculate the molecular weights (MWs) and
polydispersity indices (PDIs) of the analyzed polymers. All
analyses were repeated at least three times. The two processing
parameters used to analyze the hydrophobic copolymers were
known dn/dc ) 0.11135 and the AUX calibration constant. The
presence of <5 mol % hydrophobic comonomer does not signifi-
cantly alter the value of dn/dc.35,40 Results for the polymers used
in this study, which could be determined, are shown in Table 1.
The copolymers were characterized using a Mercury 400 MHz
1H NMR spectrometer (Varian, Palo Alto, CA) to determine the
amount of hydrophobic monomer incorporated into the copolymer
(i.e., the concentration (mol %) of hydrophobe). An assigned NMR
spectrum for a poly(acrylamide-co-hexylacrylamide) previously
existed in the literature40 and served as a guide for our NMR
assignments and hence the alkylacrylamide incorporation deter-
minations. Specifically, hydrogens found in the polymer backbone
have representative peaks at 1.7 and 2.2 ppm,40 and the area under
these peaks was compared with peak areas for the terminal alkyl
methyl hydrogens found at 0.8 ppm, to estimate the amount of
hydrophobic subunit incorporated. Table 1 shows the results.
Capillary Electrophoresis. Performance testing of HMPAM
matrixes, at a polymer concentration of 0.5 wt %, was done by
capillary electrophoresis (CE) on a Bio-Rad Biofocus 3000 (Bio-
Rad, Hercules, CA). Capillaries were dynamically coated to reduce
electroosmotic flow and protein adsorption by first rinsing with 1
M HCl, followed by a 0.5 wt % solution of hydrophilic polyDura-
mide, or poly(N-hydroxyethylacrylamide).41,42 Polymer solutions
used in biomolecule separations were loaded into the capillary
by the application of high pressure (100 psi) from a nitrogen
source for 45-60 s. This loading time was more than adequate
to pump several column volumes of these polymer solutions at
this concentration (0.5 wt %) through the capillary (75 µm inner
diameter, 25 cm length, and 20 cm to the detection window). For
each experiment, fresh polymer matrix was loaded. UV absorption
was used for simultaneous biomolecule detection at 214 and 260
nm wavelengths. Unless otherwise noted, all protein samples for
CE were dissolved in water at 1 mg/mL, from stock solutions of
10 mg/mL, and prepared freshly before the experiment. Proteins
studied were purchased from Fisher Scientific (Pittsburgh, PA)
and included bovine serum albumin (BSA), pepsin, R-lactalbumin,
â-lactoglobulins A and B, trypsin inhibitor, and acylase. DNA
samples (a 50 bp ladder, and a φX174-Hae III dsDNA digest) were
purchased from Invitrogen (Carlsbad, CA). Sample injection was
performed electrophoretically, by the application of 20 kV (800
V/cm) for 3 s. The running buffer was 1× TTE (50 mM Tris, 50
mM TAPS, 2 mM EDTA, pH 8.4), and electrophoretic separations
were performed at 400 V/cm with a ∼18 µA current.
Table 1. Summary of the Properties of Copolymers
Used in These Experiments
hydrophobe
polymer
concn (mol %)
(×M10W-6
)
PDI
LPA
0
1.38
2.59
2.93
3.13
2.34
2.03
1.96
1.54
1.53
2.25
2.19
2.17
2.78
1.52
3.83
2.10
3.44
1.41
a
1.8
1.6
1.5
1.4
1.7
1.6
1.9
1.5
1.8
1.7
1.6
1.7
1.5
a
0
LPA-co-butylacrylamide
0.06
0.08
0.14
0.18
0.21
2.84
3.01
0.21
0.23
0.27
0.42
1.18
4.29
0.12
0.14
0.19
0.26
0.27
0.27
0.31
0.31
0.42
0.45
3.28
0.06
0.12
0.13
0.17
0.19
0.19
0.21
0.22
0.23
0.24
0.25
0.27
0.29
LPA-co-hexylacrylamide
LPA-co-octylacrylamide
a
a
1.4
1.4
a
6.49
a
a
a
4.29
5.04
1.58
2.60
2.10
3.26
2.07
a
a
a
a
a
a
LPA-co-dihexylacrylamide
1.5
a
a
a
a
a
a
a
a
a
a
2.78
a
a
a
a
a
a
a
a
a
a
a
a Could not be accurately determined.
changes over 10 days, and then the polymers are freeze-dried. In
total, approximately 50 batches of copolymers with varying
molecular weights, hydrophobic moieties, and concentrations (mol
%) of hydrophobe were created for these studies. Additionally,
linear polyacrylamides (LPAs) were synthesized using essentially
the same conditions to serve as control, hydrophilic polymers.
Table 1 shows the polymers that were used in the capillary and
chip electrophoresis experiments we discuss here.
Copolymer Characterization. To determine the molecular
weights and polydispersities of the synthesized polymers, the
samples were fractionated by GPC prior to on-line MALLS and
refractive index detection, using a Waters 2690 Alliance separa-
tions module (Milford, MA) with Shodex (New York, NY) OHpak
columns SB-806 HQ, SB-804 HQ, and SB-802.5 HQ connected in
series. In this tandem GPC-MALLS mode, the effluent from the
GPC system flows directly into a DAWN DSP laser photometer
and Optilab DSP interferometric refractometer connected in series
(both Wyatt Technology, Santa Barbara, CA). A 100 µL portion
of each sample was injected into the tandem GPC-MALLS system
at a concentration of ∼0.5 mg/mL. The flow rate was 0.3 mL/
min, and the mobile phase consisted of 100 mM NaCl, 50 mM
Microchip Electrophoresis. The electrophoresis of DNA and
proteins in microfluidic chips was conducted using a system
custom-built for our laboratory by engineers at ACLARA Bio-
Sciences (Mountain View, CA) and designed to allow sensitive,
multi-color laser-induced fluorescence (LIF) detection. The instru-
ment consists of an electrical subsystem, which supplies voltage
(41) Albarghouthi, M. N.; Stein, T. M.; Barron, A. E. Electrophoresis 2003, 24,
1166-1175.
(42) Albarghouthi, M. N.; Buchholz, B. A.; Huiberts, P. J.; Stein, T. M.; Barron,
A. E. Electrophoresis 2002, 23, 1429-1440.
Analytical Chemistry, Vol. 77, No. 3, February 1, 2005 775

to the microfluidic device, and an optical subsystem, which allows
the detection of fluorescent moieties as they migrate past a
determined position in the channel. Both subsystems are operated
using a single program (from ACLARA BioSciences) written in
LabView software.
RESULTS AND DISCUSSION
Monomer and Polymer Synthesis and Purification. The
physical appearance of each alkylacrylamide differed, and the
products varied from liquids (butylacrylamide and dihexylacryl-
amide) to solids (hexylacrylamide and octylacrylamide). As the
hydrophobicity of the monomer increased, its water solubility
decreased, and this made the extraction process easier. Reaction
yields were typically greater than 90%, and the products were
determined to be pure by 1H NMR (data not shown). Copolymer
synthesis was relatively easy to accomplish in good yield, accord-
ing to the methods described. The results of the copolymer
characterization according to molecular weight and the concentra-
tion (mol %) of hydrophobe are shown in Table 1. The obtained
molecular weights were relatively large, ranging from 1.4 × 106
to 6.5 × 106, while PDIs averaged 1.6 and ranged from 1.4 to 1.9.
Copolymers including modified acrylamides with long alkyl chain
lengths (C8, C6-C6) and/or a relatively large concentration (mol
%) hydrophobe (g0.5 mol %) yielded weak light-scattering and
refractive index signals on the GPC-MALLS system, most likely
because they were not fully recovered from the columns due to
copolymer interactions with the GPC packing material or because
they had molecular weights greater than the fractionation potential
of the column. In these cases, the molecular weight and polydis-
persity could not be accurately determined. It is assumed that
the molecular weights of these copolymers are similar to those
of the other copolymers, given the identical synthesis conditions.
We are currently investigating methods that will allow the accurate
analysis of these polymers by tandem GPC-MALLS, e.g., the use
of a different solvent system. However, as discussed below, the
molecular weight of the polymers did not seem to be an important
variable for the intended application.
Capillary and Chip Electrophoresis using HMPAM and
LPA Networks. Two sets of control experiments were performed
by CE and chip electrophoresis using HMPAM and LPA networks.
The first experiments confirmed that both DNA and proteins could
be co-injected from a single sample and separated electrophoreti-
cally in a linear polyacrylamide matrix (0.5 wt %) without any
apparent detrimental effects such as aggregation, adsorption, or
peak broadening. These were done by CE with UV detection
simultaneously at 214 and 260 nm (to see protein and DNA
molecules, respectively). Since both DNA and proteins are
routinely separated by polyacrylamide gel electrophoresis and by
CE in LPA solutions, these results were not unexpected. However,
to our knowledge there has been no report of these two very
different types of analytes being co-injected and separated by CE
in a single run. DNA peaks elute first, followed by proteins (data
not shown). Some size-based separation of DNA occurs, and the
native proteins are separated by their intrinsic differences in
electrophoretic mobility. Protein peak efficiencies and resolution
were similar to those obtained by free-solution electrophoresis in
a polyDuramide-coated capillary, i.e., without the LPA separation
matrix. The second control experiment involved the separation
of a mixture of dsDNA molecules of different sizes through
HMPAM solutions by both CE and chip electrophoresis. As shown
in the microchip electropherogram provided in Figure 2, a dsDNA
sizing ladder is nicely resolved in an HMPAM matrix in less than
125 s; peak efficiencies for this separation averaged 5 million
plates/m. Additionally, we believe that the hydrophobic moieties
The power subsystem consists of a high-voltage power supply
with independent control of four electrodes. Each electrode can
be set to a desired voltage between 0 and 4.5 kV or, alternatively,
may be floated (disconnected) from the circuit. The software
allows the user to set and hold the voltage state of all four
electrodes for an arbitrary duration. Multiple steps may be
programmed for novel separation schemes or complex chip
functions. Additionally, four controllable, low-voltage outputs allow
for the digital control of external devices, e.g., a heater. The optical
subsystem is a confocal, epifluorescence system in which fluoro-
phores are excited by a JDS Uniphase Series 2214-30sl single-
line, 488 nm argon ion laser (San Jose, CA). Mirrors direct the
laser beam into a TE200 inverted, epifluorescence microscope.
The beam is then passed through a band-pass filter, reflected off
a dichroic mirror, and focused through a Nikon 10×/0.45
microscope objective into the center of the microfluidic channel,
producing a laser spot size of ∼10 µm. Emitted fluorescence is
collected through the same objective and is passed through the
dichroic mirror, followed by a second, wide-band-pass filter
(Chroma Technology, Brattleboro, VT). The spectrum of the
filtered light is measured by directing it through a transmission
grating and focusing it onto a high-quantum-efficiency, 532 × 64
pixel charge-coupled device (CCD) cooled to -15 °C (Hamamatsu
Corp., Bridgewater, NJ). Pixel binning is applied to the data from
the CCD camera to quantify the intensity of emission over a
particular range of wavelengths, calibrated by Raman scattering
lines of solvents. Data collection can be accomplished at rates up
to 50 Hz. The CCD output is collected, binned, low-pass filtered,
and stored using a program written in LabView.
Experiments were carried out with T3550 glass microchips
(Micronit, Enschede, The Netherlands) dynamically coated to
reduce electroosmotic flow and protein adsorption by first rinsing
with 1 M HCl followed by a 0.5 wt % solution of polyDuramide, or
poly(N-hydroxyethylacrylamide) synthesized in our laboratory.41,42
The chips have a standard, 100 µm “offset T” injector,43-45 a 3.5
cm separation distance, and 50 µm wide by 20 µm tall channels.
Samples were injected by grounding the sample well and applying
300 V on the sample waste for 20 s while floating the remaining
electrodes. Pullback during the separation step was at 350 V on
both sample and sample waste wells, with 1400 V on the analysis
waste, providing a separation field strength of 350 V/cm.
Proteins for chip analysis were labeled with FITC, a reagent
purchased from Fisher Scientific (Pittsburgh, PA), by following a
standard labeling protocol, and purified via elution from a packed
Sephadex G-25 column. The final protein sample concentration
was estimated to be ∼0.5 mg/mL.
(43) Manz, A.; Fettinger, J. C.; Verpoorte, E.; Ludi, H.; Widmer, H. M.; Harrison,
D. J. Trends Anal. Chem. 1991, 10, 144-149.
(44) Effenhauser, C. S.; Manz, A.; Widmer, H. M. Anal. Chem. 1993, 65, 2637-
2642.
(45) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z. H.; Effenhauser, C. S.; Manz, A.
Science 1993, 261, 895-897.
776 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

Figure 3. Capillary electropherograms comparing BSA peak elution
in polyacrylamide (blue) and poly(acrylamide-co-dihexylacrylamide)
(red). Injection of protein was at 800 V/cm for 3 s. Significant
adsorption of BSA onto the HMPAM occurs (26000-1600 mAU min).
The copolymer molecular weight is 1.5 million, and the copolymer
contains 0.122 mol % dihexylacrylamide. The polyacrylamide mo-
lecular weight is 1.4 million. The field strength was 400 V/cm, with
typically an 18 µΑ current.
Figure 2. A control experiment: microchip electropherogram of a
50 bp dsDNA ladder, freely migrating through a 2 wt % HMPAM
copolymer solution (MW ) 4.3 × 106 with 0.31 mol % octylacryl-
amide). The separation is both rapid (<125 s) and highly efficient
(average of 5.5 million plates/meter). The field strength was 240 V/cm
with a 3.5 cm separation distance and a 2.5 µA current.
in the copolymer may interact with each other to help stabilize
the copolymer network, thus enhancing DNA separation. Separa-
tion of the same DNA sample in an LPA network, roughly matched
in molecular weight and used at the same concentration, was
substantially poorer (data not shown). Further, comparative
studies of DNA separations in HMPAM vs LPA matrixes are
ongoing.
in the copolymer (Figure 4). The effects of alkyl chain length and
the overall concentration (mol %) of hydrophobe in the copolymer
on protein adsorption are shown in Figure 4A for BSA and Figure
4B for â-lactoglobulins A and B. These microchannel electro-
phoresis experiments reveal that N,N-dihexyl- and N-octylacryl-
amide copolymers enable the most significant protein adsorption,
as might be expected given that these are the most hydrophobic
monomers studied. N-Hexylacrylamide copolymers showed mod-
erate adsorption, while butylacrylamides exhibited essentially
none. Because both BSA and the lactoglobulins adsorb to the
copolymer matrixes in the same manner (compare parts A and B
of Figure 4), there appears to be a generalized trend that
increasing the hydrophobicity of the copolymer increases the
extent of protein adsorption. Differences in the molecular weights
of the copolymers were found to not significantly alter the
migration time of the proteins nor the amount of adsorption
observed, using solutions of 0.5 wt % dissolved polymer (data not
shown).
A sample containing a mixture of proteins and dsDNA was
injected under analytical conditions, with the application of 10 kV
(400 V/cm) for 8 s, and analyzed by capillary electrophoresis. In
particular, the sample consisted of a dsDNA digest at 100 µg/mL
and BSA at 1 mg/mL in water. A comparison of the results
obtained for the CE separation of this mixed DNA/protein sample
through a linear polyacrylamide network and a hydrophobically
modified polyacrylamide network (co-dihexylacrylamide) is shown
in Figure 5. Dual detection at 214 and 260 nm was used to
distinguish and detect both DNA and proteins during the same
experiment. In the protein-sensitive electropherograms, Figure
5A, protein peaks appear large relative to those of the DNA, while
in the DNA-sensitive electropherograms, Figure 5B, both DNA
and protein peaks have moderate heights (due to protein absorp-
tion of UV light resulting from the presence of aromatic amino
acid side chains). DNA, having a high electrophoretic mobility,
elutes first, followed by the protein. We found that complete
adsorption of the BSA is accomplished using this dihexylacryl-
amide-containing copolymer matrix, as determined by the absence
of a peak for the injected BSA. In the protein-sensitive electro-
Adsorption of Proteins onto HMPAM Networks. In the
next set of CE experiments, a relatively large amount of protein
sample was injected into the capillary (i.e., under nonanalytical
conditions) to allow accurate determinations of the extent of
protein adsorption onto or load capacities of the various copolymer
networks. The extent of protein adsorption onto the copolymer
networks was determined by comparing the peak height and peak
area of the eluted proteins relative to benzoic acid, which acted
as an internal standard. We found that the extent of protein
adsorption onto the HMPAMs was governed by both the amount
of hydrophobic subunit incorporated and the type of hydrophobic
moiety. Typical CE electropherograms obtained after BSA migra-
tion through both an LPA matrix (with no hydrophobic content)
and a hydrophobically modified polyacrylamide copolymer (with
0.122 mol % co-dihexylacrylamide), under identical CE injection
conditions, are compared in Figure 3. As seen in the figure, both
the peak height and the peak area for BSA are dramatically
reduced in the HMPAM trace, compared to the results obtained
with the acrylamide homopolymer (peak areas of 1600 vs 26000
mAU min, respectively).
Capillary electrophoresis experiments using BSA and â-lacto-
globulins A and B were conducted to determine how the type
and amount of hydrophobic moiety in the copolymers affect
protein adsorption. Data obtained from six electropherograms
were averaged for each copolymer matrix of varying concentration
(mol %) of hydrophobe and type of hydrophobic moiety. The
averaged protein peak areas were then normalized to the benzoic
acid peak area to correct for slight variations in the amount
injected, which could occur even though identical sample con-
centrations and injection conditions were used. The protein peak
area was plotted against the concentration (mol %) of hydrophobe
Analytical Chemistry, Vol. 77, No. 3, February 1, 2005 777

Figure 5. Capillary electropherograms of a protein-DNA sample
mixture, showing the essentially complete adsorption of protein (BSA)
on a hydrophobically modified polyacrylamide matrix. Protein-sensitive
UV detection is done at 214 nm (A), while DNA-sensitive detection
is at 260 nm (B). Polymer molecular weights (LPA and HMPAM) are
∼2 million, with 0.25 mol % dihexylacrylamide in the HMPAM, a 20
cm detection length, and a field of 400 V/cm. Extending the displayed
separation time does not reveal the presence of a BSA peak in the
HMPAM electropherogram. The size-based separation of DNA in
HMPAM is better than that in a homopolyacrylamide matrix of similar
molecular weight at the same concentration. Injection of the DNA-
protein mixture was done at 400 V/cm for 8 s.
Figure 4. Analysis of BSA adsorption onto copolymers containing
different alkylacrylamide comonomer units, with peak areas normal-
ized to the benzoic acid peak area at each concentration (mol %) of
hydrophobe (A). Adsorption trends for â-lactoglobulins A and B are
very similar to those for BSA migration through hydrophobically
modified copolymers (B), thus indicating a potential general trend.
Lines are drawn to aid the eye.
larger BSA, having many more lysine residues (MW ≈ 66000),
has multiple FITC labeling reaction products and consequently
appears as a relatively broad peak. Overlays of electropherograms
conducted with copolymers of varying hydrophobic content are
shown in Figure 6. The concentration of octylacrylamide in the
copolymers ranged from 0 mol % (LPA) to 0.31 mol % . In these
electropherograms, the fluorescence intensities and migration
times of both protein peaks have been normalized to those of
trypsin inhibitor (which was found, in earlier experiments, to be
nonadsorbing during CE through HMPAM networks, as further
discussed below). We see that copolymers with increasing
octylacrylamide content incrementally adsorb more BSA. In
particular, the peak height and area for BSA are reduced in
octylacrylamide matrixes when compared to those for LPA, and
the migration time of BSA becomes longer as more hydrophobic
copolymers are used. The removal of BSA from the sample,
although incomplete in this case, demonstrates the potential of
this purification technique for the selective removal of this highly
abundant blood plasma protein by HMPAMs in a microfluidic
device. Since, as shown in Figure 5, DNA peaks typically elute
well before protein peaks, a simple “slowing” of protein peaks by
HMPAM networks also could be enough to effect the purification
pherograms with detection at 214 nm (Figure 5A), a peak for BSA
is clearly visible for electrophoresis through polyacrylamide, while
it is completely absent in the HMPAM trace. In the DNA-sensitive
channel, 260 nm (Figure 5B), both the DNA and protein can be
detected after electrophoresis through the polyacrylamide, while
only DNA elution is detected in the HMPAM. Additionally the
dihexylacrylamide-containing copolymer is seen to provide sig-
nificantly better size-based separation of a dsDNA digest, possibly
due to the earlier onset of a physically cross-linked polymer
network with hydrophobic interactions between alkyl ma-
trixes.37,40,46 Thus, essentially complete adsorption of a control
protein, serum albumin (the main constituent in blood plasma),
with no adverse affect on DNA separation, has been accomplished
using a novel HMPAM network.
Implementation on a Microfluidic Device. Fluorescently
labeled trypsin inhibitor and BSA were first analyzed, each
individually, in an LPA matrix to allow for unambiguous peak
identification. Trypsin inhibitor, with seven lysine residues (MW
≈ 20000), shows one predominant product of the FITC labeling
reaction, and hence exhibits one tall, sharp peak, while the much
(46) Candau, F.; Regaldo, E. J.; Selb, J. Macromolecules 1998, 31, 5550-5552.
778 Analytical Chemistry, Vol. 77, No. 3, February 1, 2005

CONCLUSIONS
The inclusion of an upstream biosample processing technology
such as on-line DNA purification has so far been mostly over-
looked by the µ-TAS community, with a few notable exceptions.
In this work, novel copolymers have been created that adsorb
proteins via hydrophobic interactions during electrokinetic chro-
matography in both capillaries and microchips, with apparently
no dsDNA adsorption, and no adverse effects on dsDNA separa-
tion. These copolymers are synthesized by micellar polymeriza-
tion, a technique that allows otherwise water-insoluble monomers
to be incorporated into a water-soluble polymer. The concentration
1
(mol %) of hydrophobe in the copolymer can be analyzed by H
NMR. The amount and nature of the hydrophobic moiety play
important roles in separation of DNA from proteins. By increasing
both the length of the alkyl group and the total amount of
alkylacrylamide (or dialkylacrylamide), a greater extent of protein
adsorption occurs during both capillary and microchip electro-
phoresis. Our results show that dihexylacrylamide and octylacryl-
amide have the greatest potential to adsorb proteins away from a
complex biological mixture, and that conditions of electrophoresis
in HMPAMs can be manipulated to freely allow dsDNA migration
with complete adsorption of serum albumins, the most abundant
proteins in blood plasma. Success in this work could aid in the
miniaturization and automation of time-consuming tasks that
presently cannot easily be implemented on a portable, rapid-
analysis microfluidic device. Clearly, the elimination of these steps
has useful implications for DNA-based detection of biological
pathogens and other medical diagnostics in a traditional clinical
setting, as well as in nontraditional medical environments.
Figure 6. Electropherogram overlays showing the results of BSA
migration through 0.5 wt % octylacrylamide (OA) containing copolymer
matrixes. The separation field strength was 350 V/cm, with a ∼3 µA
current, 1× TTE buffer, and a detection length of 3.5 cm. Note that
the separation is completed within 50 s. Fluorescence is normalized
to the peak area for fluorescently labeled trypsin inhibitor, which acts
as an internal standard. The extent of BSA adsorption is seen to
increase with increasing content of the hydrophobic OA monomer in
the copolymers.
of DNA away from BSA and other proteins (one could simply turn
off the field after DNA peaks elute).
As described above in the discussion of Figure 4, similar trends
for the adsorption of BSA and â-lactoglobulins A and B onto
hydrophobically modified polyacrylamides indicate that retention
on HMPAM networks would be observed for many different
proteins. However, two of the proteins we investigated (trypsin
inhibitor and R-lactalbumin) were found not to adsorb strongly
to HMPAM matrixes at 20 °C, the standard run condition we used
in these studies. These results, though still anecdotal since we
have only studied the migration and adsorption of a few proteins,
indicate that it may be more difficult to remove low-molecular-
weight and/or hydrophilic proteins by adsorptive electrophoresis.
This is a logical result since native (folded) proteins have mostly
non-surface-exposed hydrophobic residues which may remain
unavailable for adsorption; however, even hydrophilic proteins
typically have numerous hydrophobic monomers buried inside
their folded structures. The use of higher temperature as an agent
to denature proteins should thus expose hydrophobic amino acid
residues, making them available for adsorption onto HMPAM
networks. Higher temperature also typically strengthens the
hydrophobic effect, which is entropically driven. We will explore
this phenomenon further using HMPAMs with a more complex
mixture of proteins, on temperature-controlled microfluidic de-
vices. The polymer formulations and electrophoresis conditions
we present here, however, may be of great interest to the
proteomics community since this is a method for the removal of
serum albumins, which may facilitate analyses aimed at less
abundant proteins.
ABBREVIATIONS
ACVA, 4,4′-azobis(4-cyanovaleric acid); BA, butylacrylamide;
BSA, bovine serum albumin; DHA, dihexylacrylamide; GPC, gel
permeation chromatography; HA, hexylacrylamide; HMPAM,
hydrophobically modified polyacrylamide; LPA, linear polyacryl-
amide; MALLS, multiangle laser light scattering; OA, octylacryl-
amide; PDI, polydispersity index; SDS, sodium dodecyl sulfate;
SPE, solid-phase extraction.
ACKNOWLEDGMENT
We gratefully acknowledge Dr. Alexander P. Sassi and Dr. Pin
Kao for useful discussions and technical support. Financial support
was provided by the Air Force Office of Scientific Research
(AFOSR), the Defense Advanced Research Projects Agency
(DARPA DURINT Grant No. F46920-01-1-0401), and the NSF
through the Northwestern University Nanoscale Science and
Engineering Center (Award No. EEC-0118025).
Received for review July 8, 2004. Accepted October 29,
2004.
AC049000Y
Analytical Chemistry, Vol. 77, No. 3, February 1, 2005 779