291

Bioe:ec~whemist~~ and BiaenergeHcs, 20 (1988; 291-2%

A section ol J. Ekwroanaf Chem., and constituting Vol. 254 (1988)
Ekvier Scquch S-A., Lausanne - Printtd in The Netherlands

The influence of the ionic streq$h cbnemyme solubilization

iu water-in-oil microeuudsions *

D. Bratko
Depcurmenr of Physic& Chemistg, Fucu&v of Natural Sciences, Universi& E. Karde@, Muraikuva 6.
I&b&no (Yugoslavia)

A. Pnaru
IJW~W#~
of Micr~biofo~, Medical Facu@, University E. Ku&e&, L#b&ana (Yugoslavia)

S.H. Chen
Deparfmen! qf Nuciear Engineering Massachusetts fnstitute of Tech,roiogy. Cambridge, MA 02139 (U.S.A.)
(Received 23 member 1987; in foal form 15 Seplcmber 1988)

Solubilization of enzymes in water&-oil microemulsions [l-9] is important in

various biotechnological processes ranging from selective protein extraction [7] to
biocatalyzed synthesis [S]. It is of particular interest that the activity of enzyme.~
accomodated in the aqueous core of inverted micelles often remains similar to that
in pure aqueous solution [4]. The solubility of ionizing proteins in the microemul-
sion can be regulated by varying different parameters like the detergent concentra-
tion, the detergent-to-water ratio, the temperature, the pH, and the simple electto-
lyte activity in the system [4,8]. In the present study, the influence of the ionic
strength is analyzed in terms of electrostatic contributions to the thkmodynamics of
enzyme/inverted mkeilc complexation. The shell-and-core model [1,3,8] for the
inverted micelle is used in &e Poisson-Bohzmann approximation [lo] to calculate
the free energy of transfer of the enzyme from the bulk aqueous solution to the
aqueous core of an inverted micelIe. The mechanism of the enzyme solubilizatik in
the microemulsion is shown k Fig. 1, The shell-and-core model pictures the inverted
micelles as spherical tintities consisting of the aqueous core and of the detergent
shell of radius rM_ The ionizing miceEed detergent contributes to the ‘micelle
charge zM spread over the idealized smooth micek surface, the surface charge

* Presented al the 9th BES Symposium on Bioeltctrochemist~y and Biwnergetics, Szeged (Hungary),
l-5 September, 1987.

0 1988 E&&r Sequoia S.A.

Fig. 1. ‘I%e model for aqueous enzyme sAtion (left! and for the enzyme-containing microemulsion
(l-i&?).

density being (JM = tMe0/(4mRt,). The water pool in the micelle may host a
hydrophilic enzyme which is assumed to lie in the centre of the micelle. The globular
enzyme is modelled as a uniformly charged sphere of radius rE, charge za, and
surface charge density on = zEe0/(47&. The aqueous solution of the enzyme
maintaining equilibrium with the microemulsion is represented by the spherical cell
model of colloidal solutions [ll-13j. According to this model, the solution is
represented by a large number of equal spherical cells, each containing an enzyme
molecule and the neutralizing simple ions. The volume of the cell V, and its radius rC
depend on the enzyme concentration cE = l/(N,&) where V, = 47r&3. Water in
the cell as well as that in the micellar core is treated as a dielectric continuum of
permittivity t. For the sake of simplicity, the image interactions [14] and the
structural changes of water at the interfaces [15] are not taken into account. The
electric field and the distribution of the simple ions are determined via the
Poisson-Boltzmann (PB) equation

v23,= -f-f&inS exp -F eo\l,

( 1
(1)
i

where $J is the local electrostatic potential, e. the proton charge, k the Bokmann
constant and T the absolute temperaiure. Further, np is the number density of the
z,-valent. simple ion at the surface of the cell, r = rc, or for the micelle, r = rM,
where the potential is set equal to zero for convenience. The boundary conditions
(dWdr),=o = (d\t/dr).= = 0, (d+/dr),, = -GE/C and (d#/dr)r, = CM/c follow
directly from the Gauss theorem. The choice of the appropriate values of the
reference number densities n p and the numerical solution of the PB equation (I),
 293

WEBanalogous to that described earlier [121 in copaection with the modified PI3
equation A more deraikd description will be given in the forthcoming study [16].
The free energy of the enzyme/inverted micclle complexation, AG is defined as
the change is the free energy due to the transfer of an enzyme mokcute from the
bulk aqueous solution to the water pool of an inverted micelle. During this process,
A&.4, molecules of the simple electrolyte MX move from the micelle to the aqueous
solution irl order to satisfy the equilibrium requirement that the electrolyte activity,
aMx = c+c_y& in both the water pool of the enzyme-free inverted micelle and in
the enzyme/tnicelle complex remain equal to that in the bulk solution. Here, c+
and c_ are the cation and the anion concentrations. The activity coeffitient of the
electrolyte, y *, in each of the three different subsystems is approximated by 1171:

y,(k) = I<exp(--go WWXexP(+WOjl-l (2)

where the angular brackets denote the volume average, k = 1 corresponds to the
enzyme-free inverted micelle, k = 3 to the eIlzy_me-containingelementary cell of the
bulk solution, k = 3 to the same ceil after removal of thr: enzyme, and k = 4 to the
enzyme/ inverted micelle complex.
The free energy of complexation is then approximately given by:
AG = G(4) + G(3) - G(2) - G(1) (3)
where ‘G(k) is the free energy of the subsystem k as specified above. Following the
approach of ref. 17, the free energy of the system k, G(k), is estima?ag as the sum
of the pure coulombic term and the configurational entropy term

as described in full detail elsewhere 117,181.Here, p is the local charge density due
to the presence of all ionized species of local Ilumber densities nj, and the sum is
over all solutes, including the protein. The standard chemical potential terms C,
= c i~i(k)Crp were not taken into account in the calculations since they cancel
when the difference AG of eqn. (3) is formed. The energy of enzyme/inverted
micelle complexation AU is defined in an anologous manner as AG in cqn, (3), i.e.
AU = U(4) - U(2) - U(1). Here, U(k) is equal to the coulombic term of the
subsystem k, eqn. (4), multiplied by the factor (1 -k d In E/d In T) [17], and
U(3) = 0 in the present approach.
Jn Fig. 2 we show the ionic strength dependence of the calculated energy and the
free energy of’ the enzyme/inverted micelle compiexation, AU and AG. &Themodel
parameters usr:d in these calculations were chosen to mimic the system involving
cytochrome c in water at cn = 0.0OOlti hJ and in the ~utionic detergent sodium
di-2-ethylhexyl sulfosuccinate (AOT) water-in-isooctane microemulsion at 25 o C.
This system has been character&d re:ently by small angle neutron diffraction
measurements IS]. The rounded radii of ‘rhe enzyme and of the inverted micelle were
taken to be rE = 1.5 nm and pM = 3.0 nm, while r, was 13.5 nm. The charge of the
ionized enzyme zE = + 10 [I!!] and that of the detergent shell in the inverted micelle
 , , , , ,

0.2 0.4 0.6 0.0

Ilm01dm'~

l’ig. 2. The dependence of the energy AU (- -- --) and free energy AG (- ) of enzyme/inverted
micelle complexation on the ionic strength I of the bulk aqueow sohtion.

Z,= -20. This value is consistent with the apparent degree of ionization of ACT
micelles in aqueous solution [20]. The polydispersity and the micellar growth or
shrinking that may accompany the addition of the electrolyte or polyelectrolyte such
as the ionized enqme [4,6,gj were disregarded in the present, preliminary study.
As shown in Fig. 2, the calculated energy of complexation AU is negative over
the whole range OF the ionic strength I of the aqueous enzyme solution in contact
with the microemulsion, hence the process is exothermic at all conditions. With
increasing ionic sPrength AU tends to zero due to the increasing shielding effects in
the ionic atmosphere next to the charged colloidal particles. The calculated AG
grows with I from strongly negative values at low ionic strength to positive values at
high electrolyte concentration. The enzyme solubility in the microemulsion is
therefore high at ionic strengths I below 0.18 M but drops almost to zero at
I > 0.35 M where AG is already highly positive.
The coulombic contribution to AG is proportional to the energy of complexation
AW from which it differs by the factor (i-t d ln c/d ln T). As (1+ d ln c/d In T)
= -0.372 in water at room temperature we note that the coulombic term in AG is
positive over tha whole range of I in the system studied. The negative free cncrgy of
compJexation bG is therefore due mainly to the considerable increase in the
configurational eatropy of the simple ions that leave the water goal of the inverted
micelle during complexation. Clearly, qon entrapment of the positively charged
enzyme in the water pool of the inverted mice-de, an equivalent number of positively
charged counterions move to the aqueous solution since they are no more bound to
the interior of the micelle by the electroneutrality condition. This process is
entropically favorable as long as the corlnterion concentration in the micelJe exceeds
that in the bulk aqueous phase.
A quantitative estimate of the enzyme distribution between the bulk aqueous
phase and the microemulsion is a.Jso of interest. Considering that there are (E) .ways
 295

0 0.2 0.4 0.6

Fig. 3. The fraction of the enzyme-containing inverted micelles in the microemulsion equilibrated with
the aqueous solution of the enzyme as a function of its ionic strength I at zE = 10, zM = - 20, r, = 1.5
nm, rM= 3.0 nm and r, = 13.5 nm or with the same parameters but with rE = 1.g nm (11, zM = - 45 (Z),
RM = 2.4 nm (3), and ZM= - 180 (4).

of distributing N enzyme molecules in the microemulsion contianing M available

inverted micelles, the additional configurational entropy is equal to kT In(!).
Following closely the derivation of the Langmuir adsorption isotherm [21,22], the
equilibrium fraction 8 = N/M of the inverted micelles occupied by the enzyme can
be related to the free energy of complexation of a single enzyme, AG:

8_ exp(-AG/kT)._
1+ exp( - AG/kT)
(5)
Here, the enzyme concentration cE has already been taken into account in the
determination of AG.
In Fig. 3, the calculated fractions of enzyme-containing inverted micelles in a
microemulsion equilibrated with aqueous enzyme solutions of different ionic
strengths are shown. In order to estimate the effect of varying the characteristic
parameters of the model, a few curves for different sizes and charges of the reverse
micelle and the enzyme are included. The solid curve describes the model system
which mimics the cytochrome c i.n d. *OT *zater-in-oil microemulsion spe.&ed above.
in accordti~ti with recent small angle neutron diffraction measurements [S], a high
solubility of this enzyme is predicted at low concentration of the simple electrolyte,
I +z 0.15 M, whereas the enzyme is only poorl, v soluble at I B 0.35-0.4 M, Although
the model calculations correspond to a somewhat different physical situation than
considered in the experiment, the substantial agreement between the measurement
and the calculation appears to justify the thFaretica1 approach proposed above. It is
particularly important that the observed enzyme solubilization may be interpreted
satisfactorily solely in terms of non-specific, coulombic interactions. Possible refine-
ments of the model and the calculation will be discussed later [he] but further
applications of the proposed treatment to other protein/microemulsion systems are
claerly encouraged in view of the present results.
Jm!cNOWLEDoEMENT

This pork wfts supparted by 13SF tirrough the I3Jiotechnology Process Engineering
Center at M.I.T. and through the United States-Yugoslav Joint Fund for Scientific
Cooperation under Grant No. 8711845. Support from the Research Community of
Slovenia is also acknowledged.

REFERENCES

1 F.J. Bonncr. R Wolf and P.L. Luisi, J. Solid Phase Kochem., 5 (1980) 255.
2 A.V. Levasbov, Y.L. Kht~elnitsky, N.L. Klyachko. V.Y. Chemyak and 1;. Martin&, 3. Colbid
;‘ntcrface Sci., 88 (1982) 444.
3 P.L. L&i and R Wolf in K.L. Mittal and J.H. Fendler (Ws.), Solution Behaviour of Surfactants,
vol. 2, Plenum, New York 1982, p. 867.
4 J.H. Fendlcr, Membrane Mimetic Chemistry, Wiley, New York, 19S2, p. 259.
5 P. Luthi and P.L. Luisi, J. Am. Chem. Sot., 106 (1984) 7285.
6 M.P. Pile& T. Zemb and C. Petit, Chcm. Phys. Lett, 118 (1985) 414.
7 ICE GoHen and T.A. Hatton, BiotechnoL Rag., 1 (1985) 69.
8 BY. Shcu, ICE Golden, T.A. Hatton and S.H. Chen, Biotechnol. Prog.. 1 (1986) 175.
9 D. Cbatenay, W. Urbach; C. Nicot, M. Vacher and M. Waks, J. Phys. Chem., 91 (1987) 2198.
10 R Kubii H.F. Eicke and 8. Jonsson, Helv. Chim. Acti, 65 (1982) 170.
11 G.M. Bell and AJ. Dunniug. Trans. Faraday See.. 66 (1970) 500.
12 M. Mille and Cr. Vanderkoci, -7. Colloid Interface Sci., 59 (1977) 211.
13 D. Bra&r, Bioelectrochem. Bioenerg., 13 (1984) 459.
14 D. Bratka, B. J&xson and H. Wennerstram, Chem. Phys. Lett., 128 (1986) 449.
15 A. Luzar, S. Svetina and B. &k& Bioelectrochem. Biceuerg., 13 (1984) 473,
16 D. Bratko, A. Lwar and S.H. Chen, to be published.
17 RA. Marcus, J. Chcm. Phys., 23 (1954) 1057.
18 D. Bratko and D. Dolar, J. Chem. Phys., 80 (1984) 5782.
19 A.L. Lehninger, Biochemistry, Worth Publishers, New York, 1975, p. 102.
20 E.Y. Sheu. S.H. Chm and J.S. Huang, J. Phys. Chem., 91 (1987) 3306.
21 T.L. Hill, An Introductior: to Statistical Thermodynamics, Addison-Wesley, Reading, MA, 1962,124.
22 A.W. Adamson, Physical Chemistry of Surfaces, Wiley, New York, 1976, p. 554.