Protein Engineering vol.10 no.8 pp.847–850, 1997
FORUM
Solubility, crystallization and chromatographic properties of
macromolecules strongly depend on substances that reduce the
ionic strength of the solution
Y.Papanikolau and M.Kokkinidis1
University of Crete, Department of Biology and Institute of Molecular
Biology and Biotechnology (IMBB), PO Box 2208, GR-71409 Heraklion,
Crete, Greece
1To
whom correspondence should be addressed
Keywords: chromatography/crystallization of macromolecules/
ionic strength/solubility
The differential solubility of macromolecules governs their
crystallization and is thus of great importance to crystallography (Blundell and Johnson, 1976; McPherson, 1982,
1985a,b, 1990; Feher and Kam, 1985; Giegé and Mokol, 1989;
Ollis and White, 1990; Giegé and Ducruix, 1992; Giegé
et al., 1994). In addition, precipitation is widely applied as a
fractionation method for the purification of macromolecules
(Ingham, 1984, 1990; Scopes, 1987; Harris, 1989; Roe, 1989;
Englard and Seifter, 1990; Kennedy, 1990; Rossomando, 1990).
The literature offers several observations of factors which
influence macromolecular solubility, among which organic
solvents and polyethylene glycol (PEG) play a major role: it is
known that salts have a ‘poisoning effect’ in PEG crystallization
experiments (McPherson, 1982, 1985b, 1990). Dioxane
(McPherson, 1982), PEG 400 (Ray, 1992) and glycerol (Sousa
and Lafer, 1990) may enhance salting-out crystallization with
ammonium sulfate. Additionally, after ethanol precipitation,
more salt is needed for salting-out precipitation (Harris, 1989).
In this paper, we attempt to rationalize these empirical
observations and to derive some basic principles that determine
macromolecular solubility in aqueous electrolytic solutions
that contain organic solvents or PEG. Understanding these
principles may help in the development of more rational
strategies in macromolecular crystallization and in the use of
ion-exchange and hydrophobic interaction chromatography.
For conciseness, organic solvents or PEG will be termed ‘ionic
strength reducers’ or simply ‘reducers’.
The solubility of macromolecules in electrolytic aqueous
solutions depends on the ionic strength of the solute as
expressed by Green’s general equation (Arakawa and
Timasheff, 1985; Ries-Kautt and Ducruix, 1992):
logS 5 logSW 1 kI(I)1/2 2 kOI
(1)
where SW is the solubility of macromolecules in pure water,
kI and kO are the salting-in and salting-out constants, respectively, and I is the ionic strength of the solute, which is given by
I 5 1/2 (Σ CiZi2)
(2)
Ci and Zi are the concentration and the valence of each different
ion i present in the solute. The factor kI(I)1/2 derives from the
Debye–Hückel theory (Bockris and Reddy, 1970; Ries-Kautt
and Ducruix, 1992) and expresses the contribution of the
© Oxford University Press
electrostatic interactions, which mainly influence salting-in.
The factor 2kOI reflects the net balance between salting-out
through hydrophobic interactions and the salting-in of dipolar
macromolecules (Edsall and Wyman, 1958). The negative sign
is chosen since, in most cases, hydrophobic salting-out is
predominant.
The solubility, SCO, of a macromolecule in a mixed organic
solvent–water or PEG–water solution usually decreases (SCO ,
SW); this is due to a large extent to the decrease in the dielectric
constant of the solution caused by the addition of a given
concentration CO of organic solvent or PEG (Arakawa and
Timasheff, 1985). Assuming this to be the principal factor, the
change in solubility when a macromolecule is transferred from
water to a mixture is given by (Arakawa and Timasheff, 1985)
log(SCO/SW) 5 (A/RT)(1/εW 2 1/εCO)
(3)
where SW and SCO are the solubilities of macromolecules in
pure water and in mixed aqueous solution, respectively, A is
a constant and εW and εCO are the dielectric constants of pure
water and mixed aqueous solution, respectively.
According to the ion-pair theory (Bockris and Reddy, 1970;
Padova, 1972; Vaslow, 1972), a decrease in the dielectric
constant of an electrolytic solution promotes the formation of
ion pairs. The formation of neutral ion pairs decreases the net
charge of the solute and thus reduces the ionic strength of
the solute.
Fig. 1. Schematic representation of Equations 1 and 4. Curve A represents
the logarithm of the solubility of macromolecules as a function of the ionic
strength of the electrolytic aqueous solution according to Green’s general
equation (Equation 1). SW is the solubility of macromolecules in pure water.
The area above curve A represents the supersaturation zone and the area
below curve A represents the undersaturation zone. Curve B is the
logarithm of the solubility of macromolecules as a function of the ionic
strength of the mixed reducer electrolyte aqueous solution according to the
modified Green’s general equation (Equation 4). SCO(SCO , SW) is the
solubility of macromolecules in the ionic strength reducer–water solution.
The area above curve B represents the supersaturation zone and the area
below curve B represents the undersaturation zone. X is a soluble condition
for curve A, but insoluble for curve B. Y is a soluble condition for both
curves A and B. Z is a soluble condition for curve B, but insoluble for
curve A.
847
Y.Papanikolau and M.Kokkinidis
For a concentration CO of organic solvent or PEG, we assign
r as the fraction of the ionic strength active (r , 1). By
replacing SW and I with SCO and rI in Green’s general equation
(Ries-Kautt and Ducruix, 1992), we obtain
logS 5 logSCO 1 kI(rI)1/2 2 kOrI
(4)
Equations 1 and 4 (Figure 1) lead to the prediction that for
a given concentration of electrolyte, ionic strength reducers
decrease the solubility of macromolecules under salting-in
conditions and increase the solubility of macromolecules under
salting-out conditions. Addition of salt has opposite effects.
These predictions have been tested experimentally and
typical results are given in Tables I and II. 2-Methylpentane2,4-diol (MPD) and PEG 6000 promote the precipitation of
ribonuclease A, an effect that can be reversed by the addition
of NaCl (Table I). This is a salting-in phenomenon. By contrast
(Table II), ammonium sulfate (AS) and NaCl promote saltingout precipitation of lysozyme; this can be reversed by the
addition of MPD and PEG 6000. Other ionic strength reducers,
e.g. PEG 200 and glycerol, produce similar results. These
experiments clearly indicate that MPD, PEG and glycerol
reduce the effective salt concentration. The combined effects
of ionic strength reducers and electrolytes on the solubility of
a hypothetical macromolecule, at a given concentration of the
macromolecule, are shown schematically in Figure 2.
Table I. Reduction of ribonuclease A precipitation with increasing ionic
strength of the solution in the presence of MPD and PEG as precipitating
agents
40%
50%
60%
70%
25%
30%
35%
40%
v/v
v/v
v/v
v/v
w/v
w/v
w/v
w/v
MPD
MPD
MPD
MPD
PEG
PEG
PEG
PEG
6000
6000
6000
6000
No additive
0.1 M NaCl
0.2 M NaCl
0.248 6 0.001
0.191 6 0.006
0.002 6 0.001
0.013 6 0.001
0.248 6 0.002
0.253 6 0.002
0.158 6 0.005
0.020 6 0.001
0.253 6 0.005
0.252 6 0.007
0.201 6 0.006
0.021 6 0.001
No additive
0.5 M NaCl
0.137 6 0.002
0.045 6 0.001
0.029 6 0.001
0.012 6 0.001
0.232 6 0.001
0.235 6 0.004
0.237 6 0.004
0.204 6 0.004
Experimental: each solution mixture was vigorously agitated and
subsequently incubated for 2 h at 16°C. The mixtures were then centrifuged
at 15 000 g for 15 min at 16°C. Aliquots of 200 µl were carefully
withdrawn and diluted in 4.8 ml of water. The absorption of the diluted
solution was measured at 280 nm. The solubility of the protein in various
mixtures was monitored by measuring the OD280 of diluted aliquots from
400 µl mixtures of ribonuclease A (10 mg/ml) with the indicated
composition, buffered with 10 mM sodium acetate, pH 5.0. Values are the
average OD280 and the r.m.s. deviation of three independent experiments.
Fig. 2. Schematic representation of the solubility, the interactions with ionexchange and HIC matrices and the nucleation and metastable zones of a
hypothetical macromolecule in a mixed aqueous solution of ionic strength
reducers and electrolytes. The concentration of a hypothetical
macromolecule is considered constant. All other parameters such as pH,
temperature and pressure are considered constant. S is the zone where the
macromolecule is entirely soluble. PI and PO (P 5 precipitation) are zones
where the macromolecule is not entirely soluble. PI and PO are the
precipitation zones for salting-in and salting-out, respectively. The thick
lines that separate the PI and PO areas from area S are the conditions of
infinitesimal precipitation. Ii and Io are the soluble zones where the
macromolecule interacts with the ion-exchange and HIC matrices
respectively. The electrostatic interactions play the most important role in
the II area and the polar and hydrophobic interactions play the most
important role in the IO area. NI and NO define nucleation zones in the
precipitation zones PI and PO, respectively. MI and MO define metastable
zones in the interaction zones II and IO, respectively. X is a zone, where
more complex phenomena (e.g. two-phase systems) occur, which at this
stage require further analysis.
Table II. Reduction of lysozyme precipitation with increasing concentrations of the ionic strength reducers PEG 6000 and MPD in the presence of a
precipitating agent, NaCl or ammonium sulfate
1.50 M NaCl
2.00 M NaCl
2.50M NaCl
3.00 M NaCl
3.50 M NaCl
1.40
1.75
2.10
2.45
2.80
M
M
M
M
M
AS
AS
AS
AS
AS
No additive
1% v/v MPD
5% v/v MPD
10% w/v PEG 6000
1.900 6 0.010
1.131 6 0.006
0.717 6 0.004
0.470 6 0.002
0.309 6 0.002
2.070 6 0.020
1.840 6 0.005
1.265 6 0.001
0.859 6 0.002
0.629 6 0.002
2.020 6 0.020
2.056 6 0.008
2.050 6 0.010
2.039 6 0.005
2.049 6 0.008
2.020 6 0.040
1.950 6 0.030
2.050 6 0.010
1.950 6 0.050
1.970 6 0.020
No additive
5% v/v PEG 200
5% v/v MPD
30% v/v glycerol
1.900 6 0.040
0.470 6 0.050
0.095 6 0.001
0.019 6 0.002
0.008 6 0.001
2.007 6 0.008
1.980 6 0.020
1.500 6 0.020
0.347 6 0.002
0.072 6 0.001
2.070 6 0.030
2.040 6 0.010
0.675 6 0.010
2.050 6 0.010
2.040 6 0.010
2.050 6 0.010
Experimental: each solution mixture was treated as described in Table I. The OD280 of diluted aliquots from 1.0 ml mixtures of lysozyme (10 mg/ml) with the
indicated composition, buffered with 20 mM sodium acetate, pH 5.0, was used to monitor the solubility of the macromolecule. Values are the average OD280
and the r.m.s. deviation of three independent experiments.
848
Solubilities of macromolecules
(Athanasiadis et al., 1997), the Escherichia coli polyamine
induced protein and various designed variants of the ColE1
Rop (repressor of primer) protein (details will be published
elsewhere).
Applications of these concepts can be extended to purification methods such as ion-exchange chromatography and hydrophobic interaction chromatography (HIC), expanding their
potential (Figure 3B). In the presence of ionic strength reducers,
a macromolecule requires a higher salt concentration in order
to be eluted from an ion-exchange chromatographic matrix.
Alternatively, for a given concentration of salt and ionic
strength reducer, a macromolecule can be eluted from an
ion-exchange matrix by a decreasing ionic strength reducer
gradient. Additionally, a macromolecule, in the presence of
ionic strength reducers, can be eluted at higher salt concentrations from HIC matrices. Alternatively, a macromolecule can
be eluted from an HIC matrix by an increasing gradient of
ionic strength reducer at a constant salt concentration (data
will be presented elsewhere).
In conclusion, ionic strength reducers increase electrostatic
interactions and decrease polar interactions between macromolecules in solution. It should be emphasized that each
macromolecule imposes its characteristic features on a mixed
aqueous solution containing both an ionic strength reducer and
an electrolyte, exhibiting a unique behaviour. Knowledge
of these characteristics can be of great help in designing
chromatographic approaches, precipitation procedures and
crystallization experiments.
Acknowledgements
Fig. 3. (A) Crystallization pathways: in salting-in crystallization
experiments, pathways A, B and C indicate various types of changes
induced in the macromolecular solution which lead from the soluble zone
(S) to the nucleation zone (NI); pathways D, E and F lead from the soluble
zone (S) to the metastable zone (MI). In salting-out crystallization
experiments, pathways A9, B9 and C9 lead from the soluble zone (S) to the
mucleation zone (NO); pathways D9, E9 and F lead from the soluble zones
(S) to the metastable zone (MO). (B) Chromatographic pathways: for ionexchange matrices the elution pathway indicated by arrow B shifts the
elution to higher salt concentration compared with the elution pathway
indicated by arrow A. Pathway C indicates the elution from ion-exchange
matrices with a decreasing ionic strength reducer concentration gradient and
constant salt concentration. For hydrophobic interaction matrices the elution
pathway indicated by arrow E shifts the elution to higher salt concentration
compared with the elution pathway indicated by arrow D. Pathway F
indicates the elution from hydrophobic matrices with an increasing ionic
strength reducer concentration gradient and constant salt concentration. PI,
PO, II, IO and X have been defined in Figure 2.
The influence of ionic strength reducers in both salting-in
and salting-out offers versatile approaches to crystallization.
At a constant concentration of an ionic strength reducer, a
decrease in the salt concentration by dialysis can promote the
crystallization of macromolecules in salting-in experiments
(Figure 3A). Additionally, at a constant concentration of salt,
a decrease in the reducer concentration by dialysis can promote
the crystallization of macromolecules in salting-out experiments (Figure 3A). These approaches have given excellent
results in crystallization experiments on macromolecules which
were otherwise very difficult to crystallize, e.g. the M.BseCI
DNA methyltransferase from Bacillus stearothermophilus
We thank Drs S.Timasheff, D.Yphantis, G.Thireos, C.Katerinopoulos,
D.Alexandraki, K.Petratos and V.Bouriotis for critically reading the manuscript
and helpful suggestions. This work was supported by a grant (BIOT-CT910262) in the framework of the European Union BRIDGE programme. Y.P.
was partially supported by a grant (BMH1-CT93-1454) from the BIOMED
programme.
References
Arakawa,T. and Timasheff,S.N. (1985) Methods Enzymol., 114, 49.
Athanasiadis,A., Papanikolaou,Y., Rina,M., Dauter,Z., Petratos,K., Bouriotis,V.
and Kokkinidis,M. (1997) Acta Crystallogr., D53, 477–479.
Blundell,T.L. and Johnson,L.N. (1976) In Protein Crystallography. Academic
Press, London, pp. 59–82.
Bockris,J. and Reddy,A.K.N. (1970) In Bockris,J. and Reddy,A.K.N. (eds),
Modern Electrochemistry, Vol. 1. Plenum Press, New York.
Edsall,J.T. and Wyman,J. (1958) Biophys. Chem., 1, 282.
Englard,S. and Seifter,S. (1990) Methods Enzymol., 182, 285.
Feher,G. and Kam,Z.M. (1985) Methods Enzymol., 114, 77.
Geiǵe,R. and Ducruix,A. (1992) In Ducruix,A. and Giegé,R. (eds),
Crystallization of Nucleic Acids and Proteins. A Practical Approach. IRL
Press, Oxford, pp. 1–18.
Giegé,R. and Mokol,V. (1989) Tibtech, 7, 277.
Giegé,R., Lorber,B. and Théobald-Dietrich,A. (1994) Acta Crystallogr.,
D50, 339.
Harris,E.L.V. (1989) In Harris,E.L.V. and Angal,S. (eds), Protein Purification
Methods. A Practical Approach. IRL Press, New York, pp. 125–174.
Ingham,K.C. (1984) Methods Enzymol., 104, 351.
Ingham,K.C. (1990) Methods Enzymol., 182, 301.
Kennedy,R.M. (1990) Methods Enzymol., 182, 339.
McPherson,A. (1982) In McPherson,A. (ed.), Preparation and Analysis of
Protein Crystals. Krieger, Malabar, FL.
McPherson,A. (1985a) Methods Enzymol., 114, 112.
McPherson,A. (1985b) Methods Enzymol., 114, 120.
McPherson,A. (1990) Eur. J. Biochem., 189, 1.
Ollis,D. and White,S. (1990) Methods Enzymol., 114, 646.
Padova,J. (1972) In Horne,R.A. (ed.), Water and Aqueous Solutions: Structure,
Thermodynamics and Transport Processes. Wiley, New York, pp. 109–174.
Ray,W.J. (1992) Proteins, 14, 300.
849
Y.Papanikolau and M.Kokkinidis
Ries-Kautt,M. and Ducruix,A. (1992) In Ducruix,A. and Giegé,R. (eds),
Crystallization of Nucleic Acids and Proteins. A Practical Approach. IRL
Press, Oxford, pp. 195–218.
Roe,S. (1989) In Harris,E.L.V. and Angal,S. (eds), Protein Purification
Methods. A Practical Approach. IRL Press, New York, pp. 175–244.
Rossomando,E.F. (1990) Methods Enzymol., 182, 309.
Scopes,R.K. (1987) In Scopes,R.K. (ed.), Protein Purification. Principles and
Practice. Springer, New York, 1987.
Sousa,R. and Lafer,E.M. (1990) Methods: Companion Methods Enzymol.,
1, 50.
Vaslow,F. (1972) In Horne,R.A. (ed.), Water and Aqueous Solutions: Structure,
Thermodynamics, and Transport Processes. Wiley, New York, pp. 465–518.
Received March 11, 1997; accepted April 9, 1997
850