11111111

ELSEVIER Fluid Phase Equilibria 110 (1995) 115- 128

Measurement and correlation of partition coefficients of hydrolytic enzymes for dextran+poly (ethylene glycol) +water aqueous twophase systems at 20C
T. Furuya, Y. Iwai, Y. Tanaka, H. Uchida, S. Yamada and Y. Arai

Department of Chemical Engineering, Kyushu University 6-10-1 Hakozaki, Higashi-ku, Fukuoka 812 (Japan)
Keywords: experimental, liquid-liquid, proteins, dextran, poly(ethylene glycol), water, m-amylase, [~-amylase, glucoamylase, partition coefficient, modified Flory-Huggins equation

ABSTRACT The partition coefficients of hydrolytic enzymes, x-amylase, 13-amylase and glucoamylase for the dextran (DEX)+poly(ethylene glycol) (PEG)+water aqueous two-phase systems with various polymer molecular weights were measured at 20C. The partition coefficients obtained were correlated by using a modified Flory-Huggins equation which was empirically proposed for aqueous systems. The interaction parameters required were determined by fitting the model to the experimental data. The partition coefficients could be correlated with good agreement. INTRODUCTION There is growing interest in efficient methods for the large-scale recovery and purification of bioproducts with the increasing importance of modem biotechnology. Aqueous two-phase systems provide a successful method for separating mixtures of biomolecules (Albertsson, 1986; Walter et al., 1985). For design of such a separation system, the liquid-liquid equilibria (LLE) of aqueous two-phase systems and the partition coefficient (Kp) are needed as fundamental knowledge. In recent years, LLE and Kp for proteins have been reported (Diamond and Hsu, 1989; Diamond and Hsu, 1990; Forciniti et al., 1991; Forciniti et al., 1992; King et al., 1988). However, the partition coefficient data are not enough to elucidate the partitioning mechanism of biomolecules. In this study, therefore, partition coefficients of three hydrolytic enzymes, m-amylase, 13-amylase and glucoamylase, in the dextran+poly(ethylene glycol)+water aqueous two-phase systems with various molecular weights were
0378-3812/95/$~.50 O 1995- Elsevier Science B.V.All rights reserved S S D I 0378-3812(95)02746-7

116

T. Furuya et al., / Fluid Phase Equilibria 110 (1995) 115 - 128

measured at 20C. A successful model for correlating the LLE of aqueous two-phase system and K of biomolecules is needed to design the separation process. The UNIQUAC equation was used by Kang et al. (1989), Kang and Sandler (1987) and Hartounian et al. (1993), and the osmotic viriai equation was adopted by Gaube et al. (1993) and King et ai. (1988) to correlate the LLE of aqueous two-phase system. For the partition coefficient, the osmotic virial equation (King et al. 1988) and the Flory-Huggins equation (Diamond and Hsu, 1989; Diamond and Hsu, 1990; Waiter et al., 1985) were used. In this study, a modified Flory-Huggins equation developed for aqueous systems (Furuya et al., 1994) was adopted to correlate the partition coefficients obtained.

EXPERIMENTAL
Ma~ria~

Poly(ethylene glycol) (PEG) samples with two different molecular weights were purchased from Wako Pure Chemical Ind. as PEG4000 (lot.No.LKP0209) and PEG6000 (lot.No.PTL1876). Dextran (DEX) T40 (lot. No.NH04575), DEX T70 (lot.No.PC08625) and DEX T500 ((a) lot.No.PA08143, (b) lot.No.QC11346) were purchased from Pharmacia Inc. DEX 140 (lot.No.DSL2657) was purchased from Wako Pure Chemical Ind. The number- and weight-average molecular weights of DEX and PEG are shown in Table 1. The number- and weight-average molecular weights of PEGs and DEX 140 were determined by using size exclusion chromatography (SEC). Those of DEX T40, T70, T500(a) and T500(b) were reported from manufacturer. ~amylase (lot.No.23100) was purchased from Seikagaku Corp., ~-amylase (lot.No.59F 8100) was purchased from Sigma Chemical Company and glucoamylase (lot.No. 125060) was purchased from Funakoshi K.K., respectively. The molecular weight, isoelectric point (pI) and specific volume of enzymes are shown in Table 2. All reagents were used without further purification. Water was purified with a Millipore Milli-Q system.
Me~o~

In four equilibrium glass cells, polymer solutions with the same concentration were prepared by weighing from the stock solutions of PEG and DEX. For three equilibrium glass cells, lg enzyme stock solution (lmg/g) was added. One without enzyme was served as blank. The total concentration of enzyme was about 0.2mg/g. The pH was adjusted as 7 by adding phosphate buffer (10mM/L). The pH was measured in each phase by means of microelectrode. The pH difference between the two phases was negligible. The solutions thus prepared were shaken for 12 hours and settled for 2 days. After the phase separation had taken place, each phase was withdrawn by using a syringe. Protein (enzyme) concentrations were measured by using Bradford test (Bradford, 1976). All measurements were performed at 20.0C+0.1 C.

T. Furuya et al., / Fluid Phase Equilibria 110 (1995) 115 - 128

117

TABLE 1 Molecular weights of polymers Polymer DEX T40 DEX T70 DEX 140 DEX T500(a) DEX T500(b) PEG4000 PEG6000
Mn Mw Mw/Mn

18800 37900 35600 191000 170300 2900 8500

38000 66700 139800 476000 503000 3100 8800

106

2.02 1.76 3.93 2.49 2.95 1.07 1.04

TABLE 2 Properties of enzymes a Enzyme s-amylase p-amylase glucoamylase Origin

Bacillus subtiles Sweet Potato Rhizopus niveus

Molecular weight 50000 50000 70000

pI 5.4 4.8 8.5

Specific volume (cm3/g) 0.72 0.74 0.75

a Isoelectric point and specific volume data are cited from Seikagaku Data Book 1 (Nippon Seikagakukai ed., 1979).

Results

The relation between partition coefficients of enzymes and tie-line length for the DEX T40+PEG4000+water, DEX T70+PEG4000+water, DEX 140+PEG4000+water, DEX T500(a)+PEG4000+water and DEX T500(b)+PEG6000+water aqueous twophase systems are shown in Tables 3 to 7, respectively. The LLE (tie-line) of the same aqueous two-phase systems have been reported (Furuya et al., 1994). The partition coefficient, K , and tie-line length are defined as follows, respectively. The concentration of protein in PEG-rich top phase [mg / g] = The concentration of protein in DEX-rich bottom phase [mg / g]
(. TOP _ IA~BOT'~2 ' " DEX! (hi/TOP _ blIBOT. 2 / 1 / 2 --PEG-' / ~," " PEG

Kp

(1)
(2)

Tie-Line Length =~,,WDE x

where w denotes the weight fraction. The accuracy of the experimental partition coefficient data is considered to be within +10% from the reproducibility. As the partition coefficients of enzymes are smaller than unity, these three hydrolytic enzymes are partitioned to DEX-rich bottom phase. The partition coefficient of a-

118
TABLE 3

T. Furuya et al., /Fluid Phase Equilibria 110 (1995) 115- 128

Partition coefficients of hydrolytic enzymes in DEX T40+PEG4000+water aqueous two-phase system at 20"C a

Tie-line length (wt.%) 13.33 18.16 21.99 24.84

re
ct -amylase 0.28 0.17 0.12 0.10 1~ -amylase 0.35 0.17 0.10 0.07 glucoamylase 0.44 0.29 0.20 0.15

a LLE of this aqueous two-phase system are as same as Table 3 in the previous paper (Furuya et al., 1994).

TABLE 4 Partition coefficients of hydrolytic enzymes in DEX T70+PEG4000+water aqueous two-phase system at 20"C a

Tie-line length (wt.%) 14.97 18.88 22.28 25.25

Kp
a -amylase 0.29 0.19 0.15 0.11 13 -amylase 0.38 0.23 0.15 0.11 glucoamylase 0.43 0.33 0.20 0.16

a LLE of this aqueous two-phase system are as same as Table 4 in the previous paper (Furuya et al., 1994).

TABLE 5 Partition coefficients of hydrolytic enzymes in DEX 140+PEG4000+water aqueous two-phase system at 20"C a

Kp
Tie-line length (wt.%) 13.57 17.60 20.37 22.68 a-amylase 0.32 0.23 0.18 0.14 13 -amylase 0.60 0.42 0.26 0.20 glucoamylase 0.55 0.39 0.36 0.26

a LEE of this aqueous two-phase system are as same as Table 5 in the previous paper (Furuya et al., 1994).

T. Furuya et ui., / Fluid Phase Equilibria 110 (1995) 115 - 128

119

TABLE 6 Partition coefficients of hydrolytic enzymes in DEX T500(a)+PEG4000+water aqueous two-phase system at 20"C a

rp
Tie-line length (wt.%) s-amylase B-amylase glucoamylase 11.94 0.48 0.97 0.69 13.33 0.39 0.92 0.63 17.00 0.26 0.73 0.50 21.99 0.18 0.37 0.31 a LLE of this aqueous two-phase system are as same as Table 6 in the previous paper (Furuya et al., 1994).

TABLE 7 Partition coefficients of hydrolytic enzymes in DEX T500(b)+PEG6000+water aqueous two-phase system at 20C a

rp
Tie-line length (wt.%) 12.17 14.55 16.68 18.93 o~-amylase 0.44 0.34 0.27 0.20 B-amylase 0.64 0.44 0.31 0.18 glucoamylase 0.54 0.39 0.30 0.23

a LLE of this aqueous two-phase system are as same as Table 7 in the previous paper (Furuya et al., 1994). amylase is smaller than those of other two enzymes. With increasing the tie-line length, the partition coefficients of enzymes are decreased. The partition coefficients of txamylase and B-amylase in DEX+PEG4000+water aqueous two-phase systems are shown in Figs. 1 and 2. In the aqueous two-phase system containing higher molecular weight polymer, the phase compositions are different. Therefore, phase compositions are different even though the aqueous two-phase systems are equal in tie-line length. As shown in Fig. 1, the partition coefficients of tx-amylase are almost independent of DEX molecular weight. On the other hand, as shown in Fig.2, the partition coefficients of 13-amylase are increased (partitioned to PEG-rich top phase) with increasing DEX molecular weight. The partition coefficients of glucoamylase are also increased with increasing DEX molecular weight. The magnitude of the dependency of K on DEX molecular weight for glucoamylase is moderate between tx-amylase and 13P-amylase. The partition coefficients of t~-amylase and 13-amylase in DEX T500(a)+PEG4000 and DEX T500(b)+PEG6000+water aqueous two-phase systems are shown in Fig.3. The partition coefficients of tx-amylase are almost independent of PEG molecular weight though those of ~-amylase are decreased with increasing PEG molecular

120

T. Furuya et al., / Fluid Phase Equilibria 110 (1995) 115 - 128

1.0

0.8
.T- o.6
v

Q.

0.4
0.2
0,0 I I I

10

15 20 25 30 Tie-Line Length (wt.%)

Fig. I. Partition coefficients of (x-amylase in Dextran+PEG4000+water aqueous two-phase systems at 20C: ( O ) DEX T40; ( / ~ ) DEX T70; ( [] ) DEX 140; ( ~ ) DEX T500(a); ( ) smoothed lines.

1.0 0.8 Ivy0-6

13.

'~ 0.4 0.2 0.0 10

I I I

15

20

25

30

Tie-Line Length (wt.%)

Fig.2. Partition coefficients of 13-amylase in Dextran+PEG4000+water aqueous two-phase systems at 20C: ( O ) DEX T40; ( /k ) DEX T70; ( [] ) DEX 140; ( ~ )DEX T500(a); ( ) smoothed lines.

T. Furuya et al., / Fluid Phase Equilibria 110 (1995) 115

128

121

weight. Those of glucoamylase are also decreased. However, the magnitude for glucoamylase is smaller than that for 13-amylase. CORRELATION As shown in Figs. 1 to 3, the values of K occasionally depend on the molecular weight of polymers. We should explain and correlate quantitatively the behavior of K P to design a separation equipment. The tendency of the molecular weight dependency of K for l - a m y l a s e shown in Figs. 2 and 3 can be qualitatively explained by the conventional Flory-Huggins equation (Kang and Sandier, 1987); In K~ = (1 - m4[ m l ) ~ l / 1 - ( 1 - m 4 ] m2)~l/2+ other terms assuming that the DEX concentration is negligible in the PEG-rich phase and the PEG concentration is negligible in the DEX-rich phase. However, to correlate quantitatively, we need a suitable model for aqueous polymer solutions.

Modified Flory-HugginsEquation
When aqueous two-phases are in equilibrium in each other, the chemical potential differences (A~ti _- ~ti - ~t~ur~) of component i in both phases are identical.
(A~J,i) TOP [ . =[a~t~)~BOT ,

i=l,2,-..n

(3)

1.0 0.8 -T-0.6 ~" 0.4 0.2

0.0 I I

10

15 20 25 Tie-Line Length (wt.%)

Fig.3. Partition coefficients of o~-amylaseand 13-amylasein DEX T500(a)+PEG4000+water and DEX T500(b)+PEG6000+water aqueous two-phase systems at 20C: ((~)) o~-amylase for PEG4000 system; (/x ) [~-amylasefor PEG4000 system; ( ) s-amylase for PEG6000 system; ( A ) 13-amylasefor PEG6000 system; ( ) smoothed lines.

122

T. Furuya et al., / Fluid Phase Equilibria 110 (1995) 115 - 128

Simultaneous solution of eqn.(3) for all components gives equilibrium compositions for aqueous two-phase system, when a suitable chemical potential equation can be obtained. The expression for partition coefficient is also derived from eqn.(3). The UNIQUAC equation (Hartounian et al., 1993; Kang and Sandier, 1987; Kang et al., 1989) has been adopted to calculate the liquid-liquid equilibria for the aqueous twophase system. The osmotic virial equation (Gaube et al., 1993; King et al., 1988) and the Flory-Huggins equation (Diamond and Hsu, 1989; Diamond and Hsu, 1990; Kang and Sandier, 1987; Walter et al., 1985) have been adopted to calculate the liquid-liquid equilibria and the partition coefficient of biomolecules. However, the Flory-Huggins equation is not necessarily suitable for the systems which have specific interactions such as hydrogen bonding (King et al., 1988; Yu et al., 1993). Therefore, a modified Flory-Huggins equation, which takes into account of deviations from random mixing caused by specific interaction such as hydrogen bonding, has been proposed (Furuya et al., 1994). In this study, the modified Flory-Huggins equation was used for the partition coefficient. The modified Flory-Huggins model for the chemical potential differences of component i is shown as follows (Furuya et al., 1994).

m~J-i ~j . ~ (atY (Xji -l)RT = In I1/i+ 1 - mi~ ~ + mi~ )~ij ijl]l i j Illj ] ""3
-- m i ~j ~j ~jt, lllTJklll:kJ(O~j,+ (Xtj- l)

(4)

where ~i and m i denote the volume fraction of component i and the ratio of the molar volume of component i to a reference volume, R is the gas constant and T is the temperature, respectively. Here, reference volume was taken to be that of component 3, water (thus m3=l). Xij is interaction parameter between components i andj. ~j is the exponent that is empirically introduced to take into account of deviations from random mixing caused by specific interaction such as hydrogen bonding. When c~j =1 and ~j~ =1, eqn.(4) becomes the conventional Flory-Huggins equation. To derive the expression for partition coefficient, eqn.(4) is adopted to DEX(1)+ PEG(2)+water(3)+enzyme(4) four components system. Because the enzyme concentration is very dilute, the volume fraction of enzyme, ~/4 ' was approximated to be zero (infinite dilution). To simplify the equation, (~4i and Cf,i4 w e r e set to unity. The partition coefficient is derived as follows.

(v:P /
In K' r : In / VnB-----~J

=_m4{~41(~./1 TOP_~ll BOT) "4-~42(~I~2TOP. I#/2 )"1"Z43[~1/3 O P . V3BOT~ "]" m , C I BOTh.. [. T -)f

(5)

where

T. Furuya et al., /Fluid Phase Equilibria 110 (1995) 115- 128

123

C=
/

.~1TO...o+ + ~I/2TOP--~ BOT + .o...od/ --" Iq/1 -11/3 - --~lJ3

ms m2 m3 ) SOT 0c12 BOT ct21~

,~ f[,,tTOP'~Ctl2I',IttTOP'~OC21

"l- (~'13 "l- {~'31--I )XI3{('~IJ'ITOP)OtI3(~TOP)Gt31- (~I~pOT)~I3('~J'P j' OT)Ot3t}

(6)

T TO '4" (1ff,,23"st"~'32--1)~,Z3"~, (~1/2o P) + "('1413 P)

+ -- ('t]J'2T ) + "(~/3.OT )

/1
The values of Zt2, )~13' X23 and ~o have been reported in a previous paper (Furuya et al., 1994). Therefore, the composition of DEX(1), PEG(2) and water(3) in top and bottom phases can be calculated by using eqns.(3) and (4). By using the relation ~3 = 1 - (~t~ + ~2), the expression for the partition coefficient can be finally derived as follows.

lng'p-m4C

( xl/~P - ~t/~OX/
(7)

m4 (xl/2 TP- V2 Bx) = (~43-)~41) ~~2---~_ ~2----~/ at- ()~43- X42)

From eqn.(7), the parameters (X43- X4~)and (X43 -- ~42) can be determined as the slope and the intercept, respectively, based on the straight line relationship between the left hand side and (~t TP - ~l BT) / (~2TP - ~2BOT). The partition coefficient obtained in this experiment is defined by weight fraction as given by eqn.(1). Therefore, the partition coefficients expressed in weight fraction were converted to the volume fraction-based P as follows.

K"

I]~rnl(W~T/Ml)+m2(w~TIM2)+m3(w~ v//14_+)1 K'p=~40T= ~,rnl(WrlOPi )~+m3(w~Or/M3)IK p Ml

~t~P

(8)

Results
To obtain the partition coefficients of enzymes expressed in volume fraction (K/ p)' using eqn.(8), the specific volumes of DEX, PEG, water and enzymes are required to calculate the molar volume ratio (m). In this study, 0.626cm3/g and 0.832cma/g were used for the specific volumes of DEX and PEG, respectively (Kang and Sandier, 1987). The specific volume of water was calculated from the density at 20*C. Therefore, in this

124

T. Furuya et al., / Fluid Phase Equilibria 110 (1995) 115 - 128

calculation, m 2 (PEG4000) = 142.87, m 2 (PEG6000) = 405.57, m3=l (reference),

M2=Mw(PEG) and M3= 18.02. However, DEXs used here show wide molecular weight
distributions. As shown in the previous paper (Furuya et al., 1994), DEXs are approximated by the F-distribution function and the continuous thermodynamic approach was adopted. Therefore, ~. ml(wl / MO in eqn.(8) are given by summarized values of mt(w t ! Mr) of pseudo-components (in this case 8 components). The values of ~,~ and tx0 and the calculation procedure for LLE of aqueous two-phase system have been reported in the previous paper (Furuya et al., 1994). The volume fractions of polymers in top and bottom phases can be given by the LLE calculation. The parameter C can be calculated by using eqn.(6). The plots of eqn.(7) for DEX T500(a)+PEG4000+water aqueous two-phase system, as an example, is shown in Fig.4. In the present study, each volume fraction ~i in top and bottom phases and parameter C were first given from LLE calculation based on the previous paper (Furuya et al., 1994). Then, (~1 TOP -- I~IBOT) / (~I/2TOP -- 1~2BOT can be easily ) calculated and the tie-line length can also be obtained. The experimental K ~exp) corresponding to this tie-line length was found by interpolating the experimental relationship between K and the tie-line length. And the K ~exp)was converted to K" texp~ P P by using eqn.(8). The value of (In K "t~xp~- m4C) / m4(~2T~ _ ~ 2 BOT) is plotted against p ( ~ ToP_ tV BOT) / (~2TOP _ ~2BOr) in Fig.4. From this figure, the straight line relationships are given and therefore the values of (~43 -- ~41 ) and (Z43 - Z42) can be determined as the slope and the intercept, respectively, for each enzymes. These parameters determined are presented in Table 8 as the parameter set I. A typical calculated results with the parameter set I are also shown in Fig.5(a). The parameter set II for (~43 -- ~42) are redetermined to give the same value because (~43 -- )~42) should be identical for the same PEG4000 system. Further, the parameter set II for (~43 -- X41 ) are given for the DEX T500 systems. As shown in Table 8, the parameter set II for ()~43 -- ~4I ) are slightly decreased with increasing DEX molecular weight. As an example, the calculated partition coefficient in DEX T500(a)+ PEG4000+water aqueous two-phase system used parameter set II are shown in Fig.5(b). The calculated partition coefficients with the parameter set I show a good agreement and the parameter set II are useful to predict qualitatively the partition coefficients.

CONCLUSIONS The partition coefficient data of or-amylase, [3-amylase and glucoamylase in the dextran+poly(ethylene glycol)+water aqueous two-phase systems with various polymer molecular weights were reported at 20C. These enzymes were partitioned to DEX-rich bottom phase selectively and the partition coefficients were different in each other. The partition coefficient of ~t-amylase was slightly changed with increasing polymer molecular weight. Those of [3-amylase and glucoamylase increase with increasing DEX molecular weight and decrease with increasing PEG molecular weight. The partition coefficients were correlated by using the modified Flory-Huggins equation previously proposed. The interaction parameters were determined by fitting

T. Furuya et al., / Fluid Phase Equilibria 110 (1995) 115 - 128

125

-0.21

. . . .

5
I I
e~

-0.23

.....iiii.. ...
/ i | i i [ | i i i

~~,~ .0.25 -0.27

-1.60
1

-1.55
-. BOT TOP

-1.50

Fig.4. Plot of eqn.(7) for DEX T500(a)+PEG4000+water aqueous two-phase system: t~-amylase : ( ( ] ) ) exp., ( . . . . . ) calc.; 13-amylase : ( /x ) exp., ( ..............) calc.; glucoamylase : ( [ ] ) exp., ( ) calc..

TABLE 8 Interaction parameters for partition coefficient of enzymes(4) in the DEX(1)+PEG(2)+water(3) aqueous two-phase systems at 20C DEX T40( 1)+ PEG4000(2) ){43-~41 ]~43-~42 0.478 0.486 0.449 0.442 0.438 0.426 0.398 0.365 0.474 0.487 0.432 0.423 DEX T500(a)(1)+ PEG4000(2) ~43-~41 or-amylase (I) (II) ~-amylase (I) (II) (I) glucoamylase (II) 0.414 0.444 0.380 0.389 0.416 0.428 )~43-~42 0.396 0.442 0.351 0.365 0.405 0.423 DEX T70( 1)+ PEG4000(2) ~43-~41 ~43-~42 0.455 0.454 0.447 0.442 0.439 0.432 0.395 0.365 0.429 0.420 0.431 0.423 DEX T500(b)(1)+ PEG6000(2) ~43-~41 0.375 0.444 0.320 0.389 0.379 0.428 ~43-~42 0.367 0.479 0.279 0.393 0.378 0.457 DEX 140( 1)+ PEG4000(2) )~43-~41 0.466 0.446 0.412 0.393 0.447 0.429 ~43-~42 0.472 0.442 0.393 0.365 0.449 0.423

s-amylase

(I) (II) ~-amylase (I) (II) glucoamylase (I) (II)

126

T. Furuya et al., /Fluid Phase Equilibria 110 (1995) 115- 128

1.0 0.8 I~ 0.6 0.4 0.2 0.0

0

A~..

'~% %-.% "-.

.............
"o.. ..........
I I

15

20

25

Tie-Line Length (wt.%)

Fig.5(a). Calculated partition coefficients with parameter set I in DEX T500(a)+PEG4000 +water aqueous two-phase system at 20C: t~-amylase : ( O ) exp., ( . . . . . ) calc.; p-amylase : ( / x ) exp., ( .............. ) calc.; glucoamylase : ( [ ] ) exp., ( ~ ) calc..

1.0 0.8 "-'0.6 I

v
Q. 0

..~ %.... ",.~

[]

"-.. "-...
[]

..................

0.4 0.2
0.0 I I

10

15 20 Tie-Line Length (wt.%)

25

Fig.5(b). Calculated partition coefficients with parameter set II in DEX T500(a)+PEG4000 +water aqueous two-phase system at 20C: t-amylase : ( O ) exp., ( . . . . . ) calc.; 13_amylase : ( / x ) exp., ( .............. ) calc.; glucoamylase : ( [ ] ) exp., ( ) calc..

T. Furuya et al., / Fluid Phase Equilibria 110 (1995) 115 - 128

127

the model to the experimental data and the parameter sets I and II are presented. The correlated results with the parameter set I show a good agreement with experimental data and the parameter set II is useful to predict qualitatively the tendency of partition coefficients.

LIST OF SYMBOLS C K M Mn Mw m pI R T
w
P

parameter in eqn.(5) partition coefficient of enzyme defined by weight fraction (= w4VP/ w4Bx) partition coefficient of enzyme defined by volume fraction (= ~4Tv / ~4ar) molecular weight number-average molecular weight weight-average molecular weight molar volume ratio of polymer to water isoelectric point gas constant (J mo1-1 K-~) temperature (K) weight fraction

Greek letters
t)

~t

exponent in eqn.(4) Flory-Huggins interaction parameter chemical potential volume fraction

Superscripts

TOP BOT pure

top phase (PEG rich phase) bottom phase (DEX rich phase) pure component

Subscripts
1

2 3 4 i,j

dextran (DEX) poly(ethylene glycol) (PEG) water enzyme components i and j

128

T. Furuya et al., / Fluid Phase Equilibria 110 (1995) 115 - 128

ACKNOWLEDGMENTS We gratefully acknowledge the financial support provided by the Grant-in-Aid for Encouragement of Young Scientists (1993, 05750676), The Ministry of Education, Science and Culture, Japan. We also thank Mr. K. Nakada, and Ms. J. Zhu for their support in this experiment.

REFERENCES Albertsson, P.A., 1986. Partition of Cell Particles and Macromolecules.(3rd), John Wiley & Sons,New York. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal.Biochem., 72: 248-254. Diamond, A.D. and Hsu, J.T., 1989. Fundamental studies of biomolecule partitioning in aqueous twophase systems. Biotechnol.Bioeng., 34: 1000-1014. Diamond, A.D. and Hsu, J.T., 1990. Protein partitioning in PEG/Dextran aqueous two-phase systems. AIChE J., 36: 1017-1024. Forciniti, D., Hall, C.K. and Kula, M.-R., 1991. Protein partitioning at the isoelectric point: influence of polymer molecular weight and concentration and protein size. Biotechnol.Bioeng., 38: 986-994. Forciniti, D., Hall, D.K. and Kula, M.-R., 1992. Electrostatic effects on protein partitioning: simultaneous effect ofpH and polymer molecular weight. Chem. Eng. Sci., 47: 165-175. Furuya, T., Iwai, Y., Tanaka, Y., Uchida, H., Yamada, S. and Arai, Y., 1994. Measurement and correlation of liquid-liquid equilibria for dextran-poly(ethylene glycol)-water aqueous two-phase systems at 20C. Fluid Phase Equilibria, in press. Gaube, J., Pfennig, A. and Stumpf, M., 1993. Thermodynamics of aqueous poly(ethylene glycol)dextran two-phase systems using the consistent osmotic virial equation. Fluid Phase Equilibria, 83: 365-373. Hartounian, H., Floeter, E., Kaler, E.W. and Sandier, S.I., 1993. Effect of temperature on the phase equilibrium of aqueous two-phase polymer systems. AIChE J., 39: 1976-1984. Kang, C.-H., Lee, C.-K. and Sandier, S.I., 1989. Polydispersivity effects on the behavior of aqueous two-phase two-polymer systems. Ind.Eng.Chem.Res., 28:1537-1542. Kang, C.-H. and Sandier, S.I., 1987. Phase behavior of aqueous two-polymer systems. Fluid Phase Equilibria, 38: 245-272. King, R.S., Blanch, H.W. and Prausnitz, J.M., 1988. Molecular thermodynamics of aqueous twophase systems for bioseparations. AIChE J., 34: 1585-1594. Nippon Seikagakukai ed., 1979. Seikagaku Data Book 1. Tokyo Kagaku Dojin,Tokyo, p.92-135. (in Japanese) Walter, H., Brooks, D.E. and Fisher, D., 1985. Partitioning in Aqueous Two-Phase Systems Theory,Methods, Uses, and Application to Biotechnology -. Academic Press, Inc.,New York. Yu, M., Nishiumi, H. and Arons, J.D.S., 1993. Thermodynamics of phase separation in aqueous solutions of polymers. Fluid Phase Equilibria, 83: 357-364.