Intelligent thermoresponsive polymeric stationary phases for

aqueous chromatography of biological compounds

Akihiko Kikuchi, Teruo Okano*

Institute of Advanced Biomedical Engineering and Science, Tokyo Women's Medical University, 8-1Kawadacho, Shinjuku,

Tokyo 162-8666, Japan

Received 5 November 2001; revised 21 January 2002; accepted 25 January 2002

Abstract

Sensitive polymers with external physical, chemical, and electrical stimuli are termed as `intelligent materials'

and currently have been used in variety ®elds of engineering, and medicine. Numerous research papers utilizing

stimuli-responsive intelligent materials are found in the literature to date. In this manuscript, the authors described

several applications of surfaces and interfaces modi®ed with stimuli-responsive polymers for stimuli-responsive

surface property alteration and their application for the separation sciences. The special attention is paid to the

temperature responsive polymers, poly(N-isopropylacrylamide) (PIPAAm) and its derivatives as surface modi®ers

for novel `green' chromatography in which only aqueous mobile phase was utilized for separation of bioactive

compounds. Several factors were investigated and discussed the effects on separation of bioactive compounds;

these include the effects of the temperature-responsive hydrophilic/hydrophobic changes, copolymer composition,

graft polymer molecular architecture and the incorporation of charged groups. Furthermore, application of

PIPAAm-grafted surfaces for af®nity separation of proteins will be discussed. The technique has superior char-

acteristics in reducing organic wastes and costs to run chromatographic separation, and thus must be an envi-

ronmentally friendly separation tool. q 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Stimuli responsive polymer; Surface function; Biocompatibility; Hydrophilic/hydrophobic property alteration;

Charge density alteration; Chromatography; Penetration control; Aqueous mobile phase; Bioactive compounds; Separation;

Dye af®nity chromatography; Protein

Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1166

2. Molecular separation and permeation control through the stimuli-responsive pore size regulation . . . .1167

3. Thermoresponsive chromatography using poly(N-isopropylacrylamide) as column matrix . . . . . . . . . .1169

4. Temperature-responsive wettability changes of PIPAAm-modi®ed surfaces . . . . . . . . . . . . . . . . . . . .1172

Prog. Polym. Sci. 27 (2002) 1165±1193

0079-6700/02/$ - see front matter q 2002 Elsevier Science Ltd. All rights reserved.

PII: S0079-6700(02)00013-8

www.elsevier.com/locate/ppolysci

* Corresponding author. Tel.: [phone]x30233; fax: [phone].

E-mail address: [e-mail] (T. Okano).

5. Detachment control of cultured cells from PIPAAm-grafted culture dishes . . . . . . . . . . . . . . . . . . . . .1175

6. Novel `Green' chromatography utilizing PIPAAm-modi®ed surface . . . . . . . . . . . . . . . . . . . . . . . . . .1176

6.1. Effects of hydrophilic/hydrophobic property changes of PIPAAm surfaces on elution behavior . .1176

6.2. Effect of surface morphology of PIPAAm modi®ed matrix on the separation . . . . . . . . . . . . . . .1181

6.3. Effect of surface charged groups on analyte separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1183

7. Selective adsorption/desorption of bioactive proteins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1185

8. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1190

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1190

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1190

1. Introduction

Considerable researches have been focused on polymeric materials those change their structure and

functions responding to external physical, chemical, and electrical stimuli (light, temperature, pH,

substance concentration, solvent composition, and electric ®elds, etc.). These materials termed `intel-

ligent materials', sense one or more external stimuli (sensor) as signals, judge the magnitude of these

signals (processor), and change their structure and functions in direct response (effecter). The response

of the intelligent materials toward above described stimuli induce several kinds of changes such as

phase, shape, surface energies, permeation rates, reaction rates, and molecule recognition. Introduction

of stimuli-responsive polymeric materials as switching sequences into both arti®cial materials and

bioactive compounds (peptides, proteins, nucleic acids, and others) permit modulation of their structure

induced by corresponding external stimuli. `On±off' switching of their respective functions, thus, can be

achieved at molecular level [1±5]. Intelligent materials embodying these concepts might contribute to

establish fundamental principles for fabrication of novel systems.

In the present review, several approaches to create stimuli-responsive polymer-modi®ed surfaces and

interfaces are described. Special attention will be paid to the temperature-responsive polymer, poly(N-

isopropylacrylamide) (PIPAAm), which shows lower critical solution temperature (LCST) around 32 8C

in water [6,7]. As shown in Fig. 1, PIPAAm is soluble in water below the LCST, while precipitation of

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1166

Fig. 1. Structural formula of poly(N-isopropylacrylamide) (PIPAAm) and appearance of its aqueous solution below and above

the LCST.

polymer and solution turbidity are observed above this temperature. Introduction of PIPAAm onto solid

surfaces will produce intelligent surfaces those show temperature-responsive hydrophilic/hydrophobic

changes as described in detail in Section 4. Furthermore, this surface property alteration would be

utilized to modulate solute interaction in the chromatographic separation by changing temperature in

sole aqueous condition, creating novel liquid chromatography systems. These systems might be an

alternative of reversed-phase chromatography in which separation is controlled through the regulation

of the interaction between octadecylsilica stationary phase surfaces and solute molecules by mobile

phase composition and polarity.

Kanazawa et al. [8] recently reviewed results for temperature-responsive hydrophobic chromatogra-

phy in aqueous mobile phase from the standpoint of analytical science. In this review, the point is

focused on the intelligent interface preparation using stimuli-responsive polymeric materials and the

prepared surface application in the separation sciences.

2. Molecular separation and permeation control through the stimuli-responsive pore size

regulation

Polymer chain coil/globule transition with chemical and/or physical stimuli are utilized to control

pore size and applied for the stimuli-responsive molecular valve of substance diffusion and permeation

through the pores.

Iwata and Matsuda [9] reported the modi®cation of porous membranes with either poly(acrylic acid)

(PAAc) or polyacrylamide (PAAm) through air plasma-induced polymerization of corresponding mono-

mers. In case of PAAc as the membrane modi®er, the polymer molecules on the surfaces acted as the

environmental pH-sensing device, at the same time, as the molecular valve to regulate permeation

through the pores. Filtration rate at pH 5.2 was approximately one-tenth (1/10) of that at pH 1.4.

With decreasing the ®ltration rate modi®ed membrane showed ultra®ltration character for high mole-

cular weight substances, such as albumin and dextran. They further investigated the PAAc-modi®ed

membranes by means of atomic force microscope (AFM) to visualize pH-dependent pore size control

[10]. At lower pH than pK

a

of carboxyls in PAAc, pores on the modi®ed membrane are clearly observed

under AFM, while at higher pH carboxyl groups were deprotonated to become carboxylate anions and

polymer chains extended out of pores to make hill-and-valley structure in the AFM views. Thus, pH-

dependent polymer chain conformational changes caused open/close changes of the membrane pores,

regulating substrate permeation as schematically shown in Fig. 2a. The pH-dependent membrane

characteristic changes are reversible and reproducible, PAAc-modi®ed membrane could be utilized in

various membrane technologies.

Ito et al. [11,12] utilized same concept for pH-controlled drug delivery systems. They immobilized

PAAc onto the cellulose membrane. Using side chain carboxyl groups, glucose oxidase was co-immo-

bilized onto surface grafted PAAc chains via amide bond formation. Inside the membrane, insulin as

bioactive drug was reserved. At physiological pH of 7.4, where PAAc grafts are ionized through

deprotonation and existed in expanded conformation, drug permeation was almost stopped because

membrane pores were in closed state with expanded polymer chains. When glucose is presented in

the medium, glucose oxidase mediates the glucose conversion to gluconic acid, which reduces the

environmental pH inducing protonation of PAAc. With the protonation, PAAc chains form compact

conformation, which open the membrane pores. Thus, membrane permeation rate changes and

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1167

eventually insulin inside the modi®ed membrane was released out of the membrane. There are many

points to be solved before utilizing this material to in vivo application, such as in vivo response toward

glucose concentration, and prevention of denaturation and immune response toward conjugated glucose

oxidase. Further optimization of this system to in vivo application will provide blood glucose level for

diabetes mellitus patients.

By utilizing temperature as a stimulus, similar permeation control is achieved.

Okahata et al. [13] reported thermoresponsive drug release capsules (Fig. 2b). The capsule was made

from porous nylon membrane on which PIPAAm chains were grafted covalently. Sodium dinaphthalene

sulfonic acid was used as model drug molecules. Release of drug was controlled by temperature; release

enhanced at lower temperature where surface grafted PIPAAm was hydrated and extended form, while

signi®cant drug release suppression was achieved by raising temperature more than PIPAAm's LCST.

This was due to the dehydration of PIPAAm chains above its LCST. PIPAAm molecules precipitated out

on the nylon capsule surfaces those prevented drug molecule permeation through the pore. Thus,

temperature responsive open/close changes of membrane pores were achieved.

Iwata et al. [14] reported PIPAAm-grafted porous membranes for water ¯ow control through the pore

by temperature changes. They treated porous poly(vinylidene di¯uoride) membrane with Ar plasma

followed by IPAAm polymerization onto plasma-treated membrane surfaces. Above the PIPAAm's

transition temperature, 10 times higher ®ltration rate for water was observed while negligible ®ltration

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1168

Fig. 2. (a) Stimuli-responsive porous membrane to control pore size. (b) Diffusion control using temperature-responsive

PIPAAm-grafted nylon membrane capsules.

was occurred at lower temperature. In their case, shrunken polymer molecules on the membranes open

channels of pore for water to pass through. Thus, Fig. 2a represents this model.

These two examples showed sharp contrast in terms of thermoresponsive polymer conformation,

although both membranes were grafted with PIPAAm. The difference is probably the density of

PIPAAm chains on the membrane surfaces and/or the modi®cation sites, inside or outside the pores

as schematically shown in Fig. 2a and b.

In all cases described here, however, complete `on±off' control of molecule diffusion or penetration

was not achieved. Instead, background diffusion was apparent (1±2 type control). To achieve `0±1 type

on±off' control, more precise design of membranes should be necessary in terms of graft chain density,

length (molecular weight), and con®guration on the substrate surfaces.

Recognition site of substrate in enzymes could also be regulated with a similar concept [15±17] but

not discussed in detail here.

3. Thermoresponsive chromatography using poly(N-isopropylacrylamide)as column matrix

Above described concept of pore size control with stimuli-responsive polymers can be utilized for

chromatography systems, especially in size exclusion chromatography mode [18±21].

Gewehr et al. [18] prepared PIPAAm-grafted porous glass beads as column packing materials for gel

permeation chromatography. PIPAAm was end-functionalized through telomerization polymerization

of IPAAm in the presence of mercaptopropionic acid (MPA) used as chain transfer agent. By this

method, each chain end of PIPAAm molecules possessed carboxyl group. Porous glass beads were

®rstly aminated with 3-aminopropyltriethoxysilane followed by conjugation of PIPAAm with active

ester chain ends through amide bond formation. Bare porous glass beads with 156, 171, 237, and 408 AÊ

pores were used as base materials. The PIPAAm-modi®ed glass beads were packed into column

(200 mm long with 7 mm inner diameter) and elution of dextran having a variety of molecular weights

was examined changing column temperature. In GPC mode, smaller molecular weight substances show

longer retention time due to the permeation through the matrix. In case of PIPAAm-grafted porous glass

bead-packed columns, similar effects were seen; i.e. the smaller the molecular weight of dextran used as

model samples, the longer the retention was observed. Temperature change induced PIPAAm

chain conformational changes; hydrated, and extended conformation at low temperature below the

LCST to dehydrated, and shrunken aggregates above the LCST. Due to this conformational

change on the PIPAAm-grafted porous glass bead surfaces, effective pore size changed depending

on the original pore sizes. With smaller pore beads, large pore size change by temperature

change was occurred. In case of beads with 237 AÊ pores, discontinuous elution time shift was

observed for all dextran samples in the temperature range of 25±35 8C as shown in Fig. 3. The

results strongly suggested the temperature-modulated pore size changes, and thus, successful resolution

of wide molecular weight range of dextran in GPC mode. As was pointed out, balance of PIPAAm

molecular lengths and pore size of base materials are important to control permeation of substances to be

separated.

Hosoya et al. [19,20] developed an in situ surface-selective modi®cation to introduce PIPAAm into

porous polymer beads. They used porous polystyrene beads (1 mm diameter) as seed particles and

IPAAm polymerization was then carried out using either cyclohexanol or toluene as porogen agent in

water. In case of cyclohexanol used as porogen, propagating PIPAAm radical is soluble in cyclohexanol

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1169

and thus entire surfaces (both inside and outside) of the porous beads were covered with PIPAAm. On

the other hand, PIPAAm was grafted only on the external bead surfaces with toluene used as porogen

since PIPAAm cannot penetrate into toluene-®lled pores of the beads due to the existence of non-

solvent, toluene for PIPAAm. In this way, PIPAAm was selectively grafted on either entire surfaces

or only external surfaces of the beads. Microscopic appearance of the beads was relatively homogeneous

for modi®ed beads with cyclohexanol as porogen, and heterogeneous, rough surface morphology was

observed for modi®ed beads with toluene as porogen.

These surface structures in¯uenced on the elution behavior of dextran in aqueous systems. They

carried out size exclusion chromatography using these two sets of modi®ed beads as column matrixes.

When cyclohexanol was used as porogen agent, temperature increase affects to longer elution time for

higher molecular weight dextran. At temperature below PIPAAm's LCST, PIPAAm chains exist in

expanded conformation due to hydration of PIPAAm chains, restricting dextran molecules to penetrate

into deep inside of the pores. At higher temperature, PIPAAm dehydrated, and shrunken at the interface,

opening pores wide, which permit dextran to penetrate into the pores. Thus, the elution times were

retarded. Contrary to the case for beads modi®ed with cyclohexanol, temperature-dependent dextran

elution was opposite for beads modi®ed in the presence of toluene as porogen agent. The result indicates

that at higher temperature than LCST, faster elution was achieved than that at lower temperature due to

the reduction of pore size through the shrinkage of surface grafted PIPAAm chains. These results suggest

that the in situ surface selective modi®cation methods were achieved. It is indicated that separation mode

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1170

Fig. 3. Temperature dependence of the elution time of various molecular weight dextran on the PIPAAm-modi®ed porous silica

beads. Pore size of original silica beads is 237 AÊ. (Reprinted with permission from Ref. [18]. Copyright (1992) Wiley±VCH

GmbH.)

can be controlled through the modi®cation route even the modi®er is the same PIPAAm. Obtained beads

could be applied for a number of chromatographic applications.

Hosoya et al. [20] further investigated prepared PIPAAm-modi®ed particles as chromatographic

matrix for separation of several drug molecules in reversed-phase chromatography mode using aqueous

acetonitrile with different composition. They compared elution of several drug molecules with different

functionalities, possessing either hydrophobic or hydrophilic groups. They considered that as amide

groups in PIPAAm chains might be exposed to outward below the LCST, speci®c interaction should

be occurred with drug molecules with hydrophilic, polar functional groups (amine, amide, carbonyl,

hydroxyl, nitrile, and other functions). In fact, as shown in Fig. 4, when separation selectivity was

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1171

Fig. 4. Comparison of separation selectivity in 60% aqueous acetonitrile on base and PIPAAm-grafted packing materials.

Solute: 1, acetanilide; 2, acetophenone; 3, methyl benzoate; 4, cyanobenzene; 5, aniline; 6, nitrobenzene; 7, m-dinitrobenzene;

8, phenol; 9, p-dinitrobenzene; 10, bromobenzene (open circle, alkylbenzenes). (Reprinted with permission from Ref. [20].

Copyright (1995) American Chemical Society.)

compared, stronger interaction of drugs with more polar functional groups was observed at lower

temperature, at which PIPAAm exists in expanded conformation for packing matrix with PIPAAm

modi®ed on internal surfaces of pores. No selectivity was found on base packing materials. Furthermore,

they found the co-existing polypeptide, albumin, could be eluted out before exclusion volume for the

column packed with PIPAAm modi®ed on external bead surfaces and baseline separation of drugs is

achieved in reversed-phase chromatographic mode. Thus, with the surface-selective modi®cation with

PIPAAm, temperature-dependent separation of drug molecules is performed. These packing materials

could be selectively utilized for medical diagnosis; drug monitoring in the blood samples of the patients

without interference from serum proteins.

Go et al. [22] used silane coupler with methacryloylpropyl groups to introduce polymerizable groups

onto silica beads. Then, they carried out IPAAm polymerization to produce PIPAAm-grafted porous

silica beads for reversed-phase chromatography in GPC mode. Introducing PIPAAm onto silica beads

could shorten elution times, and show separation of dextrans with different molecular weight in aqueous

mobile phase. There is only a small difference in elution time change of dextran at lower temperature;

however, relatively larger elution time was observed above the polymer transition temperature. They

further analyzed lower molecular weight benzene derivatives with different polar functions in reversed-

phase chromatography using aqueous methanol solution. They showed the preferential retention of

hydrogen bond acceptors at low temperatures in low amount of methanol containing aqueous solution,

while hydrophobic interaction is dominant factor for solutes at higher temperature and in the mobile

phase with higher content of methanol. These results suggest that, at low temperature where PIPAAm

exists in expanded conformation, polar groups in the sample analyte exhibit relatively strong interaction

with amide side chains in PIPAAm and at higher temperature than PIPAAm LCST the hydrophobic

interaction regulates the solute retention.

The problem arose in their methods may be the use of organic solvent in the mobile phase preparation,

which may be disadvantageous from the environmental reason. Thus, separation of drug substances

should be carried out in the column system utilizing sole aqueous mobile phase.

Lakhiari et al. [21] introduced diethylaminoethyl (DEAE)-dextran onto silica beads through electro-

static interaction followed by cross-linking reaction of dextran hydroxyls with 1,4-butanediol diglyci-

dylether (BDGE). On this matrix surface, carboxyl terminated PIPAAm was further grafted using

BDGE. The PIPAAm-modi®ed bead packed column was characterized with high performance size

exclusion chromatography using aqueous mobile phase. Although the modi®ed column showed

temperature-responsive elution changes for lower molecular weight substances, proteins, and polysac-

charides, elution was almost same for both sample types with more than 10 000 molecular weights

regardless of temperature. For some proteins, elution time was retarded at higher temperature, probably

due to the hydrophobic nature of proteins those lead the hydrophobic interaction of proteins with

dehydrated, hydrophobized matrix surfaces. Although the base silica bead matrix has pore of 1000±

1250 AÊ, after the modi®cation with cross-linked DEAE-dextran and PIPAAm grafts the pore size and its

distribution might change dramatically. Therefore, surface characteristics of modi®ed beads were signif-

icantly altered as size exclusion chromatography matrix. Pore size control should be an important factor

to design thermoresponsive size exclusion chromatography matrix.

4. Temperature-responsive wettability changes of PIPAAm-modi®ed surfaces

As mentioned in Section 1, PIPAAm molecules in aqueous solution show temperature-responsive

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1172

soluble/insoluble changes due to the hydration/dehydration of side groups. Hydrophobic hydration

around isopropyl groups plays a role in this phenomenon. Above described reports are mainly focused

on PIPAAm's thermoresponsive characteristics; expanded coils at lower temperature than the LCST to

dehydrate globules above the LCST, to control matrix pore sizes. Other than such temperature-respon-

sive conformational changes to control pore sizes, the PIPAAm-modi®ed surfaces should show hydro-

philic/hydrophobic changes in response to temperature change in aqueous environment. In fact,

PIPAAm-grafted surfaces show temperature-dependent water contact angle changes.

Takei et al. [23] investigated the effect of graft chain conformations on the temperature-responsive

wettability changes of PIPAAm-grafted surfaces. They conjugated end-carboxyl PIPAAm and poly(N-

isopropylacrylamide-co-acrylic acid) (P(IPAAm-co-AAc)) to aminated glass cover slips with poly(styr-

ene-co-aminomethylstyrene) coatings independently to form free end linear PIPAAm-grafted surfaces

and multipoint attached PIPAAm surfaces, respectively, at 4 8C in water. They then investigated surface

wettability using Wilhelmy plate technique changing temperature. A large contact angle change was

obtained for PIPAAm-end grafted surfaces around 24 8C while smaller contact angle changes was

observed for P(IPAAm-co-AAc)-grafted surfaces with a wide temperature range. The transition

temperature was lower than PIPAAm's LCST for both surface types, probably due to the in¯uence of

base coating with polystyrene derivative as well as density of PIPAAm chains. Small contact angle

changes of the PIPAAm-multi-point attached surfaces might be due to the restricted chain conformation

of the grafted polymers.

The effects of the graft polymer chain conformation were further investigated on temperature-respon-

sive wettability changes of PIPAAm-grafted surfaces [24,25]. Four model surfaces were prepared and

these are schematically shown in Fig. 5; (a) free end linear PIPAAm-grafted surface, (b) multi-point

grafted looped PIPAAm surface, (c) free end PIPAAm grafted onto PIPAAm looped chain modi®ed

surface, and (d) PIPAAm thin hydrogel layer modi®ed surface. When PIPAAm was immobilized onto

the surfaces with multi-point fashion (type b) in Fig. 5) in good solvent for PIPAAm, PIPAAm

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1173

Fig. 5. Schematic representations of four types of PIPAAm-modi®ed surfaces with different molecular architecture.

molecules were attached with extended chain conformation, resulting in linear and small increase in

contact angle with temperature increase. On the contrary, immobilization medium contained poor

solvent to PIPAAm (two volume fractions of toluene to eight volume fractions of good solvent,

dioxane), PIPAAm chains were grafted onto the surface with looped chain con®guration with relatively

mobile nature. Thus, graft chain density of the looped PIPAAm should be increased than the previous

report [23]. For modi®cation of PIPAAm-grafted surfaces, amination of the base glass cover slips was

carried out. Considering from the surface density of amine functions and reaction with polymer mole-

cules, all prepared surfaces have higher PIPAAm density than the surfaces reported by Takei et al. [23].

The temperature-responsive contact angle changes measured using Wilhelmy plate technique are

indicated in Fig. 6. All the surfaces were ¯at enough to measure contact angles as con®rmed by

AFM mean roughness analysis. All surfaces showed temperature responsive contact angle changes,

low contact angles (high cos

u

A

) at lower temperature and higher contact angle at higher temperature.

Thus, the PIPAAm-grafted surfaces show temperature-responsive hydrophilic/hydrophobic changes.

With a multi-point graft chain conformation, transition temperature decreased slightly than PIPAAm

LCST in water. Thus, only a small restriction of the chain freedom was obtained. With increasing

polymer density on the surface by grafting free end PIPAAm on PIPAAm loops, large change in contact

angle was observed around polymer transition temperature. For PIPAAm hydrogel grafted surfaces,

however, smaller contact angle change was seen, probably due to the restricted chain mobility arising

from the cross-linking. Since the graft chain conformation has great in¯uence on the surface property

alteration in terms of wettability to water, the property modulation should affect on the interaction with

biomolecules and cells.

Morra and Cassinelli [26] utilized UV irradiation in the presence of benzophenone as photosensitizer

to polymerize and graft PIPAAm molecules onto the surfaces. They dissolved IPAAm monomer in the

presence of 0.1% benzophenone into 2-propanol at concentration ranging 5±40 wt%. UV light irradiation

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1174

Fig. 6. Temperature-dependent water contact angle changes for four types of PIPAAm-grafted surfaces with different molecular

architecture.

at 365 nm wavelength was carried out for 5±30 min. Surface analyses were done using ESCA with take-

off angle of 458. ESCA analysis revealed the existence of repeating units of PIPAAm on the modi®ed

surfaces. They measured surface wettability of the modi®ed surfaces. Volpe et al. [27] investigated the

contact angle changes of PIPAAm modi®ed surfaces using Wilhelmy plate measurements. They ®rstly

applied air plasma treatment of polystyrene sample surfaces followed by the polymerization of IPAAm

(10 wt%) in water at 25 8C in the presence of 0.2% ammonium cerium nitrate as initiator. ESCA

measurements revealed that after polymerization atomic ratio of the carbon, nitrogen and oxygen was

close to the theoretical ones, suggesting complete modi®cation with PIPAAm on the surfaces. They then

investigated the surface wettability toward water by means of Wilhelmy plate technique in detail,

changing immersion speed and temperature. Immersion speed was critical factor to change surface

wettability, though, it could be concluded that the PIPAAm-grafted surfaces show temperature-respon-

sive wettability changes; at 20 8C the surface is hydrophilic with low water contact angle, and at 37 8C

the surface become hydrophobic.

Liang et al. [28] prepared glass surface modi®ed with silane coupling agent with dithiocarbamate

groups and photopolymerization was carried out in the presence of IPAAm to form PIPAAm layer on the

glass surfaces, glass plate, and glass capillary tubing. These surfaces showed temperature responsive

wettability changes due to the hydration/dehydration changes of PIPAAm graft chains. Furthermore,

when PIPAAm was photopolymerized on the capillary tube surfaces, temperature responsive meniscus

height changes was observed; higher meniscus at lower temperature below the LCST and lower menis-

cus at higher temperature; 7 mm difference was seen for meniscus between 20 and 40 8C for capillary

with 2 mm diameter. Modi®cation was simple and easily controlled by UV irradiation times for poly-

merization. Therefore, the authors concluded the usefulness of the modi®cation methods to apply for

anti-fouling surface for biomolecules, and micro-channel for separation membranes. Their system

should be useful to modify UV transparent base materials, however, it would be dif®cult to modify

the sample with UV opaque materials thus, further modi®cation should be necessary. Pan et al. [29]

recently utilized plasma polymerization of IPAAm monomer gas in the radio-frequency plasma poly-

merization reactor. IPAAm monomer ¯ask was heated to 72 8C to evaporate monomer gas and intro-

duced into reactor vessel that contained base samples; Si wafers, and glass capillary tubing. Both FTIR

spectra and ESCA analysis of the modi®ed samples con®rmed successful modi®cation of samples with

PIPAAm. Thermoresponsive wettability change was observed by the meniscus height changes. A

difference of the meniscus is 0.7 cm between 5 and 55 8C.

5. Detachment control of cultured cells from PIPAAm-grafted culture dishes

Above-mentioned thermoresponsive wettability changes could be applied for cultured cell detach-

ment control. Takezawa et al. [30,31] utilized thermoresponsive soluble polymers for surface property

changes to recover cultured ®broblasts. They failed to culture cells on PIPAAm coating on tissue culture

polystyrene dishes because of dif®culty to get enough hydrophobicity at 37 8C. They actually mixed cell

adhesive protein, collagen, to achieve cell adhesion. As PIPAAm is solubilized by decrease in tempera-

ture, adhered cells are detached from the surface. However, detached cell suspension contained

solubilized PIPAAm and collagen as contaminants, those may interfere further use of detached cells.

Furthermore, detached cells form cell agglomerates and never succeeded to recover as a cell monolayer

sheet.

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1175

It is reported that the PIPAAm-grafted cell culture surfaces modi®ed by means of electron beam

irradiation to polymerize and graft PIPAAm on the surfaces covalently [32±34]. Recovery of cultured

cells was succeeded for various cell types from PIPAAm-grafted culture dishes by only lowering

temperature treatment without any use of harmful enzymes [33,35,36]. PIPAAm-grafted culture dishes

were utilized to construct novel tissue constructs in vitro [35±42]. Details are not discussed in this

manuscript and thus refer to above-mentioned references. Following these reports, several researchers

presented similar PIPAAm modi®cation of the surfaces to control and manipulate cultured cells

[26,27,43,44].

Thermoresponsive hydrophilic/hydrophobic changes of PIPAAm-modi®ed surfaces should also be

applied for novel hydrophobic chromatography as described in Section 6.

6. Novel `Green' chromatography utilizing PIPAAm-modi®ed surface

6.1. Effects of hydrophilic/hydrophobic property changes of PIPAAm surfaces on elution behavior

In Section 5, PIPAAm-modi®ed matrix was introduced to show thermoresponsive pore size control to

separate high molecular weight substances in GPC manner [18±20]. The authors group proposed that

thermoresponsive surface wettability changes of PIPAAm-grafted surfaces could be utilized to separate

hydrophobic bioactive compounds in aqueous mobile phase.

Kanazawa et al. [45] grafted end-carboxyl PIPAAm onto aminopropylsilica beads with 5 mm

diameter and used as packing materials for high performance liquid chromatography column. End

carboxyl PIPAAm was prepared according to the reports by Takei et al. [23,46] and utilizing radical

telomerization reaction in the presence of MPA as telogen. Initiator radicals attack thiol groups of MPA,

which initiated IPAAm polymerization. Synthetic route of the end carboxyl PIPAAm is shown in Fig. 7a.

Number average molecular weight of the PIPAAm was 6800 for surface modi®cation. End carboxyl

group of PIPAAm was then active esteri®ed with hydroxysuccinimide followed by reaction with

aminopropylsilica beads. Thus, PIPAAm-grafted silica beads with thermoresponsive property alteration

were obtained. Temperature-responsive surface property changes for PIPAAm-modi®ed beads are

shown in Fig. 7b. PIPAAm-modi®ed silica beads were packed into stainless steel column

(150 mm £ 4.6 mm F) by a slurry method. Column was then connected to HPLC system with a column

jacket connected to thermostated water bath with temperature control within 0.1 8C. Hydrophobic

steroids with different hydrophobicities were selected as the ®rst examples for separation. These steroids

were dissolved in Milli-Q water and injected into the PIPAAm-modi®ed column. As shown in Fig. 8, at

lower temperature than PIPAAm's LCST steroids cannot be separated nicely. With increasing tempera-

ture above the LCST of PIPAAm, retard retention times of the steroid molecules were observed. The

elution order was in accordance with the increasing order of the hydrophobicities of steroids as judged

from the log P values of the steroids where, P is the partition coef®cient of the sample material in 1-

octanol/water system [47]. Since no ef®cient separation was observed for unmodi®ed silica bead-packed

column in aqueous mobile phase, separation mechanism is totally different from the conventional

reversed-phase chromatography. At higher temperature above the PIPAAm's LCST, the surface of

the column packing materials becomes hydrophobic due to the dehydration of surface bound PIPAAm

molecules and precipitated on the interface, it would be plausible that the driving force of the steroid

interaction with the PIPAAm-modi®ed surfaces is the hydrophobic interaction. The discussion was

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1176

further con®rmed by evaluating the effect of hydrophobic moieties into the surface grafted PIPAAm

molecules through copolymerization. Kanazawa et al. [48] investigated the effect of surface hydropho-

bicity on the separation behavior of steroid molecules. Hydrophobic butyl methacrylate (BMA) was

introduced into the end-functionalized PIPAAm during the copolymerization with changing composi-

tion. The resultant copolymers showed lower LCSTs than PIPAAm homopolymer with increasing BMA

content in the copolymer. These modi®ers were then introduced to aminopropylsilica beads and modi-

®ed beads were packed into stainless steel column. With increasing hydrophobic BMA content in the

modi®er polymer, retention time of steroids with higher log P values retarded even at 5 8C. On silica

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1177

Fig. 7. (a) Synthesis of an end-carboxyl PIPAAm. (b) Schematic representation of PIPAAm-grafted silica beads (surface type

A).

beads modi®ed with P(IPPAm-co-BMA) of 3.2 mol% BMA content, sharp peaks of each steroid were

observed at 5 8C. With increasing temperature, longer retention times of steroids on each copolymer-

modi®ed surfaces were seen and baseline separation of steroids was achieved for BMA containing

copolymer-modi®ed surfaces at 30 8C where insuf®cient separation was observed for PIPAAm column.

As BMA is hydrophobic regardless of temperature, hydrophobic steroids have stronger interaction with

BMA moiety within the surface modi®ed copolymer, even though the copolymers were hydrated, and

expanded conformation at lower temperature than LCST. The thermoresponsive polymer-modi®ed

columns were then compared with conventional reversed-phase chromatography column, octadecylsi-

lane (ODS) column for steroid separation. Capacity factors of steroids on each column were evaluated

experimentally. The capacity factors for steroids on BMA-containing copolymer-modi®ed columns

were higher than those observed on PIPAAm-modi®ed column. As BMA content increases from 0.6

to 3.2 mol%, capacity factors for steroids dramatically increased at a given temperature; capacity factor

of testosterone changed from 2.45 for P(IPAAm-co-BMA) (BMA 0.6 mol%) to 7.81 for P(IPAAm-

co-BMA) (BMA 3.2 mol%) at 5 8C and increased from 18.5 to 33.8 at 50 8C, respectively. On the

contrary, when ODS column was used to elute hydrophobic steroids in water no elution of steroids

was achieved. Thus, the elution was performed with 50% methanol/water mixed solvent system, and

signi®cant decrease in capacity factors was apparent with temperature increase from 5 to 30 8C, indicat-

ing decreased interaction with temperature. This is sharp contrast with the case of column packed with

thermoresponsive polymer modi®ed silica.

With increasing hydrophobicity of the column modi®er, successful separation of hydrophobic

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1178

Fig. 8. Chromatograms of ®ve steroids and benzene at different column temperatures with water as a mobile phase from

PIPAAm-modi®ed column (surface type A). Peaks: 1, benzene; 2, hydrocortisone; 3, prednisolone; 4, dexamethasone; 5,

hydrocortisone acetate; 6, testosterone. (Reprinted with permission from Ref. [45]. Copyright (1996) American Chemical

Society.)

compounds in aqueous mobile phase was achieved. However, it also increases total elution times. As the

surface property alteration with temperature is reversible phenomenon, step temperature gradient might

effect to shorten analyzing times. Fig. 9 shows the effect of step temperature gradient on P(IPAAm-co-

BMA) (BMA 3.2 mol%). For steroids with lower partition coef®cients were insuf®ciently separated at

5 8C on P(IPAAm-co-BMA) (BMA 3.2 mol%) (Fig. 9a), while baseline separation was achieved at

30 8C with relatively longer analysis time (Fig. 9b). Thus, after the separation of ®rst three molecules

with lower hydrophobicities at 30 8C, column temperature was decreased to 5 8C by changing thermo-

stated waterbath set at 5 8C through three-way stopcocks. Retention times of steroids with higher

hydrophobicities were shortened with sharper peaks (Fig. 9c). This result indicated the effectiveness

of `thermoresponsive', and `thermoreversible' property alterations of PIPAAm-modi®ed surfaces,

in¯uencing hydrophobic interaction with sample molecules.

Proposed temperature-responsive liquid chromatography can be used with an aqueous mobile phase

without mixing any organic solvents, like acetonitrile and methanol in conventional reversed-phase

chromatography. For protein separation as well as puri®cation, the use of sole aqueous mobile

phase should be an advantage because denaturation of protein and following decrement of bioactivity

is likely caused by the use of organic solvent in the mobile phase. To explore the possibility for

separation of proteins in thermoresponsive chromatography using aqueous mobile phase, mixture

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1179

Fig. 9. Effect of step-temperature gradient on steroid elution from P(IPAAm-co-BMA) (BMA 3.2 mol%)-modi®ed column.

Isocratic elutions using water as mobile phase at (a) 5 8C and (b) 30 8C were indicated to compare with (c) step-temperature

gradient.

solution of three peptides were applied to P(IPAAm-co-BMA) (BMA 3.2 mol%) column varying

column temperature [48,49]. Used peptide samples were bovine insulin chain A, human b-endorphin

fragment (1±27aa), and bovine insulin chain B. Those samples have approximately same molecular

weight (around 3000) with different amount of hydrophobic amino acid residues. Fig. 10 shows the

elution pro®les of three peptides. At 5 8C, insuf®cient elution of three peptides was obtained. However,

as shown in this ®gure, baseline separation of three peptides were obtained with elution order of insulin

chain A , b-endorphin fragment , insulin chain B. When amino acid sequence of these peptides was

compared, the number of relatively hydrophobic amino acids (leucine, isoleucine, phenylalanine, tryp-

tophan, tyrosine and valine) was increased with the order of insulin chain A , b-endorphin fragment ,

insulin chain B, which directly corresponds to their respective retention times. Thus, it could be

concluded that at least peptides with molecular weight of approximately 3000 are separated through

the hydrophobic interaction with thermoresponsive polymer modi®ed stationary phase and eluted with

the order of their hydrophobicity.

Thermoresponsive P(IPAAm-co-BMA)-modi®ed column is also applicable to the separation and

analysis of phenylthiohydantoin-amino acids (PTH-amino acids), those are formed during amino acid

sequence analysis of the proteins and peptides by applying proteins to automated system for Edman-

based phenylisothiocyanate degradation [50]. Fig. 11 shows the elution pro®les of 18 PTH-amino acids

from P(IPAAm-co-BMA) (BMA 3.2 mol%) column. The elution times of most of hydrophobic PTH

amino acids were extended at 30 8C than those at 5 8C from P(IPAAm-co-BMA) (BMA 3.2 mol%)

column. Since lysine has two amino groups to be reacted with phenylisothiocyanate, PTH-lysine showed

slowest elution time among the PTH amino acids examined. Step temperature gradient is also revealed to

be effective to separate PTH-amino acids within a shorter time period.

Temperature-responsive polymer-modi®ed column was further utilized to analyze endocrine disrup-

ter, bisphenol A in aqueous mobile phase at 40 8C recently [51]. Yamamoto et al. detected bisphenol A

in the extracts from a coated drink can made from steel.

Above described results strongly indicated the advantage of thermoresponsive polymer-modi®ed

column for analysis of a wide range of samples in aqueous mobile phase by changing only temperature,

no further change in elution condition is needed such as mobile phase composition. Since organic

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1180

Fig. 10. Chromatograms for the mixture of three peptides on P(IPAAm-co-BMA) (BMA 3.2 mol%)-modi®ed column with

0.5 M NaCl solution (pH 2.1) at (a) 5 8C and (b) 30 8C. Peaks: 1, insulin chain A; 2, b-endorphin fragment (1-27aa); 3, insulin

chain B. (Reprinted with permission from Ref. [48]. Copyright (1997) American Chemical Society.)

solvent is not necessary in this chromatographic system, reducing cost for preparation and waste of the

mobile phase is advantageous. Furthermore, it would be desirable for environmental reasons.

6.2. Effect of surface morphology of PIPAAm modi®ed matrix on the separation

As described in Section 4, graft conformation of PIPAAm on the surface greatly in¯uences the

temperature-responsive wettability changes. Thus, it is possible to consider that the graft conformation

of the PIPAAm on the column matrix surfaces show different elution behavior of steroids. Thus,

temperature-dependent elution behavior of steroids was examined on PIPAAm looped chain grafted

surfaces, freely PIPAAm grafted onto PIPAAm loops, and PIPAAm thin hydrogel grafted surfaces ([25]

for hydrogel surface). Fig. 12 shows the chromatograms of steroids from three types of PIPAAm-

modi®ed surfaces at 5 and 45 8C, respectively. On each PIPAAm-modi®ed surface, low separation

ef®ciency is apparent from the chromatograms at 5 8C. With increasing temperature to 45 8C, elution

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1181

Fig. 11. Chromatogram for the mixture of 18 PTH-amino acids using water as a sole mobile phase at 30 8C. Peaks: 1, Ala; 2,

Asn; 3, Asp; 4, Cys; 5, Gln; 6, Glu; 7, Gly; 8, His; 9, Ile; 10, Leu; 11, Lys; 12, Met; 13, Phe; 14, Pro; 15, Thr; 16, Trp; 17, Tyr;

18, Val. (Reprinted with permission from Ref. [50]. Copyright (2000) American Chemical Society.)

of hydrophobic steroid retarded and longer retention times were observed. As surface grafted polymer

layers have higher graft chain density than freely mobile linear PIPAAm-grafted surfaces as discussed

earlier, analyte partitioning within PIPAAm layers should affect the extension of steroids retention times

and peak broadening. As PIPAAm with free end chains were grafted onto PIPAAm loop chains, thick-

ness of PIPAAm layer increased simultaneously with the grafting reaction. Thus, longer retention was

seen on the surface with free end PIPAAm grafted onto PIPAAm loops than the surface with PIPAAm

loops alone. Signi®cant peak broadening was seen in the case of PIPAAm thin hydrogel grafted surface.

As PIPAAm hydrogel on the matrix surfaces have three-dimensional cross-linked structure, partitioning

of analyte molecules should be signi®cant. Furthermore, the mobility of the PIPAAm chains are

restricted due to the cross-linked structure, in¯uencing peak broadening as well as the retardation of

elution times of steroids. These considerations were supported by the elution pro®les of steroids after

applying step temperature gradient. Step temperature gradient from 45 to 15 8C was applied during the

elution of two steroids, cortisone acetate and testosterone. Fig. 13 shows the change in retention time

with step temperature gradient. Horizontal axis in this ®gure represents time at which column tempera-

ture was changed to 15 8C and time 0 (zero) indicates the sample injection time. Upper and lower lines of

the shaded area in the ®gure represents the isocratic elution time at 45 and 15 8C, respectively, for each

tested samples. Earlier elution of the steroids was observed than that obtained by isocratic elution at

45 8C. For each column, backpressure of the column became equilibrated values within 5 min regardless

of the surface con®guration of PIPAAm graft chains. In the case of PIPAAm loop-grafted surfaces,

retention did not return to the equilibrium retention time at 15 8C with isocratic elution even though the

column temperature was changed before sample injection. This may be due to the restricted graft chain

mobility. After that retention times changed linearly, indicating relatively rapid surface property

alterations because the slow surface property change might produce the retarded retention and retention

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1182

Fig. 12. Chromatograms of steroids from PIPAAm-modi®ed surfaces at 5 and 45 8C using water as mobile phase. (a) PIPAAm

loop grafted, (b) free PIPAAm grafted onto PIPAAm loops, (c) PIPAAm thin hydrogel grafted surfaces. Peaks: 1, cortisone; 2,

prednisolone; 3, dexamethasone; 4, cortisone acetate; 5, testosterone.

curve changes convexly. Thus, mobile phase temperature would change rapidly in this chromatographic

condition, which results in the alteration of surface properties of PIPAAm-grafted matrix. Two column

types; freely PIPAAm grafted onto PIPAAm loops, and PIPAAm thin gel layer grafted matrix showed

relatively large changes with temperature. This may be due to the thickness of the grafted polymer layer

to partitioning analyte samples. Although the present results need longer analysis time, system could be

improved to shorten analysis time with good separation factors. Molecular design of the matrix interface

with thermoresponsive polymer modi®er is promising approach to modulate interactions of and separate

analyte samples.

After Kanazawa's reports, Teal et al. [52] reported the temperature mediated hydrophobic modulation

chromatography for puri®cation of biomolecules. They utilized PIPAAm microgels instead of solid

phase materials modi®ed with PIPAAm as in the reports by Kanazawa et al. They used the prepared

PIPAAm microgel packed column to separate several sets of amino acids, proteins, and nucleotides.

Although they showed changes in elution of these biomolecules, they mixed PIPAAm microgels with

commercially available Sephadex bead matrix. This is probably because the avoidance of thermoresponsive

void volume changes inside the column since the PIPAAm gels are known to show signi®cant volume

transition with temperature changes [7,53,54]. Kanazawa et al. [45,48±50] and Yakushiji et al. [25] used

solid silica beads as base material for PIPAAm modi®cation. Furthermore, introduced PIPAAm molecules

produced signi®cantly thinner surface layers on the silica beads, minimum void volume changes is apparent

as was judged by the no change in marker molecule elution time with temperature. It is concluded that

the molecular architecture of PIPAAm as the column matrix should be one of important points for designing

aqueous temperature responsive chromatography columns for separation of diverse biomolecules.

6.3. Effect of surface charged groups on analyte separation

In Sections 6.1 and 6.2, mainly separation of hydrophobic substances was discussed using thermally

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1183

Fig. 13. Effects of step-temperature gradient on steroid elution from thermoresponsive columns with three different molecular

architectures using water as mobile phase. Sample was injected at time 0 (zero) in the ®gure and step-temperature gradient was

applied at given temperature indicated in horizontal axis. Tes., testosterone; C.A., cortisone acetate.

responsive and reversible solid phase surfaces with a sole aqueous mobile phase. In this section, recent

approach is described to separate charged biomolecules using charged, thermoresponsive matrix.

A wide variety of biomolecules possess both charge, and hydrophobic nature in the molecules. In

general, biomolecules are separated with reversed-phase chromatography, ion exchange chromatogra-

phy systems and these combinations. Although these procedures are currently utilized in wide ®elds, use

of organic solvents may limit further application of separated molecules due to the denaturation,

especially proteins. Thus, if electrostatic and hydrophobic interactions were modulated simultaneously

with temperature in aqueous mobile phase, the separation system would have potential application in

separation science. Described approach is to introduce weakly charged functional groups, carboxyl or

amine groups, into thermoresponsive PIPAAm derivatives onto base silica beads matrix.

Cross-linked poly(IPAAm-co-N,N-dimethylaminopropylacrylamide (DMAPAA)) with 20 mol%

DMAPAA was grafted onto silica beads to obtain cationic thermoresponsive column matrix [55].

This matrix could potentially be applied to ion exchange column in which eluent pH is changed in a

wide range. Silica base materials, however, are labile at higher pH like pH 10. Thus, the polymer

modi®ed beads as well as the base aminopropylsilica beads were immersed into alkaline solution at

pH 10 for 21 days and checked diameter change by means of scanning electron microscope (SEM) and

surface chemical composition by ESCA. Bare silica beads showed decrease in bead diameter from 5.2 to

4.2 mm, while negligible change was observed on polymer-modi®ed beads without change in chemical

composition at the surface as determined by ESCA measurement even at pH 10. Therefore, polymer

coating on the beads improved the surface stability for at least 21 days at pH 10. In the practical use, pH

of the medium is mostly neutral in the following experiments; there would be negligible size as well as

chemical composition changes during chromatographic analysis. The cationic matrix was used as

column matrix and separation of anionic substances was carried out at pH 7. With increasing tempera-

ture, elution time of anionic analytes was shortened. IPAAm sequence in the copolymer matrix dehy-

drated and surfaces became hydrophobic. Under this condition, micro environmental polarity around

amino functions decreased, which induce deprotonation of amino groups, leading to weak electrostatic

interaction with samples. Thus, retention time was shortened with temperature. Feil et al. [56] reported

the effects of pH on the swelling behavior of pH-/temperature-responsive hydrogels with temperature

changes. Swelling transition pH for pH-/temperature-responsive hydrogels was found to change with

temperature. The results strongly supported the above-mentioned discussion.

Anionic temperature-responsive polymer modi®ed column was also prepared to examine the modula-

tion of electrostatic interaction between catecholamines [57]. For this purpose, we prepared copolymer

of IPAAm and acrylic acid (AAc) with various chemical compositions. As AAc content in the copoly-

mer increased, LCST also was increased. Suf®cient soluble/insoluble phase transition was observed

below the 3 mol% AAc containing polymers. Catecholamine derivatives were used to examine

P(IPAAm-co-AAc)-modi®ed column as thermoresponsive chromatography. At lower pH where AAc

exists in protonated carboxyl, low retention with poor separation of samples was observed. At pH 7,

however, stronger interaction occurred and longer retention was observed for positively charged

samples. Electrostatic interaction is the primary force for retention of catecholamines. Since hydro-

phobic steroids have stronger interaction and increased retention time was observed with temperature at

pH 5 on P(IPAAm-co-AAc) column, sample hydrophobicity also affected on the catecholamine retention

because catecholamine also has hydrophobic benzene ring. Hence both electrostatic and hydrophobic

interactions simultaneously in¯uence partitioning of basic catecholamine derivatives. These two interaction

forces compete with each other, those interaction forces modulate retention of cationic substances.

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1184

Charged groups should be hydrophilic in aqueous solution, introduction of these groups increases the

transition temperature of matrix PIPAAm condition as described earlier. Co-incorporation of hydro-

phobic monomer, butyl methacrylate (BMA), suppresses increase in polymer transition temperature,

thus, ternary copolymer, cross-linked P(IPAAm-co-DMAPAA-co-BMA) was prepared on the silica

bead surfaces, and the modi®ed bead matrix was used for separation of adenosine nucleotides [58].

Temperature-dependent amine pK

a

shift in the copolymer was examined. Amine pK

a

decreased with

increasing temperature, especially around polymer transition temperature a drastic decrease in pK

a

was

observed. Furthermore, surface potential of the beads indicated that the amine-containing thermo-

responsive polymer-modi®ed beads showed lightly positive surface potentials below the polymer transi-

tion temperature than PIPAAm-modi®ed beads, due to the incorporation of cationic amine sites. Above

the transition temperature, surface potential decreased to negative value. These results indicated that

deprotonation of the protonated amino groups below the transition temperature occurred by only

increasing temperature. Urry et al. [59,60] reported the pK

a

shift of amino acid carboxyl groups in

hydrophobic polypeptides with temperature. They claimed that the hydrophobicity and micro environ-

mental polarity changes strongly in¯uenced to the protonation/deprotonation changes of carboxyl

groups in polypeptides. Their results strongly supported the results of the temperature-responsive charge

density control of P(IPAAm-co-DMAPAA-co-BMA)-modi®ed surfaces. Ternary polymer-modi®ed

silica beads were then used in the separation of adenosine nucleotides. Adenine nucleotides; AMP,

ADP, and ATP, play a role in energy metabolism in cellular organism, and are important with respect to

the bioanalytical and biochemical research. A mixture of adenosine nucleotides was injected into the

thermostated P(IPAAm-co-DMAPAA-co-BMA) column and monitored elution pro®les at different

temperatures. With increasing amount of phosphate group in the molecular structure, increased retention

time was observed regardless of temperature. Poor resolution of three nucleotides is observed on the

PIPAAm-modi®ed column. Thus, the electrostatic interaction dominates between positively charged

matrix surfaces and negatively charged samples. Increase in temperature resulted in the shorter retention

times and narrower peaks. The van't Hoff plots indicated the discontinuous changes in capacity factors at

temperature corresponding to the polymer transition temperature. In this case, the step temperature

gradient is effective to modulate elution of sample analytes similar to the gradient elution in

reversed-phase chromatography by altering solvent composition. Although, the system still needs longer

time for complete analysis, column condition could be optimized through chemical composition of

copolymer modi®er, size of column and silica bead diameter, and so on. Adenosine nucleotides are

currently analyzed with reversed-phase column in the absence of [61,62] or in the presence of additive

alkylamine (ion-pair RPC) [63±65] to modulate sample property and interaction with stationary phase

surfaces and thus elution times. The use of intelligent materials should prove valuable in the design of

novel aqueous `green' chromatography systems.

7. Selective adsorption/desorption of bioactive proteins

Selective protein separation and puri®cation are usually carried out using af®nity columns [66]. On the

af®nity columns, biomolecules with speci®c binding ability to target molecules are immobilized. Most

commonly used biomolecules conjugated onto the af®nity matrix are antibody. In the af®nity separation,

recovery of target proteins is commonly achieved through changing eluent composition, pH and/or ionic

strengths, or rather, urea to actively denature adsorbed proteins altering binding strength to antibody.

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1185

However, the recovery procedure frequently causes irreversible denaturation of recovered proteins,

which limits further application in biomedical purposes. Thus, af®nity separation with simple adsorp-

tion/desorption procedure is needed.

Galaev et al. [67,68] utilized temperature-responsive polymer, poly(N-vinylcaprolactam) (PVCL) as

eluent modi®er for elution of lactate dehydrogenase (LDH) bound to dye-af®nity chromatographic

column, Blue Sepharose. As above noted, to recover proteins adsorbed on the dye-conjugated column

matrixes elution with high salt concentration in buffer is frequently utilized to weaken protein±dye

interactions. However, in their system, PVCL was added to the eluent to recover proteins on the column

matrix without changing salt concentration in the eluent solution [67]. They synthesized Blue Sepharose

from Sepharose conjugated with Cibacron Blue. Then, LDH was applied onto the Blue Sepharose

packed columns. Polymer elution was carried out with 1% PVCL solution. They found that 1%

PVCL solution is a more ef®cient non-speci®c eluting agent than 1.5 M KCl solution. PVCL contains

amide polar group which interacts strongly with matrix-conjugated Cibacron Blue, therefore, PVCL

could be displaced column bound LDH. Since PVCL is a thermoresponsive polymer with a LCST

around 35±40 8C, PVCL in a polymer-enriched phase could be separated with raising solution tempera-

ture after the LDH elution with PVCL. Increasing temperature resulted in precipitation of PVCL, thus

with a low speed centrifugation precipitated polymer was separated. PVCL was completely recovered

from aqueous phase during PVCL precipitation with temperature treatment. Since the temperature was

not exceeded above 40 8C, minimum denaturation of separated enzymes was con®rmed. Recovered,

precipitated PVCL can be re-dissolved by lower temperature treatment and re-used for further chroma-

tographic runs. Galaev et al. [68] further examined by modifying Cibacron Blue-conjugated column

matrix with PVCL to shield dye molecules. Column shielding was performed with 1% PVCL solution

(ca. 50 column volumes) followed by column washing with 1.5 M KCl and re-equilibration with

appropriate buffer solution before chromatographic runs. Fig. 14 shows the comparison of LDH elution

from unmodi®ed column and PVCL-shielded column at 23 and 40 8C. As shown in Fig. 14a, LDH

elution was observed only after the addition of 1.5 M KCl, which is required for elution of dye-bound

enzymes regardless of temperature. By strong contrast, PVCL shielding greatly affected LDH elution

from the column (Fig. 14b). At 40 8C, only a slight elution was observed, and majority of the enzymes

bound to Cibacron Blue dye on column matrixes is eluted out by the addition of 1.5 M KCl. At 23 8C,

however, elution is successful with 0.1 M KCl. This is a sharp contrast from the unmodi®ed column.

Thus, they further investigated the effect of temperature change during LDH elution with 0.1 M KCl as

eluent (Fig. 15). By raising temperature virtually no elution was obtained at 40 and 45 8C (above the

LCST of the PVCL). Only cooling column temperature is the necessary and simple method for elution of

column-bound enzymes. PVCL bound to the Cibacron Blue molecules on the column matrix surfaces

with multivalent manner. At lower temperature where PVCL is hydrated, loose coils, af®nity of the

PVCL to Cibacron Blue compete with LDH and slight decrease of LDH binding ef®ciency with 0.1 M

KCl elution is suf®cient to elute LDH from the column. By contrast, PVCL dehydrates, shrunk at higher

temperature above the LCST, which cause exposure of more amount of LDH toward Cibacron Blue on

the matrix surfaces. Therefore, stronger interaction occurs between LDH and Cibacron Blue, which lead

to the necessity of higher salt concentration of KCl for LDH elution. Galaev et al. then utilized PVCL-

shielded Blue Sepharose column for puri®cation of LDH from crude porcine muscle extracts. Crude

extract was applied to the PVCL-shielded column at 40 8C until breakthrough. Then the foreign proteins

in the crude extract were eluted out with 0.1 M KCl at 40 8C. After that the column temperature was

lowered to 23 8C, resulting in successful elution of LDH. During this procedure, the recovery of LDH

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1186

was 90%. In conclusion, shielding of Cibacron Blue dye molecules in the Blue Sepharose column with

PVCL is the simple, ef®cient method to modulate LDH elution with temperature changes.

In their method, there is a possibility of co-elution of surface bound PVCL molecules during the

protein elution. Galaev et al. concluded that, as PVCL is temperature-responsive polymer, raising

temperature of the ef¯uent followed by low speed centrifuge is suf®cient to separate co-eluted PVCL.

This might be tedious procedure if the method extends to commercial use.

PIPAAm's soluble/insoluble changes have been recently utilized to control af®nity of target molecule

by modulating the surrounding temperature [69]. In this system, adsorbed biomolecules could be

completely recovered by lowering temperature through `kicking-out' effect of samples with expanding

PIPAAm chains (Fig. 16). To do this, Cibacron Blue is immobilized onto amino functionalized poly-

methacrylate bead matrixes with spacers having different spacer sizes. Cibacron Blue is known to have

higher af®nity to serum albumin molecules and used in dye-af®nity chromatography [70,71]. The matrix

surfaces are co-immobilized with end-carboxyl PIPAAm. End-carboxyl PIPAAm was synthesized

through radical telomerization using MPA as chain transfer agents in N,N-dimethylformamide (DMF)

[46]. The PIPAAm chain lengths are controlled by the molar ratio of thiol compounds to IPAAm

monomer. In our study, the molecular weight of PIPAAm was 1900 with molecular weight distribution

index of 3.8 as determined by terminal carboxyl quanti®cation by titration and GPC measurement.

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1187

Fig. 14. Elution pro®le of LDH at 23 8C (open circle) and 40 8C (closed plot) from (a) unmodi®ed and (b) PVCL-shielded Blue

Sepharose column with 0.1 M KCl, followed by elution with 1.5 M KCl. Arrows in the ®gure indicate when elution with 0.1 or

1.5 M KCl was begun. (Reprinted from Yu, et al. Temperature-induced displacement of proteins from dye-af®nity columns

using an immobilized polymeric displacer. J Chromatogr A 1994;684:37±43 with permission from Elsevier Science.)

Polymethacrylate beads with epoxy side chains are aminated with 1,6-hexamethylenediamine for further

conjugation of spacer molecules and active esteri®ed PIPAAm molecules. Spacers used were 1,3-

butadiene diepoxide and ethylene glycol diglycidylether. Those spacer molecules were used to change

the distance between bead surfaces and af®nity molecule, Cibacron Blue. Assuming that the surface-

grafted PIPAAm is fully expanded below the LCST, its length is comparable or slightly longer than the

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1188

Fig. 15. Elution pro®les of LDH from PVCL-shielded Blue Sepharose with 0.1 M KCl at different temperatures. Arrows in the

®gures indicated when elution was interrupted, the column was cooled to 23 8C and elution was continued at this temperature.

(Reprinted from Yu, et al. Temperature-induced displacement of proteins from dye-af®nity columns using an immobilized

polymeric displacer. J Chromatogr A 1994;684:37±43 with permission from Elsevier Science.)

Cibacron Blue bound with 1,3-butadiene diepoxide. However, in case of ethyleneglycol diglycidylether

as the spacer, the length of Cibacron Blue is much longer than the fully expanded PIPAAm molecules.

Thus, by changing temperature, the number of af®nity molecules on the bead surfaces could be regu-

lated. In fact, no albumin adsorption onto PIPAAm-grafted surfaces was evident, however, temperature

dependent adsorption was obvious for Cibacron Blue co-immobilized beads surfaces. At higher

temperature than PIPAAm's LCST, albumin adsorption is large, and the amount decreased with lower-

ing temperature below the LCST. Adsorbed albumin at higher temperature was easily desorbed with low

temperature treatment, where PIPAAm molecules hydrate and expand to outward. Expansion of

PIPAAm molecules induced albumin conjugated to Cibacron Blue to be pushed out. Almost all albumin

molecules on the Cibacron Blue immobilized with 1,3-butadiene diepoxide spacers were desorbed with

only decreasing temperature. No modi®cation to eluent pH, or ionic strength is applied. When ethylene

glycol diglycidylether was used as spacer molecule for Cibacron Blue immobilization, the limited

number of albumin adsorbed on the bead surfaces can be desorbed with lowering temperature. This

is probably due to the insuf®cient expansion of PIPAAm molecules at lower temperature since the

expanded chain length of hydrated PIPAAm is calculated to be shorter than Cibacron Blue with ethylene

glycol diglycidylether spacers. Thus, the size of spacer molecules and PIPAAm chain length are both

dominant factors for controlled albumin adsorption/desorption behavior.

The human sera were then used to separate human serum albumin from the sera (Fig. 17). At higher

temperature (37 8C), only one peak was appeared. Using SDS-PAGE, this peak contains two different

molecular weight substances other than albumin. After 3 ml elution, column temperature was decreased

to 20 8C and a single, major peak was obtained. In this peak only one molecular weight protein was

found in SDS-PAGE, which corresponds to the human albumin. This result strongly supports the new

thermoresponsive chromatography with `selective catch and release' mechanism other than hydrophobic

interaction or pore size control with thermoresponsive PIPAAm grafts.

Under this procedure, no contamination of PIPAAm was achieved because PIPAAm was covalently

bound on the matrix surfaces and unbound molecules were extensively washed out during matrix

preparation. Thus, this method is feasible to extend a large number of proteins.

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1189

Fig. 16. Schematic representation of the concept for selective adsorption/desorption control with thermoresponsive PIPAAm

co-grafted with af®nity dye, Cibacron Blue.

8. Conclusion

In this review, surface modi®cation with stimuli-responsive `intelligent materials' is introduced. Modi®ed

surfaces and interfaces were utilized to control permeation, diffusion of substances to be separated and/or

released. Special attention was made on the thermo-responsive polymer, PIPAAm, as a surface modi®er, and

on PIPAAm-modi®ed surfaces for possible use in chromatographic matrix with sole aqueous mobile

phase. Successful modulation of interaction between PIPAAm-modi®ed surfaces and samples was

apparent with alteration of surface property through temperature control in aqueous solution. This

feature should be feasible to develop `green' chromatography systems with sole aqueous mobile phase.

Acknowledgements

Part of the present work on thermoresponsive aqueous chromatography was carried out by collabora-

tion with Professors Hideko Kanazawa and Yoshiazu Matsushima of Kyoritsu College of Pharmacy, Drs

Yukio Hasegawa, Yasuro Shinohara, Kimihiro Yoshizako, and Yoshikatsu Akiyama of Amersham

Biosciences Co., and Dr Taiji Yakushiji, and Professor Kiyotaka Sakai of Waseda University. Financial

support was provided from the Japan Chemical Innovation Institute (JCII) under the New Energy and

Industrial Technology Development Organization (NEDO) from ®scal year of 1996±2000 to which the

authors would express sincere acknowledgement.

References

[1] Okano T, Yoshida R. Intelligent polymeric materials for drug delivery. In: Tsuruta T, Hayashi T, Kataoka K, Ishihara K,

Kimura Y, editors. Biomedical applications of polymeric materials. Boca Raton, FL: CRC Press, 1993. p. 407±28.

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1190

Fig. 17. Effect of step temperature gradient on selective `catch and release' of target molecule, human serum albumin, for

PIPAAm-co-grafted Cibacron Blue af®nity column.

[2] Okano T. Molecular design of temperature-responsive polymers as intelligent materials. In: Dusek K, editor. Advances in

polymer science, responsive gels; volume transition II. Berlin: Springer, 1993. p. 179±97.

[3] Okano T, Yui N, Yokoyama M, Yoshida R. Advances in polymeric systems for drug delivery. Yverdon, Switzerland:

Gordon & Breach, 1994.

[4] Okano T, editor. Biorelated polymers and gels: controlled release and applications in biomedical engineering. Chestnut

Hill, MA: Academic Press, 1998.

[5] Hoffman AS. Intelligent polymers in medicine and biotechnology. Macromol Symp 1995;98:645±64.

[6] Heskins M, Guillet JE, James E. Solution properties of poly(N-isopropylacrylamide). J Macromol Sci, Chem A

1968;2:1441±5.

[7] Bae YH, Okano T, Kim SW. Temperature dependence of swelling of crosslinked poly(N,N-alkyl substituted acrylamides)

in water. J Polym Sci: Part B, Polym Phys 1990;28:923±36.

[8] Kanazawa H, Matsushima Y, Okano T. Temperature-responsive chromatography. In: Brown PR, Grushka E, editors.

Advances in chromatography. New York: Marcel Dekker, 2001. p. 311±36.

[9] Iwata H, Matsuda T. Preparation and properties of novel environment-sensitive membranes prepared by graft polymer-

ization onto a porous membrane. J Membr Sci 1988;38:185±99.

[10] Iwata H, Hirata I, Matsuda Y. Atomic force microscopic images of solvated polymer brushes. Langmuir

1997;13:3063±6.

[11] Ito Y, Casolaro M, Chung DJ, Imanishi Y. Design and synthesis of glucose-sensitive insulin-releasing systems. Drug

Deliv Syst 1988;3(3):391±7.

[12] Ito Y, Casolaro M, Kono K, Imanishi Y. An insulin-releasing system that is responsive to glucose. J Control Release

1989;10:195±203.

[13] Okahata Y, Noguchi H, Seki T. Thermoselective permeation from a polymer-grafted capsule membrane. Macromolecules

1986;19:493±4.

[14] Iwata H, Oodate M, Uyama Y, Amemiya H, Ikada Y. Preparation of temperature-sensitive membranes by graft poly-

merization onto a porous membrane. J Membr Sci 1991;55:119±30.

[15] Fong RB, Ding Z, Long CJ, Hoffman AS, Stayton PS. Thermoprecipitation of streptavidin via oligonucleotide-mediated

self-assembly with poly(N-isopropylacrylamide). Bioconj Chem 1999;10:720±5.

[16] Stayton PS, Shimoboji T, Long C, Chilkoti A, Chen G, Harris JM, Hoffman AS. Control of protein-ligand recognition

using a stimuli-responsive polymer. Nature 1995;378:472±4.

[17] Chilkoti A, Chen G, Stayton PS, Hoffman AS. Site-speci®c conjugation of a temperature-sensitive polymer to a geneti-

cally-engineered protein. Bioconj Chem 1994;5:504±7.

[18] Gewehr M, Nakamura K, Ise N, Kitano H. Gel permeation chromatography using porous glass beads modi®ed with

temperature-responsive polymers. Makromol Chem 1992;193:249±56.

[19] Hosoya K, Sawada E, Kimata K, Araki T, Tanaka N, Frechet JMJ. In situ surface-selective modi®cation of uniform size

macroporous polymer particles with temperature-responsive poly-N-isopropylacrylamide. Macromolecules

1994;27:3973±6.

[20] Hosoya K, Kimata K, Araki T, Tanaka N, Frechet JMJ. Temperature-controlled high-performance liquid chromatography

using a uniformly sized temperature-responsive polymer-based packing material. Anal Chem 1995;67:1907±11.

[21] Lakhiari H, Okano T, Nurdin N, Luthi C, Descouts P, Muller D, Jozefonviz J. Temperature-responsive size-exclusion

chromatography using poly(N-isopropylacrylamide) grafted silica. Biochim Biophys Acta 1998;1379:303±13.

[22] Go H, Sudo Y, Hosoya K, Ikegami T, Tanaka N. Effects of mobile-phase composition and temperature on the selectivity

of poly(N-isopropylacrylamide)-bonded silica gel in reversed-phase liquid chromatography. Anal Chem 1998;70:4086±

93.

[23] Takei YG, Aoki T, Sanui K, Ogata N, Sakurai Y, Okano T. Dynamic contact angle measurement of temperature-

responsive surface properties for poly(N-isopropylacrylamide) grafted surfaces. Macromolecules 1994;27(21):6163±6.

[24] Yakushiji T, Sakai K, Kikuchi A, Aoyagi T, Sakurai Y, Okano T. Graft architectural effects on thermo-responsive

wettability changes of poly(N-isopropylacrylamide)-modi®ed surfaces. Langmuir 1998;14(16):4657±62.

[25] Yakushiji T, Sakai K, Kikuchi A, Aoyagi T, Sakurai Y, Okano T. Effects of cross-linked structure on temperature-

responsive hydrophobic interaction of PIPAAm hydrogel modi®ed surfaces with steroids. Anal Chem 1999;71(6):1125±

30.

[26] Morra M, Cassinelli C. Thermal recovery of cells cultured on poly(N-isopropylacrylamide) surface-grafted polystyrene

dishes. Polym Prepr 1995;36(1):55±6.

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1191

[27] Volpe CD, Cassinelli C, Morra M. Wilhelmy plate measurements on poly(N-isopropylacrylamide)-grafted surfaces.

Langmuir 1998;14(16):4650±6.

[28] Liang L, Feng X, Liu J, Rieke PC, Fryxell GE. Reversible surface properties of glass plate and capillary tube grafted by

photopolymerization of N-isopropylacrylamide. Macromolecules 1998;31:7845±50.

[29] Pan YV, Wesley RA, Luginbuhl R, Denton DD, Ratner BD. Plasma polymerized N-isopropylacrylamide: synthesis and

characterization of a smart thermally responsive coating. Biomacromolecules 2001;2(1):32±6.

[30] Takezawa T, Yamazaki M, Mori Y, Yonaha T, Yoshizato K. Morphological and immuno-cytochemical characterization

of a heterospheroid composed of ®broblasts and hepatocytes. J Cell Sci 1992;101:495±503.

[31] Takezawa T, Mori Y, Yoshizato K. Cell culture on a thermo-responsive polymer surface. Biotechnology

1990;8:854±6.

[32] Yamada N, Okano T, Sakai H, Karikusa F, Sawasaki Y, Sakurai Y. Thermo-responsive polymeric surfaces; control of

attachment and detachment of cultured cells. Makromol Chem, Rapid Commun 1990;11:571±6.

[33] Okano T, Yamada N, Sakai H, Sakurai Y. A novel recovery system for cultured cells using plasma-treated polystyrene

dishes grafted with poly(N-isopropylacrylamide). J Biomed Mater Res 1993;27:1243±51.

[34] Okano T, Yamada N, Okuhara M, Sakai H, Sakurai Y. Mechanism of cell detachment from temperature-modulated,

hydrophilic±hydrophobic polymer surfaces. Biomaterials 1995;16:297±303.

[35] von Recum H, Kikuchi A, Okuhara M, Sakurai Y, Okano T, Kim SW. Retinal pigmented epithelium cultures on thermally

responsive polymer porous substrates. J Biomater Sci, Polym Ed 1998;9(11):1241±53.

[36] Kushida A, Yamato M, Konno C, Kikuchi A, Sakurai Y, Okano T. Temperature-responsive culture dishes allow non-

enzymatic harvest of differentiated Madin±Darby canine kidney (MDCK) cell sheets. J Biomed Mater Res 2000;51:216±

23.

[37] Kikuchi A, Okuhara M, Karikusa F, Sakurai Y, Okano T. Two-dimensional manipulation of con¯uently cultured vascular

endothelial cells using temperature-responsive poly(N-isopropylacrylamide)-grafted surfaces. J Biomater Sci, Polym Ed

1998;9(12):1331±48.

[38] Kushida A, Yamato M, Konno C, Kikuchi A, Sakurai Y, Okano T. Decrease in culture temperature releases monolayer

endothelial cell sheets together with deposited ®bronectin matrix from temperature-responsive culture surfaces. J Biomed

Mater Res 1999;45:355±62.

[39] Hirose M, Kwon OH, Yamato M, Kikuchi A, Okano T. Creation of designed shape cell sheets that are noninvasively

harvested and moved onto another surface. Biomacromolecules 2000;1(3):377±81.

[40] Kwon OH, Kikuchi A, Yamato M, Okano T. Rapid cell sheet detachment from poly(N-isopropylacrylamide)-grafted

porous cell culture membranes. J Biomed Mater Res 2000;50:82±9.

[41] Yamato M, Utsumi M, Kushida A, Konno C, Kikuchi A, Okano T. Thermo-responsive culture dishes allow the intact

harvest of multilayered keratinocyte sheets without dispose by reducing temperature. Tissue Engng 2001;7(4):473±80.

[42] Nandkumar MA, Yamato M, Kushida A, Konno C, Hirose M, Kikuchi A, Okano T. Two-dimensional cell sheet

manipulation of heterotypically co-cultured lung cells utilizing temperature-responsive culture dishes results in long-

term maintenance of differentiated epithelial cell functions. Biomaterials 2002;23(4):1121±30.

[43] Rollason G, Davis JE, Sefton MV. Preliminary report on cell culture on a thermally reversible copolymer. Biomaterials

1993;14(2):153±5.

[44] Ito Y, Chen G, Guan Y, Imanishi Y. Patterned immobilization of thermoresponsive polymer. Langmuir 1997;13:2756±9.

[45] Kanazawa H, Yamamoto K, Matsushima Y, Takai N, Kikuchi A, Sakurai Y, Okano T. Temperature-responsive chroma-

tography using poly(N-isopropylacrylamide)-modi®ed silica. Anal Chem 1996;68(1):100±5.

[46] Takei YG, Aoki T, Sanui K, Ogata N, Okano T, Sakurai Y. Temperature-responsive bioconjugates. 1. Synthesis of

temperature-responsive oligomers with reactive end groups and their coupling to biomolecules. Bioconj Chem

1993;4:42±46.

[47] Hansch C, Leo A, Hoekman D. Exploring QSARÐhydrophobic, electrostatic, and steric constants. Washington, DC:

American Chemical Society, 1995.

[48] Kanazawa H, Kashiwase Y, Yamamoto K, Matsushima Y, Kikuchi A, Sakurai Y, Okano T. Temperature-responsive

liquid chromatography. 2. Effect of hydrophobic groups in N-isopropylacrylamide copolymer-modi®ed silica. Anal Chem

1997;69(5):823±30.

[49] Kanazawa H, Yamamoto K, Kashiwase Y, Matsushima Y, Takai N, Kikuchi A, Sakurai Y, Okano T. Analysis of peptides

and proteins by temperature-responsive chromatographic system using N-isopropylacrylamide polymer-modi®ed

columns. J Pharm Biomed Anal 1997;15:1545±50.

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1192

[50] Kanazawa H, Sunamoto T, Matsushima Y, Kikuchi A, Okano T. Temperature-responsive chromatographic separation of

amino acid phenylthiohydantoins using aqueous media as the mobile phase. Anal Chem 2000;72(24):5961±6.

[51] Yamamoto K, Kanazawa H, Matsushima Y, Oikawa K, Kikuchi A, Okano T. Temperature-responsive chromatographic

separation of bisphenol A with water as a sole mobile phase. Environ Sci 2000;7(1):47±56.

[52] Teal HR, Hu Z, Root DD. Native puri®cation of biomolecules with temperature-mediated hydrophobic modulation liquid

chromatography. Anal Biochem 2000;283:159±65.

[53] Sato-Matsuo E, Tanaka T. Kinetics of discontinuous volume-phase transition of gels. J Chem Phys 1988;89(3):1695±703.

[54] Yoshida R, Uchida K, Kaneko Y, Sakai K, Kikuchi A, Sakurai Y, Okano T. Comb-type grafted hydrogels with rapid

temperature responses. Nature 1995;374(6519):240±2.

[55] Kobayashi J, Nakayama S, Kikuchi A, Kanazawa H, Matsushima Y, Sakai K, Okano T. Separation of acidic bioactive

compounds by aqueous chromatography using thermo-responsive cationic surfaces. Therapeu Engng 2000;12(1):670±6

(in Japanese).

[56] Feil H, Bae YH, Feijen J, Kim SW. Mutual in¯uence of pH and temperature on the swelling of ionizable and thermo-

sensitive hydrogels. Macromolecules 1992;25:5528±30.

[57] Kobayashi J, Sakai K, Kikuchi A, Okano T. Aqueous chromatography utilizing pH-/temperature-responsive polymer

stationary phases to separate ionic bioactive compounds. Anal Chem 2001;73(9):2027±33.

[58] Kikuchi A, Iwasa T, Kobayashi J, Sakai K, Okano T. Submitted for publication.

[59] Urry DW. Free energy transduction in polypeptides and proteins based on inverse temperature transitions. Prog Biophys

Mol Biol 1992;57:23±57.

[60] Urry DW, Gowda DC, Peng S, Parker TM, Jing N, Harris RD. Nanometric design of extraordinary hydrophobic-induced

pK

a

shifts for aspartic acid: relevance to protein mechanisms. Biopolymers 1994;34:889±96.

[61] Leoncini G, Buzzi E, Maresca M, Mazzei M, Balbi A. Alkaline extraction and reverse-phase high-performance liquid

chromatography of adenine and pyridine nucleotides in human platelets. Anal Biochem 1987;165:379±83.

[62] Teerlink T, Hennekes M, Bussemaker J, Groeneveld J. Simultaneous determination of creatine compounds and adenine

nucleotides in myocardial tissue by high-performance liquid chromatography. Anal Biochem 1993;214:278±83.

[63] Tekkanat KK, Fox IH. Isocratic separation of ATP and its degradation products from biological ¯uids by automated liquid

chromatography. Clin Chem 1988;34(5):925±32.

[64] Schobert B. Enzymatic synthesis of ATP analogs and their puri®cation by reversed-phase high-performance liquid

chromatography. Anal Biochem 1995;226:288±92.

[65] Blanco J, Canela EI, Sayos J, Mallol H, Lluis C, Franco R. Adenine nucleotides and adenosine metabolism in pig kidney

proximal tubule membranes. J Cell Physiol 1993;157:77±83.

[66] Matejtschuk P. Af®nity chromatogaphyÐa practical approach. Oxford: IRL Press, 1997.

[67] Galaev IY, Mattiasson B. Effect of synthetic polymers, poly(N-vinyl pyrrolidone) and poly(N-vinyl caprolactam), on

elution of lactate dehydrogenase bound to blue Sepharose. J Chromatogr 1993;648:367±71.

[68] Galaev IY, Warrol C, Mattiasson B. Temperature-induced displacement of proteins from dye-af®nity columns using an

immobilized polymeric displacer. J Chromatogr 1994;684:37±43.

[69] Yoshizako K, Akiyama Y, Yamanaka H, Shinohara Y, Hasegawa Y, Carredano E, Kikuchi A, et al. Submitted for

publication.

[70] Allary M, Saint-Blancard J, Boschetti E, Girot P. Bioseparations 1991;2:167.

[71] Matejtschuk P, Feldman PA, Rott J, More J. Af®nity separations of proteins. In: Matejtschuk P, editor. Af®nity separa-

tionsÐa practical approach. Oxford: IRL Press, 1997. p. 61±97.

A. Kikuchi, T. Okano / Prog. Polym. Sci. 27 (2002) 1165±1193

1193