Novel quasi-interpenetrating network/functionalized multi-walled carbon nanotubes double-network composite matrices for DNA sequencing by CE

Research Article

Novel quasi-interpenetrating network/functionalized multi-walled carbonnanotubes double-network compositematrices for DNA sequencing by CE

Poly(N, N-dimethylacrylamide) (PDMA)-functionalized multi-walled carbon nanotubes

(MWNT-PDMA) were prepared via atom transfer radical polymerization and then added

into quasi-interpenetrating network (quasi-IPN) composed of linear polyacrylamide

(3.3 MDa) and PDMA to form polymer/nanotube double-network composite sieving

matrices for DNA sequencing by CE. The CE results show that, compared with quasi-

IPN, the novel composite matrices can improve ssDNA sequencing performances due to

the formation of a double-network consisting of a flexible quasi-IPN polymer network

and a rigid MWNT network based on a unique tubular structure, which makes the total

sieving networks more restricted and stable and increases the apparent molecular weight

of the matrices. The effects of MWNT-PDMA concentration in matrices and molecular

weight of PDMA side chains in MWNT-PDMA on ssDNA sequencing performances

were studied in detail. Furthermore, these double-network composite matrices were also

compared with other matrices and the results indicate that they are promising ones for

DNA sequencing. The separation provided with high resolution, speediness, excellent

reproducibility and easy loading owing to the addition of MWNT-PDMA is likely to

achieve full automation, especially for capillary array electrophoresis and microchip

electrophoresis.

Keywords:

CE / DNA sequencing / Double-network composite sieving matrices / Multi-walled carbon nanotubes / Quasi-interpenetrating network

DOI 10.1002/elps.200700925

1 Introduction

Separation and sequencing of DNA are vital to reveal genetic

code and sieving matrices play an important role in the

analysis of DNA by CE [1–3], which is one of the most

significant techniques for the separation and sequencing of

DNA [4–6]. In recent years, non-gel sieving matrices (i.e.

non-cross-linking polymer solutions) have been employed

widely in CE, which usually include linear homopolymers,

copolymers, mixtures, etc. [7–15]. For example, linear

polyacrylamide (LPA) with high molecular weight (MW)

possesses high sequencing ability and long read length [10,

11]. However, high-MW LPA solution is very viscous and

has no self-coating ability. On the contrary, poly(N, N-

dimethylacrylamide) (PDMA) shows excellent self-coating

ability but offers relatively poor sieving performance.

Therefore, a non-cross-linking quasi-interpenetrating

network (quasi-IPN) consisting of an LPA with very high

MW (up to 9.9 MDa) and PDMA was prepared as a high-

performance ssDNA sequencing medium, which could

combine the high sieving ability of LPA and the dynamic

coating ability of PDMA [16]. However, the viscosity of the

matrix containing LPA (9.9 MDa) is high and the prepara-

tion of long-chain LPA is not so easy. Therefore, searching

for the sieving matrices with low viscosity, high sieving

ability and self-coating ability still remains an important

issue for high-throughput DNA analysis [17].

Recently, certain nanoparticle additives (such as

montmorillonite clay [18], gold nanoparticles [17, 19–23],

polymer nanoparticles (nanosized PEGylated-latex) [24],

Dan ZhouLiping YangRunmiao YangWeihua SongShuhua PengYanmei Wang

Department of Polymer Scienceand Engineering, University ofScience and Technology ofChina, Hefei, P. R. China

Received December 18, 2007Revised March 20, 2008Accepted March 23, 2008

Abbreviations: ATRP, atom transfer radical polymerization;

DMA, N, N-dimethylacrylamide; LPA, linear polyacrylamide;

MW, molecular weight; MWNT, multi-walled carbonnanotube; MWNT-Br, bromoisobutyrate group-functionalized MWNT; MWNT-COOH, carboxyl group-functionalized MWNT; MWNT-PDMA, poly(N, N-dimethylacrylamide)-functionalized MWNT; PDMA, poly(N,N-dimethylacrylamide); POP-6, performance optimizedpolymer; quasi-IPN, quasi-interpenetrating network; TGA,

thermogravimetric analysis; TTE, Tris-TAPS-EDTA buffer

Correspondence: Professor Yanmei Wang, Department ofPolymer Science and Engineering, University of Science andTechnology of China, 96 Jinzhai Rd., Hefei 230026, P. R. ChinaE-mail: [email protected]: 186-551-3601592

& 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.electrophoresis-journal.com

Electrophoresis 2008, 29, 4637–4645 4637

bacterial cellulose fibrils [25], carbon nanotubes [26]) were

incorporated into low-viscosity polymer solutions, which has

been proved to be a very efficient and simple method to

overcome the difficulty of filling capillaries and improve

dsDNA separation performance due to their unique prop-

erties. Although various additives used to separate dsDNA

have been investigated in the past several years, the study on

additives for ssDNA sequencing is very deficient at present.

Hence, we aim at improvement of the performances of

ssDNA sequencing by using additives. Multi-walled carbon

nanotubes (MWNT) have recently attracted considerable

attention because they possess high surface area, unique

nanostructure, excellent thermal and chemical stability,

significant mechanical strength and high electrical conduc-

tivity. However, since the crude MWNT are large molecules

with thousands of carbon atoms in an aromatic delocalized

system, they are practically insoluble in most solvents and

incompatible with chemical and biological systems [27].

Consequently, it is difficult to handle or use them in CE

directly and extensive research will be focused on the surface

modification of the crude MWNT mainly to enhance their

compatibility and dissolution properties. Several papers

about the application of acid-treated MWNT (MWNT-

COOH) in CE have been reported [26, 28–30], but the

solubility and compatibility of MWNT-COOH are still not so

good. Therefore, we consider that MWNT functionalized by

polymers, which are soluble in and compatible with biolo-

gical systems, may be the potential additives for the

separation and sequencing analysis of DNA by CE, and the

relevant reports are scarce. Living radical polymerization,

especially atom transfer radical polymerization (ATRP)

‘‘grafting from’’ technique, is the method used most to graft

polymer chains with controlled MW from substrates (such

as silicon wafers, gold particles, long polymer backbones

and MWNT) [31, 32]. The ‘‘grafting from’’ technique

involves the immobilizing of initiators onto the substrates

followed by in situ surface-initiated polymerization to

generate the tethered polymer chains [33].

In this work, in order to improve the ssDNA sequencing

properties using quasi-interpretating network (quasi-IPN)

formed by LPA with lower MW (3.3 MDa) and PDMA and

prepare novel sieving matrices, we tried to add surface-

modified MWNT (poly(N, N-dimethylacrylamide)-functio-

nalized MWNT (MWNT-PDMA)) prepared through ATRP

of N, N-dimethylacrylamide (DMA) into this quasi-IPN to

form polymer/nanotube double-network composite matri-

ces (quasi-IPN/MWNT-PDMA), similar to the double-mesh

concept described previously [25]. The reasons for selecting

PDMA to functionalize MWNT are that PDMA is soluble in

water and compatible with biological components and quasi-

IPN; furthermore, PDMA has coating ability so as to reduce

EOF and the adsorption of DNA on the capillary wall effi-

ciently. The effects of MWNT-PDMA concentration in

matrices and MW of PDMA side chains in MWNT-PDMA

on ssDNA sequencing performances were studied in detail

by CE in the bare fused-silica capillaries, and the possible

functions of MWNT-PDMA in DNA sequencing were

discussed. Additionally, quasi-IPN/MWNT-PDMA matrices

were compared with other matrices and the reproducibility

of sequencing was also investigated.

2 Materials and methods

2.1 Materials

Crude MWNT were purchased from Sun Nanotech

(Nanchang, China). HNO3, thionyl chloride (SOCl2), glycol,

a-bromoisobutyryl bromide, 2-dimethylaminoprydine,

triethylamine, tetrahydrofuran, chloroform, DMF, ether

and CuBr were obtained from Sinopharm Chemical

Reagents (Shanghai, China) and distilled or purified before

use. 5, 5, 7, 12, 12, 14-Hexamethyl-1, 4, 8, 11-tetraazama-

crocyclotetradecane was prepared according to the route [34,

35]. The details of BigDye Terminator kit V3.1 sequencing

standard DNA sample and other reagents used for the

preparation of quasi-IPN and 1� TTE buffer (50 mM Tris/

50 mM TAPS/2 mM EDTA in water) or 1� TTE/7 M urea

buffer can be seen in the previous works [36, 37].

2.2 Preparation and characterization of quasi-IPN/

MWNT-PDMA

The synthetic procedures of quasi-IPN have been reported

elsewhere [36].

The general strategy for preparing initiator (bromoiso-

butyrate group-functionalized MWNT, MWNT-Br) used for

ATRP of DMA includes four steps [31], as shown in Fig. 1:

the crude MWNT was firstly oxidated by 60% HNO3 to

1) HNO3COOH

COOH

2) SOCl23) HOCH2CH2OH

Br Br

O4)

COOCH2CH2OOCC BrCH3

CH3

COOCH2CH2OOCC BrCH3

CH3

COOCH2CH2OOCCCH3

CH3

CH2 CH Brm

C=ON

H3C CH3

COOCH2CH2OOCCCH3

CH3

CH

CH

2 CH Brm

C=ON

H3 3

DMA

CuBrMe6[14]aneN4

MWNT MWNT-COOH MWNT-Br MWNT-PDMAC

Figure 1. Schematic representation of preparation of MWNT-COOH, MWNT-Br and MWNT-PDMA via ATRP.

Electrophoresis 2008, 29, 4637–46454638 D. Zhou et al.


obtaining carboxyl-contained MWNT (MWNT-COOH),

which was then reacted with SOCl2, glycol and a-bromoi-

sobutyryl bromide in sequence to generating MWNT-Br.

The in situ ATRP approach to polymer-functionalization of

MWNT is described as follows: 50 mg of MWNT-Br initia-

tor, 14.5 mg of CuBr catalyst, 86 mg of 5, 5, 7, 12, 12,

14-hexamethyl-1, 4, 8, 11-tetraazamacrocyclotetradecane

ligand, a certain amount of DMA monomer (distilled under

reduced pressure before use) and DMF solvent were placed

in a dried tube and degassed by freeze-pump-thaw cycles

three times. Then the tube was sealed under vacuum,

immersed in a thermostated oil bath at 601C immediately

and kept stirring for 20 h. After polymerization, the mixture

was diluted with a large amount of DMF and filtrated

through a 0.45 mm PTFE membrane three times to ensure

that no any possible un-grafted polymer and free reagents

were mixed in the product. The resulting solid was then

redispersed in 5 mL of DMF and precipitated in 100 mL of

ether. The MWNT-PDMA was obtained by filtration and

drying. Different amounts of DMA monomer (250, 500 and

1000 mg) were used for preparing MWNT-PDMA with

different-MW (or length) PDMA side chains, signed as

MWNT-PDMA1, MWNT-PDMA2 and MWNT-PDMA 3,

respectively.

Various quasi-IPN/MWNT-PDMA double-network

composite sieving matrix solutions were prepared by mixing

quasi-IPN and MWNT-PDMA with 1� TTE buffer/7 M

urea to the desired concentrations: the same 2.5% w/v of

quasi-IPN but different amounts of MWNT-PDMA2, or the

same 2.5% w/v of quasi-IPN but different MWNT-PDMA

with different-MW side PDMA at the same neat MWNT

concentration of 0.038 mg/mL. In addition, quasi-IPN

solution without MWNT-PDMA and the solution with

MWNT-COOH were also prepared.

The samples were characterized using Ubbelohde visc-

ometer (intrinsic viscosity [Z] measured in 1� TTE/7 M

urea buffer at 301C), AVANCE 300 1H-NMR Spectrometer

(BRUKER BIOSPIN AG, Switzerland), Hitachi H-800

Transmission Electron Microscope (Hitachi High-Technol-

ogies, Japan), Bruker EQUINOX55 Fourier Transform

Infrared Spectrometer (Bruker, Germany), Shimadzu DTG-

60 H Thermal Analyzer (heating rate of 101C/min, nitrogen

flow of 30 mL/min, Shimadzu, Japan) and VG ESCALAB

MK II X-Ray Photoelectron Spectrometer with Mg (Ka)

X-rays (VG Scientific Instruments, England).

2.3 DNA sequencing by CE and data processing

Sequencing of standard DNA sample was carried out on an

ABI 310 PRISMTM Genetic Analyzer (Perkin-Elmer, Applied

Biosystems Division, USA) with four-color LIF detection.

Sequencing conditions: effective/total length of bare fused-

silica capillaries (Polymicro Technologies, Phoenix, AZ,

USA), 50/61 cm; id/od, 75/365 mm; sequencing electric field

strength, 150 V/cm; DNA electrokinetic injection, 41 V/cm

for 30 s; anode buffer, 1�TTE; cathode buffer, 1�TTE/7 M

urea; sequencing temperature, 501C. The four-color raw

data were collected and analyzed by ABI PRISM 310 Data

Collection Software and ABI PRISM DNA Sequencing

Analysis Software (the base-calling software of ABI 310

was not very suitable for our sequencing matrices),

respectively. The raw LIF data were also transformed

through the ABI-Browser software and Origin 7.5 software

(Microcal, Northampton, MA, USA) was used to extract

data from transformed data, which were fitted into

Gaussian peaks by using PeakFitTM 4.06 software (SPSS,

Chicago, USA). All the peak fittings had r240.99. In order

to quantify the separation performance of a matrix and

compare it with other matrices, the resolutions (R) of

selected nine pairs of DNA fragments with base length of

79/80, 98/99, 192/193, 344/345, 412/413, 514/515, 635/638,

828/830 and 924/927 were calculated according to the

equation [38], respectively.

3 Results and discussion

3.1 Preparation and characterization of quasi-IPN/

MWNT-PDMA

LPA (3.3 MDa) was produced through inverse emulsion

polymerization [36]. Quasi-IPN was formed by solution

polymerization of DMA in LPA aqueous solution, which is

a non-cross-linking network with a higher sieving ability

[16, 39, 40] different from a traditional cross-linking IPN

network. The molar ratio of acrylamide unit to DMA

unit for polymer versions in the resultant quasi-IPN,

estimated from the integral peak area ratio in 1H NMR

spectra, is about 30/1.

The crude MWNT are difficult to disperse in water and

sedimentation appears soon even after sonication. After

oxidation of MWNT with HNO3, polar carboxyl groups were

introduced onto the convex surface of MWNT; thus, MWNT-

COOH is partially soluble in water. However, MWNT-

PDMA was soluble in water or quasi-IPN solution with the

aid of ultrasonic agitation for 5 min and no precipitation was

observed from this solution even after 1 month, which

proves that MWNT-PDMA synthesized by ATRP method

enhances the solubility of MWNT. The parallel results can

also be observed from transmission electron microscopy

(TEM) images (Fig. 2): the crude MWNT are piled up

seriously, but the MWNT-PDMA is dispersed individually

and do not further agglomerate, indicating that the bundles

of original MWNT can be separated into individual tubes by

surface ATRP of DMA.

MWNT-PDMA was also characterized using 1H NMR,

and the characteristic peaks of PDMA are clearly found: the

methylene and methylidyne protons have chemical shifts of

�1.6 and�2.6 ppm, respectively, and the methyl protons have

a chemical shift from 2.8 to 3.2 ppm (split into three peaks).

FTIR spectra of the crude MWNT and MWNT-Br (figure

not shown) reveal that the characteristic stretching vibration

signal of C—O (ester carbonyl) at around 1730 cm�1 is

Electrophoresis 2008, 29, 4637–4645 CE and CEC 4639


hardly detected for the crude MWNT, but clearly observed

for the MWNT-Br. For this reason we conclude that the

initiator groups were covalently bound to the MWNT, not

just adsorbed. The characteristic absorption peaks of PDMA

such as C—O (acrylamide carbonyl) at 1630 cm�1 and C–H

at 2926 cm�1 can be found in FTIR spectra of MWNT-

PDMA, which indicates that PDMA side chains have been

grafted successfully from MWNT via ATRP.

Because the grafted moieties and MWNT have distinct

thermal stabilities, thermogravimetric analysis (TGA) results

can give further evidence regarding the content and species

of the moieties grafted on MWNT. From TGA data, the

crude MWNT sample is steady without significant weight

loss below 5001C, whereas the MWNT-Br sample displays

weight loss of �14.6%, and this weight loss results from the

losing of initiator groups on the surface of MWNT and is

consistent with the content of initiator groups on the

MWNT, which corresponds to about 0.71 mmol bromoiso-

butyrate groups per gram of neat MWNT, or about 8.6

bromoisobutyrate groups per 1000 carbons. The loss-weight

fractions (fwt) of the polymer layers for the MWNT-PDMA

ranging from about 50 to 70% (Table 1) were determined

from TGA data and increased with an increase in the weight

ratio (Rwt) of DMA monomer to MWNT-Br initiator, indi-

cating that the contents (or MW) of the polymer side

chains on the functionalized MWNT can be controlled by

Rwt. Table 1 also shows the average MW of grafted PDMA

side chains calculated from TGA data. Furthermore, the

decomposition temperature of the functionalized moieties of

MWNT-Br is higher than the boiling point of a-bromoiso-

butyryl bromide (1621C); similarly, the major decomposition

temperature (320–4501C) corresponding to the surface

grown PDMA on MWNT is higher and wider than that of

pure PDMA. All these results indicate that a-bromoisobu-

tyryl bromide and PDMA were covalently linked to MWNT,

which is consistent with the previous work [31].

In addition, X-ray photoelectron spectroscopy analysis

was employed to determine the composition of the surfaces

of MWNT-Br and MWNT-PDMA. From Fig. 3A, the peaks

at the binding energy of about 70.16, 284.59 and 533.48 eV

are assigned to Br3d, C1s and O1s, respectively. The mole

content of the ATRP initiator groups on the surface of

MWNT is about 0.74% (mol) with respect to carbon

according to the peak area of each element, which is close to

the result obtained from TGA data. All these facts also verify

that bromoisobutyrate groups have been linked onto the

MWNT. Figure 3B shows the strong N1s peak (399.52 eV),

indicating the formation of the PDMA brushes on the

MWNT, but Br3d peak does not appear and the possible

reason is that the content of Br in MWNT-PDMA is too low

to detect.

To study the effect of addition of MWNT-PDMA on

matrix solution properties, the intrinsic viscosities ([Z]) of

quasi-IPN and composite matrices in 1� TTE/7 M urea

buffer were measured, as listed in Table 2. The overlap

concentration (c�E1/[Z]), or the concentration of the

polymer in solution at which polymer chains interact with

each other in solution, is a critical measure of the extent

of physical entanglements within a polymer solution, which

in turn is critical to the DNA sequencing performance of a

matrix. The ratio of polymer concentration to polymer

Table 1. Preparation and properties of MWNT-PDMA

Sample Rwta) fwt (%)b) MWc)

MWNT-PDMA1 5:1 50 1400

MWNT-PDMA2 10:1 62 2300

MWNT-PDMA3 20:1 70 3300

a) The weight ratio of DMA monomer to MWNT-Br initiator.

b) The loss-weight fraction of grafted PDMA on MWNT-PDMA

calculated from TGA data.

c) The average MW of grafted PDMA side chains on MWNT-

PDMA calculated from TGA data: MW 5 fwt/((1�fwt)�0.71�10�3); herein, 0.71 represents the concentration of

ATRP initiator sites per gram of neat MWNTs (mmol/g). In

order to compare conveniently, we assume that every ATRP

initiator site participates in the polymerization reaction;

hence, the calculated MW is generally smaller than the actual

value.

Figure 2. TEM images of (A) crude MWNT in water and (B)MWNT-PDMA in quasi-IPN solution.

0 100 200 300 400 500 600 700 800 900

0 100 200 300 400 500 600 700 800 900

0

60000

120000

180000

240000

300000

360000

0

40000

80000

120000

160000

200000

Binding Energy (eV)

Br3d

O1s

C1s

Rel

ativ

e In

tens

ity (

cps)

MWNT-Br

O1s

N1s

C1sB MWNT-PDMA

A

Figure 3. X-ray photoelectron spectroscopy spectra of (A)MWNT-Br and (B) MWNT-PDMA.



overlap concentration (c/c�), i.e. the extent of poly-

mer–polymer entanglements, controls the lifetime of the

virtual polymer ‘‘tube’’ so that the DNA migrates through

while under the electric field [41]. Compared with that of

quasi-IPN, intrinsic viscosities of quasi-IPN/MWNT-

PDMA2 matrices (i.e. quasi-IPN/MWNT-PDMA2-I, 2-II and

2-III with the neat MWNT concentrations of 0.002, 0.038

and 0.190 mg/mL, respectively) all increased a little

(Table 2); consequently, c� decreased and c/c� increased,

which suggests the formation of the double-network and the

increase in apparent MW and stability of the total sieving

networks because of the coexistence of flexible quasi-IPN

and rigid MWNT. Moreover, the extent of inter-chain

entanglements (c/c�) will be enhanced with the increase in

MWNT-PDMA concentration. Dolnik et al. also showed that

the potentially best sieving polymers are those with high

intrinsic viscosity, which can be used for the first selection

of sieving polymers before DNA sequencing [42].

3.2 Effects of MWNT-PDMA concentration

on sequencing

In order to find out the relationship between the resolutions

and the MWNT-PDMA concentrations in the matrix,

different amounts of MWNT-PDMA2 were incorporated

into quasi-IPN solutions (2.5% w/v) to form double-network

composite matrices, i.e. quasi-IPN/MWNT-PDMA2-I, 2-II

and 2-III with the neat MWNT concentrations of 0.002,

0.038 and 0.190 mg/mL, respectively. Figure 4 shows a part

(e.g. yellow-track, base T) of four-color electropherograms of

DNA sample using quasi-IPN/MWNT-PDMA2-II as

sequencing matrix at 501C. Resolution comparisons of

selected DNA fragments using 2.5% w/v quasi-IPN as well

as three quasi-IPN/MWNT-PDMA2 matrices at 501C are

shown in Fig. 5. Obviously, the resolutions of three quasi-

IPN/MWNT-PDMA2 matrices are all higher than those of

quasi-IPN without MWNT-PDMA2. In quasi-IPN matrix,

the linear polymer chains entangle one another to form a

transient network with a certain mesh size (pore size)

exhibiting molecular sieving property. The DNA molecules

migrate, collide and entangle with the matrix polymer

chains under the action of electric field during sequencing;

thus the polymer network possibly becomes loose due to the

slide of the polymer chains, resulting in a lower sieving

ability [1]. But the sequencing performances can be

improved when MWNT-PDMA was added into quasi-IPN

matrix because of the formation of a double-network

consisting of a flexible quasi-IPN polymer network and a

rigid MWNT network based on a unique tubular structure

(Fig. 6), which can afford additional interaction sites for

DNA molecules and is similar with the previous works [25,

26]. On the one hand, MWNT-PDMA can act as additional

obstacles (stereo effect) and these two different types of

networks can coexist and interact in matrix solution and

thus make the total sieving networks more restricted and

stable, the disentanglement of the polymer chains slower

and the apparent MW higher, which are helpful for long-

chain DNA sequencing. In addition, PDMA existing on both

MWNT-PDMA and quasi-IPN can improve the compat-

ibility of the total matrix system, and the PDMA side chains

on MWNT-PDMA may entangle with homo LPA or PDMA

in quasi-IPN to further stabilize the matrix network, as

shown in Fig. 6. On the other hand, the addition of MWNT-

PDMA enhances the separation of small DNA molecules

due to the double-network exhibiting smaller pore size than

quasi-IPN. Finally, the enhancement of long and small DNA

sequencing by the addition of MWNT-PDMA may be due to

not only the stereo effect but also more restricted, stable and

smaller nanopore structure. Moreover, it is also noted that

high efficiency, reproducible separation results and better

peak shapes are partially due to minimized DNA adsorption

on the capillary wall in the presence of MWNT-PDMA.

Nanotubes can serve as large surface area platforms, which

can interact with the capillary surface [30]. Moreover, PDMA

side chains on MWNT-PDMA have coating ability; thus, the

MWNT-PDMA that was adsorbed on the wall can restrain

the interaction between DNA and the wall.

From Fig. 5, we can also find that the resolutions of

quasi-IPN/MWNT-PDMA2 increase with an increase in

MWNT concentration up to 0.038 mg/mL. It should be

noted that a rigid network may not form when it is at very

Table 2. The intrinsic viscosities, the overlap concentrations and the extent of entanglements of quasi-IPN, quasi-IPN/MWNT-PDMA

matrices and quasi-IPN-H in 1� TTE/7 M urea buffer solution

Intrinsic viscosities Overlap concentration Extent of entanglementsb)

[Z] (mL/mg) c� (mg/mL)a) c/c�

Quasi-IPNc) 0.799 1.252 20.0

Quasi-IPN/MWNT-PDMA2-I 0.832 1.202 20.8

Quasi-IPN/MWNT-PDMA2-II 0.855 1.170 21.4

Quasi-IPN/MWNT-PDMA2-III 0.864 1.157 21.6

Quasi-IPN-Hd) 1.293 0.773 32.3

a) c�E1/[Z].

b) cE2.5% w/v, i.e. 25 mg/mL.

c) Containing LPA with lower MW of 3.3 MDa.

d) Containing LPA with higher MW of 6.5 MDa.



low MWNT concentration, but in the presence of MWNT-

PDMA, the entanglement of LPA chains with the carbon

nanotubes or PDMA side chains will still lead to more stable

polymer network. Therefore, even low-concentration

MWNT-PDMA can improve DNA sequencing performances

and it is also expected that the matrix network will become

more stable as the amount of MWNT-PDMA increases due

to the gradual formation of a double-network. However, it is

not easy to determine the exact concentration at which the

rigid network is formed in our complex systems.

However, the resolutions will not always improve with

increased MWNT-PDMA concentration and further increase

in concentration (i.e. MWNT higher than 0.038 mg/mL)

even results in decreased resolutions. The very high

concentration of MWNT-PDMA in the solution probably

leads to form local aggregation with very small pores and

blocks the migration of DNA fragments [25]. This aggrega-

tion could then cause non-homogeneous matrix–DNA

interactions and decrease the efficiency in sequencing. To

some extent, the very black color of MWNT at very high

concentration will interfere with LIF detection and augment

the baseline noise [26].

The presence of MWNT-PDMA also change migration

time of DNA fragments slightly, for instance, quasi-IPN and

quasi-IPN/MWNT-PDMA2-II resulted in 1000 bases being

detected in 78.1 and 75.5 min (Table 3), respectively. One

reason for the alteration in migration time is likely that

some MWNT-PDMA is adsorbed on the capillary wall

surface, leading to the decrease in EOF and thus the

increase in DNA mobility; the other reason is that the

double-network with smaller pore size will make DNA

molecules migration slower. These two contrary factors

affect DNA movement together and result in the change in

migration time. Also, the readlength at 98% accuracy of

quasi-IPN increases (from 715 to 792 bases) because of the

addition of MWNT-PDMA (Table 3).

54 56 58 60 62 64 66 68 70 72 74 76 784500

5000

5500

6000

650028 30 32 34 36 38 40 42 44 46 48 50 52

4500

5000

5500

6000

6500

4 6 8 10 12 14 16 18 20 22 24 26 28

450050005500600065007000

Migration time (min)

846 10

00

Flu

ores

cenc

e in

tens

ity

385

526 57

6

Yellow, T139

266

Figure 4. A part (e.g. yellow-track, base T) offour-color electropherograms of BigdyeTerminator V 3.1 sequencing standardDNA sample by CE using quasi-IPN/MWNT-PDMA2-II as the sequencing matrix.Sequencing conditions: effective/totallength of bare fused-silica capillaries (Poly-micro Technologies), 50/61 cm; id/od, 75/365 mm; sequencing electric field strength,150 V/cm; DNA electrokinetic injection,41 V/cm for 30 s; anode buffer, 1� TTE;cathode buffer, 1� TTE/7 M urea; sequen-cing temperature, 501C; solution concentra-tion, 2.5% w/v quasi-IPN10.038 mg/mL neatMWNT for quasi-IPN/MWNT-PDMA2-II.

0 200 400 600 800 1000

0.4

0.6

0.8

1.0

1.2

Res

olut

ion

Base number

quasi-IPN quasi-IPN/MWNT-PDMA2-I quasi-IPN/MWNT-PDMA2-II quasi-IPN/MWNT-PDMA2-III

Figure 5. Resolution versus base number in DNA sequencing ofDNA sample by CE using quasi-IPN, quasi-IPN/MWNT-PDMA2-I,2-II and 2-III. Solution concentration: 2.5% w/v quasi-IPN10,0.002, 0.038 and 0.190 mg/mL neat MWNT for quasi-IPN, quasi-IPN/MWNT-PDMA2-I, 2-II and 2-III, respectively. Other sequen-cing conditions and DNA sample are as in Fig. 4.

+

MWNT-PDMA quasi-IPNquasi-IPN/MWNT-PDMAdouble-network matrix

Figure 6. Schematic representation of the formation of quasi-IPN/MWNT-PDMA double-network composite sieving matrix.



3.3 Effects of MW of PDMA side chains in MWNT-

PDMA on sequencing

For the purpose of studying the effects of side PDMA MW

in MWNT-PDMA on DNA sequencing performances, three

types of MWNT grafted with different-MW PDMA were

prepared by adjusting the weight ratio of DMA monomer to

MWNT-Br initiator. In addition, MWNT without PDMA (e.g.MWNT-COOH) were also used for comparison. We did not

select crude MWNT or MWNT-Br but MWNT-COOH

because the solubility of MWNT-COOH in buffer solution

is better than those of the former. It is observed from Fig. 7

that when the neat MWNT concentration in quasi-IPN/

MWNT-COOH, quasi-IPN/MWNT-PDMA1, 2 and 3 was

fixed at 0.038 mg/mL, the effects of higher PDMA MW on

resolutions are more significant than those of lower MW.

Functionalized MWNT with longer PDMA chains on the

surface possess higher compatibility with quasi-IPN system

and higher solubility in water; on the other hand, longer

PDMA chains result in easier entanglement between

MWNT-PDMA themselves or between MWNT-PDMA and

quasi-IPN system. All these factors cause the double-

network with longer PDMA side chains more stable and

thus lead to increase in resolutions. Although the functio-

nalized MWNT without PDMA side chains exhibits much

lower resolutions than those with PDMA side chains,

addition of MWNT-COOH can improve DNA sequencing

performances because of a double-network.

3.4 Comparison with other matrices

It is well known that using only LPA in bare capillary cannot

separate DNA due to LPA without coating ability, but DNA

sequencing can be obtained by addition of MWNT-PDMA

(data not shown), indicating that MWNT-PDMA possess

coating property. But the resolutions using LPA/MWNT-

PDMA are much worse than those using quasi-IPN/

MWNT-PDMA. The possible reasons are that, compared

with PDMA chains in quasi-IPN, PDMA side chains on

MWNT-PDMA in LPA/MWNT-PDMA are much shorter

and fewer and thus the efficiency of absorbing on the wall

surface and entangling with LPA chains is much lower.

Hence, quasi-IPN/MWNT-PDMA shows better DNA

sequencing performances.

The concentration and viscosity of sieving matrix are

also important for DNA sequencing performances and

matrix application. As can be seen from Fig. 8, compared

with quasi-IPN (2.5% w/v), quasi-IPN/MWNT-PDMA2

(2.0% w/v10.038 mg/mL) shows close resolutions for DNA

fragments smaller than 192 bases owing to lower quasi-IPN

concentration and viscosity unsuitable to separate small

DNA fragments, but displays obviously higher resolutions

for large DNA fragments due to the presence of the double-

network. In addition, as a result of lower concentration

and thus lower viscosity, the migration time of about 1000-

base DNA fragments using quasi-IPN/MWNT-PDMA2

was much shorter than that using quasi-IPN (about 66

and 78 min, respectively). Thus, the addition of MWNT-

PDMA into quasi-IPN can lead to more excellent

sieving performances in terms of resolution and

migration time.

Table 3. Comparison of separation conditions, migration times and obtained readlengths between prepared and existing matrices

Migration time (min)

Concentration (% w/v1mg/mL) Electric field strengtha) (V/cm) Base 620 Base 1000 Readlength at 98%

accuracy (bases)

Quasi-IPN 2.510 150 51.5 78.1 715

Quasi-IPN/MWNT-PDMA2 2.010.038 150 43.7 66.0 649

Quasi-IPN/MWNT-PDMA2 2.510.038 150 49.2 75.5 792

Quasi-IPN-H 2.510 150 52.4 80.1 785

Commercial POP-6b) – 200 107.0 – 638

a) Other sequencing conditions and DNA sample as in Fig. 4.

b) We do not make certain the concentration of POP-6; the migration using POP-6 is so slow and the resolution for base 1000 is so low

that we did not obtain the exact migration time for base 1000 by using POP-6.

0 200 400 600 800 1000

0.4

0.6

0.8

1.0

1.2

Res

olut

ion

Base number

quasi-IPN quasi-IPN/MWNT-COOH quasi-IPN/MWNT-PDMA1 quasi-IPN/MWNT-PDMA2 quasi-IPN/MWNT-PDMA3

Figure 7. Resolution versus base number in DNA sequencing byCE using quasi-IPN, quasi-IPN/MWNT-COOH, quasi-IPN/MWNT-PDMA1, 2 and 3 with the same neat MWNT concentration of0.038 mg/mL. Other sequencing conditions and DNA sample areas in Fig. 4.



The MW of LPA can determine the sieving perfor-

mances to a great extent, for example, the resolutions will

increase with LPA MW increase, but at the same time the

viscosity will also increase. Quasi-IPN containing LPA with

higher MW of 6.5 MDa (denoted as quasi-IPN-H) were also

prepared and used for further comparison and demonstra-

tion of the function of MWNT-PDMA. It is shown from

Fig. 8 that when both quasi-IPN concentrations are

2.5% w/v, the resolutions of quasi-IPN/MWNT-PDMA2

with lower LPA MW of 3.3 MDa are even slightly higher

than those of quasi-IPN-H without MWNT-PDMA although

the polymer with higher intrinsic viscosity may be a better

sieving matrix [42] (the detailed values of [Z] can be seen in

Table 2), further indicating that the double-network might

form in the composite sieving matrices, which would form a

more robust sieving matrix network, increase apparent MW

of quasi-IPN and consequently improve sieving properties

based on separation mechanisms. The readlength at 98%

accuracy of quasi-IPN/MWNT-PDMA2 is also slightly

higher than that of quasi-IPN-H (792 versus 785 bases), as

shown in Table 3. Hence, we do not have to use LPA with

high MW, which is difficultly prepared and very viscous.

Moreover, lower viscosity as a result of lower MW resulted

in close resolutions but less migration time of 1000 bases:

75.5 versus 80.1 min for using quasi-IPN/MWNT-PDMA2

versus quasi-IPN-H, respectively.

Performance optimized polymer (POP-6) from Applied

Biosystems is a commercial product used widely for ABI 310

Genetic Analyzer; thus, we select and compare it with our

double-network matrices. As can be seen from Fig. 8, quasi-

IPN/MWNT-PDMA2 shows better resolutions than POP-6

with the exception of small DNA fragments (smaller than

about 250 bases) owing to higher solution concentration of

POP-6 helpful to separate small DNA fragments. The

former possesses longer readlength (792 versus 638 bases at

98% accuracy) and much shorter migration time than POP-

6 (the migration time was 49.2 and 107.0 min for base 620,

respectively). All the above results demonstrate that quasi-

IPN/MWNT-PDMA double-network composite matrices are

potential for DNA sequencing.

3.5 Reproducibility

The RSD of the migration time and resolution using quasi-

IPN/MWNT-PDMA as matrices in the same bare fused-

silica capillary for the first ten runs are all less than 2.5%.

The high reproducibility is attributed to the self-coating

ability of PDMA and the adsorption of MWNT-PDMA on

the capillary inner wall, which can suppress EOF and avoid

interactions between DNA and capillary wall.

4 Concluding remarks

The DNA sequencing is enhanced by addition of MWNT-

PDMA prepared by ATRP into quasi-IPN composed of LPA

and PDMA to form polymer/nanotube composite sieving

matrices. Without complete optimization (such as base

calling software), quasi-IPN/MWNT-PDMA2-II yielded a

readlength of 792 bases at 98% accuracy in about 62 min by

using the ABI 310 Genetic Analyzer at 501C and 150 V/cm.

The study results show that the double-network might form

in composite sieving matrix, which consists of a flexible

quasi-IPN polymer network and a rigid MWNT network

based on a unique tubular structure. These two different

types of networks can coexist and interact in matrix solution,

prevent the polymer chains from sliding away from each

other, stabilize and restrict the total sieving networks,

increase the apparent MW of the matrices and reduce the

pore size of matrices. Furthermore, the PDMA side chains

on MWNT-PDMA may entangle with homo LPA or PDMA

in quasi-IPN to further stabilize the matrix network.

Therefore, more restricted, stable and smaller nanopore

structure in quasi-IPN/MWNT-PDMA matrix results in

more excellent properties. Additionally, minimized DNA

adsorption on the capillary wall due to MWNT-PDMA

adsorbed on the wall, leading to decrease in EOF, is also one

of the reasons for high performances.

Comparative studies between quasi-IPN or quasi-IPN/

MWNT-PDMA and other matrices (i.e. LPA/MWNT-PDMA,

quasi-IPN/MWNT-PDMA with lower quasi-IPN concentra-

tion, quasi-IPN with higher-MW LPA and commercial POP-

6) further indicate that the double-network composite

matrices seem able to combine optimal sieving ability and

dynamic coating ability with moderate viscosity. The

separation of DNA by using low-viscosity composite matri-

ces due to low-MW LPA and low concentration is provided

with high resolution, speediness, excellent reproducibility

and easy loading by addition of MWNT-PDMA, which helps

to achieve full automation, especially for capillary array

electrophoresis and microchip electrophoresis.

0 200 400 600 800 1000

0.4

0.6

0.8

1.0

1.2

1.4

Res

olut

ion

Base number

quasi-IPN (2.5%) quasi-IPN/MWNT-PDMA2(2.0%+0.038mg/mL) quasi-IPN/MWNT-PDMA2(2.5%+0.038mg/mL) quasi-IPN-H (2.5%) POP-6

Figure 8. Resolution versus base number in DNA sequencing byCE using quasi-IPN (2.5% w/v), quasi-IPN/MWNT-PDMA2(2.0% w/v10.038 mg/mL), quasi-IPN/MWNT-PDMA2 (2.5% w/v1

0.038 mg/mL), quasi-IPN-H (2.5% w/v) and commercial POP-6(electric field strength of 200 V/cm for POP-6). Other sequencingconditions and DNA sample are as in Fig. 4.



Other parameters will be optimized in the future to

further improve the DNA sequencing performances, and one

of our interests is to use functionalized MWNT to separate

proteins in CE. Furthermore, other new additives will also be

tried to promote the separation of biomacromolecules.

We greatly acknowledge the support of this work by theNational Natural Science Foundation of China (Grant No.50773074), Ministry of Science and Technology of China(Grant No. 2007CB936401), the Foundation for Developmentof Talent of Anhui Province (Grant No. 2005Z026) and theScientific Research Foundation for the Returned OverseasChinese Scholars, State Education Ministry.

The authors declare no financial or commercial conflict ofinterest.

5 References

[1] Chu, B., Liang, D. H., J. Chromatogr. A 2002, 966, 1–13.

[2] Ashton, R., Padalay, C., Kane, R. S., Curr. Opin.Biotechnol. 2003, 14, 497–504.

[3] Albarghouthi, M. N., Buchholz, B. A., Huiberts, P. J.,Stein, T. M., Barron, A. E., Electrophoresis 2002, 23,1429–1440.

[4] Sartori, A., Barbier, V., Viovy, J. L., Electrophoresis 2003,24, 421–440.

[5] Albarghouthi, M. N., Barron, A. E., Electrophoresis 2000,21, 4096–4111.

[6] Liang, D. H., Liu, T. B., Song, L. G., Chu, B., J. Chro-matogr. A 2001, 909, 271–278.

[7] Ruiz-Martinez, M. C., Berka, J., Belenkii, A., Foret, F.et al., Anal. Chem. 1993, 65, 2851–2858.

[8] Best, N., Arriaga, E., Chen, D. Y., Dovichi, N. J., Anal.Chem. 1994, 66, 4063–4067.

[9] Goetzinger, W., Kotler, L., Carrilho, E., Ruiz-Martinez,M. C. et al., Electrophoresis 1998, 19, 242–248.

[10] Salas-Solano, O., Carrilho, E., Kotler, L., Miller, A. W.et al., Anal. Chem. 1998, 70, 3996–4003.

[11] Zhou, H. H., Miller, A. W., Sosic, Z., Buchholz, B. et al.,Anal. Chem. 2000, 72, 1045–1052.

[12] Madabhushi, R. S., Electrophoresis 1998, 19, 224–230.

[13] Chang, H. T., Yeung, E. S., J Chromatogr. B 1995, 669,113–123.

[14] Gao, Q. F., Yeung, E. S., Anal. Chem. 1998, 70,1382–1388.

[15] Barron, A. E., Soane, D. S., Blanch, H. W., J. Chroma-togr. A 1993, 652, 3–16.

[16] Wang, Y. M., Liang, D. H., Ying, Q. C., Chu, B., Electro-phoresis 2005, 26, 126–136.

[17] Huang, M. F., Huang, C. C., Chang, H. T., Electrophoresis2003, 24, 2896–2902.

[18] Liang, D. H., Song, L. G., Chen, Z., Chu, B., Electro-phoresis 2001, 22, 1997–2003.

[19] Lin, Y. W., Huang, M. J., Chang, H. T., J. Chromatogr. A2003, 1014, 47–55.

[20] Chiou, S. H., Huang, M. F., Chang, H. T., Electrophoresis2004, 25, 2186–2192.

[21] Huang, M. F., Kuo, Y. C., Huang, C. C., Chang, H. T.,Anal. Chem. 2004, 76, 192–196.

[22] Lin, Y. W., Huang, M. F., Chang, H. T., Electrophoresis2005, 26, 320–330.

[23] Tseng, W. L., Huang, M. F., Huang, Y. F., Chang, H. T.,Electrophoresis 2005, 26, 3069–3075.

[24] Tabuchi, M., Katsuyama, Y., Nogami, K., Nagata, H.et al., Lab Chip 2005, 5, 199–204.

[25] Tabuchi, M., Baba, Y., Anal. Chem. 2005, 77,7090–7093.

[26] Xu, Y., Li, S. F. Y., Electrophoresis 2006, 27,4025–4028.

[27] Cui, J., Wang, W. P., You, Y. Z., Liu, C. H., Wang, P. H.,Polymer 2004, 45, 8717–8721.

[28] Wang, Z. H., Luo, G. A., Chen, J. F., Xiao, S. F.,Wang, Y. M., Electrophoresis 2003, 24, 4181–4188.

[29] Luong, J. H. T., Bouvrette, P., Liu, Y. L., Yang, D. Q.,Sacher, E., J. Chromatogr. A 2005, 1074, 187–194.

[30] Xiong, X., Ouyang, J., Baeyens, W. R. G., Delanghe, J. R.et al., Electrophoresis 2006, 27, 3243–3253.

[31] Kong, H., Gao, C., Yan, D. Y., J. Am. Chem. Soc. 2004,126, 412–413.

[32] Qin, S. H., Qin, D. Q., Ford, W. T., Resasco, D. E.,Herrera, J. E., J. Am. Chem. Soc. 2004, 126, 171–176.

[33] Zhao, B., Brittain, W. J., Prog. Polym. Sci. 2000, 25,677–710.

[34] Bodie, E. D. (Ed.), Inorganic Syntheses, Wiley, New York1978.

[35] Yang, R. M., Wang, Y. M., Wang, X. G., He, W. D.,Pan, C. Y., Eur. Polym. J. 2003, 39, 2029–2033.

[36] Zhou, D., Wang, Y. M., Zhang, W. L., Yang, R. M.,Shi, R. H., Electrophoresis 2007, 28, 1072–1080.

[37] Zhou, D., Wang, Y. M., Yang, R. M., Zhang, W. L.,Shi, R. H., Electrophoresis 2007, 28, 2998–3007.

[38] Giddings, J. C., Separ. Sci. 1969, 4, 181–189.

[39] Wang, Y. M., Liang, D. H., Hao, J. C., Fang, D. F., Chu, B.,Electrophoresis 2002, 23, 1460–1466.

[40] Song, L. G., Liu, T. B., Liang, D. H., Fang, D. F., Chu, B.,Electrophoresis 2001, 22, 3688–3698.

[41] Doherty, E. A. S., Kan, C. W., Barron, A. E., Electro-phoresis 2003, 24, 4170–4180.

[42] Dolnik, V., Gurske, W. A., Padua, A., Electrophoresis2001, 22, 692–698.



Documents

Novel quasi-interpenetrating network/functionalized multi-walled carbon nanotubes double-network composite matrices for DNA sequencing by CE