- Email: [email protected]
Isoelectric focusing using non-amphoteric buffers in free solution: II. Apparatus and measures of pH stability
Journal of Biochemical and Biophysical Methods, 13 (1986) 275-287
275
Elsevier BBM 00570
Isoelectric focusing using non-amphoteric buffers in free solution: II. Apparatus and measures of pH stability Pierre Wenger and Philippe Javet Swiss FederalInstitute of Technology, Institute of Chemical Engineering CH-1015 Lausanne, Switzerland (Received 3 June 1986) (Accepted 25 August 1986)
Summary Isoelectric focusing of amino acids or proteins requires a time-invariant pH gradient. The maintenance, in an electric field, of such profiles with simple non-amphoteric buffers is possible in multicompartment cells using well determined concentration profiles. The method for determining these concentration profiles has been proposed in a preceding paper [1]. A multicompartment cell has been built for the purpose of verifying the assumptions made. The technical characteristics of this cell will be described here. The time-stability of pH profiles obtained with sodium acetate/acetic acid buffers has been measured for concentration profiles determined so as to keep constant the transport number of each ion throughout the cell. Measured over a time span of 10 h and with a current density of 96 mA cm -2, the pH shift is, in the worst case, 0.009 pH unit/h. This corresponds to a much better stability than the one obtained with a constant concentration of sodium acetate in all compartments, and justifies the chosen concentration profiles based on the condition of constant transport numbers. The above mentioned method for the computation of the buffer concentration assumed a constant relative mobility of the ions. The experiments have shown the limits of this assumption, i.e. the variation of ionic mobilities with the ionic strength and the temperature diminishes the stability of the pH gradient. However, slight adjustments of the concentration in the end compartments allow some improvement of the stability. If the temperature could be precisely controlled in each compartment, a better pH stability than what has been achieved in these experiments should be reached. Two methods of compensation of the migration of ions in the end compartments (buffer renewal method and external recycling method [1]) have been tested and will be discussed. Key words: Isoelectric focusing: Non-amphoteric buffer; Multicompartment electrolyzer; Rheoelectrolysis
Correspondence address: Dr. P. Javet, Swiss Federal Institute of Technology, Institute of Chemical Engineering, CH-1015 Lausanne, Switzerland. List of symbols used: C, concentration (kmol/m 3); Ca, concentration of acid (kmol/m 3); C b, concentration of base (kmol/m3); F, Faraday constant (C/eq); I, electric current (A); J, ionic strength (kmol/m3); Ka, acidity constant (kraal/m3); Kw, ionic product of water (kmol2/m6); P, pressure (Pa); T, temperature (°C); Tk, transport number (-); t, time (s); 15", volumetric flow (ml/min); 1~'(~), volumetric flow in (~) (ml/min); l/C), volumetric flow in (~) (ml/min); U, electric tension (V); ~i, relative ionic mobility (-); fl, buffering capacity (mol/m3). Letters and numbers in a circle indicate a compartment of the cell. 0165-022X/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)
276
Introduction The first apparatus used for isoelectric focusing of proteins or amino acids were multicompartment cells. Svensson [2], in a review article, describes the apparatus developed until 1948. At that early time the pH was not stabilized, but was allowed to vary with time or the charge passed. During some decades, the interest in multicompartment cells was not pursued because of the success of the isoelectric focusing in gel layers using a mixture of carrier-ampholytes (such as Ampholine®, Pharmalyte*, and Servalyte® [3,4]). Recently Rilbe [5-7], Jonsson [8], Martin [9,10] or Bier [11,12] have suggested to replace these expensive ampholytes by simple non-amphoteric buffers. The apparatus using such simple buffers seems to open a very promising way for economical, continuous separations of large amounts of proteins. However, the separations seem to present problems associated with the stability of the pH gradient in an electric field, the choice of membranes, or the characterization of the separations. Many studies in progress are searching for the best stability of a pH gradient formed with simple non-amphoteric buffers. Many authors have mentioned as a stability criterion, the need to keep the transport number constant of all the ions present in the cell. In either acidic (pH < 5.0) or basic (pH > 8) media they note that this condition cannot be fulfilled with the buffers and concentrations used. In a preceding paper [1], the present authors have proposed a way of determining buffer profiles which maintain the constancy of the transport number of each ion, in all the compartments of the cell. The experimental results shown here have allowed to test the proposed concentration profiles and to verify the assumptions made in the theoretical development [1].
Apparatus An electrochemical cell based on the concept developed by Martin [9] has been built. The working diagram of this cell is given in Fig. 1. This cell consists of: Two electrode compartments ((~ and (~)) separated from the others by an ion exchange membrane. The anode is made out of Pt with 0.1 N H2SO4 as anolyte and the cathode of Fe with 0.1 N NaOH as catholyte. The anode is separated from the next compartment by a cationic exchange membrane (Ionics 61 CZL-386), and the cathode is separated by an anionic exchange membrane (Ionics 61 QZL-386). The electrolytes are recirculated from thermostatted reservoirs (L in Fig. 3) which allows the disengagement of the gas formed on the electrodes and the maintenance of a constant temperature. Five working compartments ( ( ~ to (~) in Fig. 3) separated by neutral flow-tight membranes. In these compartments five buffer solutions circulate with increasing pH from the anode to the cathode. An external heat exchanger (E in Fig. 3), which allows the maintenance of a constant temperature.
277
t
°1
Fig. 1. Working scheme of a multicompartment isoelectric focusing cell, with an external heat exchanger. fE = recirculation flow through the heat-exchanger; V = electrolyte flows for compensation of the migration in the end compartments.
Two working compartments are shown in Fig. 2. They are milled in three PVC-plates glued together and have a total thickness of 4 mm. They can be packed at will on the top of each other, to reach the desired number of separation chambers. A lateral distributor and collector allows a homogeneous flow of the buffer solution within the compartment. The electric field is perpendicular to the hydraulic flow. The cross-area of a compartment is thus: for the current 52 cm 2, for the electrolyte 3 cm 2. Membranes (W in Fig. 2) which separate these compartments are prepared by impregnation of two layers of cotton fabric with a hot gel of agarose 6.2% (w/v). This composite is pressed, still hot, between two metallic plates, and then cooled. It is hardened by a chemical treatment: 15 min in 2.5 M NaOH, then left overnight in a solution of epichlorhydrin (8%, v/v) in xylene. These membranes are tightened between compartments with soft silicone gaskets (0.5 mm thick) (V in Fig. 2). The working compartments are then placed between the electrode compartments. The isoelectric focusing cell, assembled in such a manner, is fed with a constant current of 5 A from a stabilized DC-power supply (500 V).
o~v
uoo•o ,./;
w o
V
U
0 0
Fig. 2. Working compartments: U = compartments; V = silicone gasket; W = membrane.
278
F
,ii
,J PC
I
Fig. 3. General view of the apparatus: B = magnetic stirrers; D = peristaltic pumps; E = heat-exchanger; F = isoelectric focusing cell; L = thermostatted electrode tanks; M = five diaphragm pump heads; N = cooling water; P = manometers; pH = pH probes; R = storage tanks; S = centrifugal pumps; T = temperature probes (Pt 100).
The heat exchanger (E in Fig. 3) is made out of 0.5 mm thick steel plates. The heat exchanger surface is 312 cm 2 for each circuit. With a mean temperature difference of 10°C, the dissipated power is approximately 160 W for each circuit. The buffer solutions are recirculated through this heat exchanger by a membrane pump (M in Fig. 3) with five synchron heads which allow a controlled flow through in each compartment of the isoelectric focusing cell (F in Fig. 3). The flow rate is chosen to have an effective cooling and identical pressure in each compartment. For example, with an electric current of 5 A, a throughflow of 30 1/h is required to maintain an increase of the temperature of 1-5 ° C between the inlet and the outlet of each compartment (depending on the conductivity of the buffer solution). Storage tanks of 200 ml (R in Fig. 3) are located in each recirculation loop (total volume of one loop: 500 ml). These reservoirs allow for the introduction of the buffer solution in each loop at the start-up of the system, and to withdraw samples for analysis. The content of the two tanks of the end compartments (~) and © is stirred (B in Fig. 3) to allow the addition or the withdrawal of solution to keep the buffer composition in the end compartments constant [1]. When the buffer renewal method is used, the additions are delivered by two peristaltic pumps (D in Fig. 3), and the withdrawals by overflow systems. The pH in the end compartments is measured and printed at regular time intervals. It is used to control the flow of the two peristaltic pumps that feeds basic solution to compartment (~) and acidic solution to compartment © to maintain the pH in the end compartments at a constant value [1].
279 When the method with external recycling of the buffers is used, there is no addition of acidic or basic solutions to the end compartments. A peristaltic pump is used to transfer a controlled part of the content of compartment (~) to compartment © while an overflow system between these compartments transfers an equal volume of the buffer solution © back to (~). The rate of recycling depends on the electric current and the composition of the buffer solutions in the end compartments [1]. Stability of preformed pH gradients prepared with acetic acid/sodium acetate buffers (pK a = 4.76 at infinite dilution J = 0 and T = 15°C [16]), has been measured in the cell described in section 2. As described earlier [1], the compositions of the buffer solutions are computed in such a way as to keep the transport number of each ion constant throughout the cell. This study was undertaken to verify that these calculated compositions allow a stabilization of pH gradient which maintains correct separation conditions in the isoelectric focusing cell over a sufficiently long period of time. As mentioned in our preceding paper [1], two operating modes are proposed: (i) buffer renewal method in the end compartments e.g. addition of sodium acetate solution in compartment (~) and of acetic acid solution in compartment © ; (ii) buffer recycling method from compartment (~) to © , and from © to ( ~ . These two cases will be examined separately.
Buffer renewal method Based on the theoretical expressions developed in our preceding paper [1], the compositions of the buffer solutions have been computed for constant transport numbers and the experimental conditions are given in Table 1. The following values for the relative ionic mobilities ( J = 0.05 M, T = 15°C) and the ionic product of water ( T = 15°C) have been used: ~a+=l.00;
]~A-I=0.69;
~ri+=8.15;
1~oH-1=4.45;
gw=0.6.10-~'
The values chosen for the protonic transport number (TH+ = 0.0024) and the pH in each compartment fix the concentration of sodium acetate to be added in the buffer solution. The concentration of acetic acid to be dissolved in the buffer solution to obtain the chosen pH was computed using the equation of acidity equilibrium (Ka) with the activity. The activity coefficients have been evaluated as described [14]. In this article, the transport numbers have been computed using the ionic concentrations. Since the pH values measured with glass pH electrodes give the protonic activity, the protonic and hydroxyl concentrations have been determined using computed hydroxyl and protonic activity coefficients. The obtained stability of the pH profile as well as measured conductivity are shown in Fig. 5. Fig. 5 shows the remarkable pH stability over a time span of 10 h, using the proposed concentration profiles. A small residual shift of 0.01 p H / h (0.05 pH/Faraday) is nevertheless still present in compartment (~), which is of course the most sensitive since its buffering capacity (43.2 mM) is the lowest. This residual pH shift is due to a difference in experimental transport numbers with regard to the computed values.
280
TABLE 1 BUFFER RENEWAL METHOD Circuits
0
(D
(~)
(~)
0
4.30 271.95 125.75 177.0
4.55 86.01 67.10 83.6
4.75 34.25 40.82 43.2
4.90 17.17 28.22 25.4
0.0024 3.10 - 9 0.4074 0.5092
0.0024 1.10 -8 0.4074 0.5092
0.0024 3.10 -8 0.4074 0.5092
0.0024 5.10 -8 0.4074 0.5092
Buffer compositions: pH Ca (mM) C b (mM) fl (mM)
4.15 542.57 183.96 268.5
Computed transport numbers; Tn + TOH TA TNa~
0.0024 2.10 - 9 0.4074 0.5092
~Na+=l.00; [ ~ A - 1 ~ 0 . 6 9 ; fiH +=8.15; [~on 1=4.45 Experimental conditions for the stabilization of the p H gradient illustrated in Fig. 5. Compositions and flow rates of the solutions added in the end compartments; compositions of the buffer solutions computed to keep constant transport numbers (TH* = 0.0024). I = 5 A; U = 145 V ---,130 V; t -- 10 h; I/¢'~,~ X" = 5.7 m l / m i n CCH3COONa = 0.505 M; ~-~l/t~= 65.0 m l / m i n CCH3COOn = 0.0649 M.
Fig. 4. Overview of the isoelectric focusing cell device: B = magnetic stirrers; D = peristaltic pumps; E = heat-exchanger; F = isoelectric focusing cell; L = electrode tanks; M = five diaphragm p u m p heads; N = cooling water; P = manometers; p H = p H glass electrodes; p r i m = p H meter; T = temperature probes; T m = digital thermometers; --, = loop of the buffer solutions.
281
5.0 pH 4.8
4.6
(1.72)
(1.71)
(1.72)
(2.4o)
(2.43)
<2.4 L
(3.63)
(4.09)
(4.40)
(6.52)
(6.99)
(7.19)
4.4
4.2
['" °"" °-'---4.0
0
Q~
(9~40) I
i
t20
240
o ,
o-(~)
o-
(8.96)
(~'I°I
360 Time
I
I
600
480
A
720
( min )
Fig. 5. Buffer renewal method: variations of pH and conductivity with time in the working compartments. The buffer compositions have been computed to keep a constant protonic transport number (Table 1: Tx+ = 0.0024). Electric current: 5 A. Conductivities are indicated by the numbers in parentheses and are given in # - 1. cm - I. 10- 3.
5.0 (2.81)
pH
A~
4.B
I
/="(2.54)
~--~-T----" ,-~-4.7
4.5 0
(2.40)
(2.56)
=
(2.6oi
~ O
(2.39)'~
0(2.43)
i
i
i
120
E40
360
4BO
Time ( min ) Fig. 6. Buffer renewal method: variations of pH and conductivities (I2-1. cm-}. 10S-3) with time in the compartment (~) for various buffer compositions in the compartment (~): ( ~ ) Cb('c] = 0.0282 M (value computed in agreement with our previous paper [1]); ( i ) CbtLE)= 0.0250 M; (A) Cb~ = 0.0233 M. Electric current: 5 A. The starting concentrations in compartment (~) are always the same, as in Table 1.
282 Three facts may explain this difference. (1) The small electric conductivity of the electrolyte in compartment (~) induces a heat evolution larger than in the other chambers. The physical constants, and thus the transport numbers are temperature dependent. At the temperature of this compartment, the transport numbers are not identical any more to the one prevailing in the others. (2) The numer, ical values for ionic mobilities as reported by various authors in the literature exhibit discrepancy. The values adopted here for the computation [15], are probably slightly different from those prevailing in the agarose membranes. (3) The absolute ionic mobilities are a function of the ionic strength of the solution. This last parameter is different in each compartment. The relative mobilities, assumed constant in the determination of the concentration profiles, could vary somewhat from one compartment to the other due to the different ionic strength in each of them. Fig. 6 shows the measured p H stability in compartment (~) as a function of time, after some adjustment of the computed concentration in compartment © has been made. It is to be noted that small concentration variations in compartment © result in the inversion of the slope of the residual p H shift in compartment (~) which ensures a much better p H stability. However, the remaining variations of conductivity show that the equality of transport numbers is still not completely reached. These p H stabilities have been compared with those obtained with a constant concentration of base in each compartment (namely C b = 40.8 raM). The results for these experimental conditions are shown in Fig. 7.
5.0 pH
(2.59) 4.8
(2.60)
i~2. 45 ) •
1"-1~_(2.18
4.6
, 45)
4.4
_
®
<~~ < ~
4.2
)
(2.28) ~ ~ <~...._ ~ )
(2°42) ~Z~--.... (2.40) A""A~A ~
O
...~,O ~ E ] ~ FI~E] ~ (2.49) (2,57) I
4.0
0
120
,
I
240
Q ,
I
,
I
360 480 Time ( min )
i
1
600
A,
720
Fig. 7. Buffer renewal method: variations of pH and conductivity (/2 1-cm-i-10-3) with time in the working compartments with a constant concentration of sodium acetate in the buffer solutions (Table 2: Cb = 40.82 raM). Electric current 5 A.
283 TABLE 2 BUFFER RENEWAL METHOD Circuits
@
(~)
(~)
@
4.30 96.60 40.82 61.2
4.55 54.29 40.82 52.0
4.75 34.25 40.82 43.2
4.90 24.24 40.82 36.4
0.0068 9.10 -9 0.4058 0.5874
0.0038 2.10 -8 0.4069 0.5893
0.0024 3.10 -8 0,4074 0,5902
0.0017 4.10 -8 0.4077 0.5906
(~)
Buffer compositions: pH Ca (mM) C b (mM) fl (mM)
4.15 136.50 40.82 65.5
Computed transport numbers: TH +
TOllT^ TNa +
0.0096 7-10 -9 0.4048 0.5856
fiN,+=1.00; IfiA 1=0.69; fi8+=8.15; IfioH-l=4.451 Experimental conditions for the stabilization of the pH gradient illustrated in Fig. 7. Compositions and flow rates of the solutions added in the end compartments; composition of the buffer solutions computed by keeping constant the concentration of sodium acetate (C b = 40.82 mM). I = 5 A; I/(~)= 22.56 m l / m i n CCH3COON, = 0.1215 M; I)(~) = 44.98 m l / m i n
CCH3COO H =
0.0932 M.
External recycling of the buffers between end compartments In the end compartments, the stabilization of the buffer compositions can be obtained by recycling the buffer from one compartment to the other at a flow rate
TABLE 3 B U F F E R RECYCLING METHOD Circuits
~)
(~
(~)
(~)
(~
4.40 105.7 57.45 79.8
4.80 16.76 21.34 21.9
5.20 2.66 8.09 4.8
5.65 50 500 127.2
0.003 7.10 -9 0.443 0.554
0.003 4.10 - s 0.443 0.554
0.003 3.10 -7 0.443 0.554
2.10 - 5 2.10 s 0.444 0.556
Buffer compositions: pH C. (mM) C b (mM) fl (mM)
4.00 667 158 241.1
Computed transport numbers: Tn + TollTA-
TNa+
0.003 1.10 -9 0.443 0.554
fiNa+ = 1.0; I fiA- I = 0.8; fill+ = 7.2; l fioH- I ~ 3.9 Experimental conditions for the stabilization of the pH gradient illustrated in Fig. 8. In compartments (~) to (~), the composition of the buffer solutions is computed to keep constant transport numbers (Trt+ = 0.003). In compartment (~), the buffer composition is computed to allow a stabilization of the p H gradient by the buffer recycling method [1]. I = 2 A; 1;"= 2 ml/min.
284 B.B
pH
6"2 f 5 B (21.8)
(21.0)
©
5 4 -
!-,4.1 .o/ s
o
_
I--i_
I
(0.72)
4 B -
i___.
(3.3) ~
(I .65) i _ (3.3)
A--A_A
~
A
42 _(7 9 0 o 3
B
,
o
,
@ @ @ @
/ n ~ l ~
,
I
t20
L ,
,
J
I
J
a4o Time ( min
i
1
360
I
I
I
d80
Fig. 8. External recyclingmethod: variations of pH and condactivity(J2-l. cm-1.10 3) with time. The compositions of the buffer solutions are given in Table 3. Electric current: 2 A. Recycling flow: 2.2 ml/min. established by a mass balance [1,13]. To verify the equations developed in ref. 13, the following experiment has been made: the pH of the five working compartments has been chosen and the composition of the solution of chambers (~) to (~) has been evaluated so as to keep the transport numbers constant. The composition of compartment C ) is thus fixed, but the transport numbers, as demonstrated in our previous paper [1], are not identical to those of the other compartments. The experimental conditions are given in Table 3, and the stability of pH profiles as well as the conductivities are illustrated in Fig. 8. According to the theory, the recycling flow should be 2 ml/min; experimentally, the stabilization has been reached with a slightly higher flow, 2.2 ml/min. However, even with this corrected value, the pH of compartment 3 shifts markedly and this for two reasons: (1) the buffering capacity of this compartment is the lowest (fl = 4.8 mM); (2) the transport numbers of ions in compartments (~) and ( ~ a r e not and cannot be identical [1]. However, if one reduces the pH range covered in one experiment, the stability can be increased.
Discussion The above reported results demonstrate the possibility of stabilizing a pH profile in acid medium with simple non-amphoteric buffers. In a multicompartment cell, a range of 0.75 pH unit can be stabilized for more than 10 h, with a current density of 96 rnA cm-2.
285 A criterion for the pH stability can be defined in each compartment as the inverse of the pH shift per Faraday and per unit of buffering capacity. With this criterion, the stabilities measured here are 4-20 times higher than the one given in the literature [8-10], where the condition of constant transport numbers is not satisfied. In acid medium, the pH stability of compartment (~) which is the most sensitive due to its low buffering capacity can be improved by a small adjustment of the composition of the next compartment ( ( ~ ) . This observation shows that the condition of equal ionic transport numbers is not entirely fulfilled. A possible explanation of this residual pH shift can be obtained from the experiment conducted with a constant concentration of base (Table 2). Since the ionic strength and the temperature in the cell were constant during this experiment (identical conductivity of the buffer solutions), the transport numbers can be computed for each compartment using equivalent ionic conductivities [15]. The values are reported in Table 2. With these calculated transport numbers, the mass balance over compartment (~) based on the migrational fluxes entering and leaving the compartment suggests that sodium ions are depleted, protons and acetate ions are accumulating and acetic acid is produced. Electroneutrality is, of course, maintained: the accumulation of protons is exactly balanced by the increase in sodium ions and the increase in acetate ions. This appears experimentally (Fig. 7) as a decrease in both pH and conductivity in compartments (1), (~) and (~) due to the formation of acetic acid and the depletion of sodium ions. The ionic transport numbers calculated using equivalent ionic conductivity allow thus to predict the variation of the buffer concentration in a compartment. The residual pH shift observed with concentration profiles which are supposed to keep a constant transport number (Fig. 5), may be explained by the same arguments. These concentration profiles imply profiles of both conductivity and ionic strength (0.01-0.1 M). Temperature profiles (10-20°C) between compartments (~) and © are then observed due to the heat exchanger which does not allow variable heat transfer rates from each compartment. These variations in ionic strength and temperature from one compartment to another cause slight differences of relative ionic mobilities between the compartments. Then ionic transport numbers (Table 1) are reevaluated: with these new values corresponding to the experimental conditions (ionic strength, temperature), the transport numbers of Na + and H + decrease from compartment (~) to compartment ( ~ by about 5%, while the transport numbers of acetate ion increase by 8%. The mass balance over the ionic species in the compartments allows the prediction of the increase in the concentration of sodium acetate and acetic acid. Effectively, an increase in the conductivity of the buffer solution contained in the central compartments is observed experimentally (Fig. 5). These observations indicate the necessity to limit the variations of the relative ionic mobilities to reach a better pH stability. It is possible to work only with small difference of buffer concentrations between adjacent compartments or using a buffer in which the ionic mobilities of the components vary in the same manner with the ionic strength. One could use the temperature of the buffers to adjust ionic mobilities.
286 It has been shown [13] that agarose membranes with cotton supports are not entirely neutral causing transport of water by electroosmosis from one compartment to another. The smaller the buffer concentration, the more important the water transport, which contributes also to the residual pH shift. Martin [9] proposes the utilization of amphoteric membranes to suppress electroosmosis. Such membranes have been prepared and their tests indicate that they can effectively reduce electroosmosis (inversion of water transport depending on the pH)~'However, they have not yet been tested in the isoelectric focusing cell. Finally, these experiments indicate the problems associated with the two methods of stabilizing buffer composition in the end compartments: (a) Using a fresh solution allows the stabilization of the composition and the p H of the buffer solutions as time progresses, even in the case of very dilute buffer concentrations (C b = 25 mM). The second experiment (Fig. 6) indicates how important it is to maintain the composition in the end compartments constant in order to avoid the shift in the p H and conductivity in the central compartments. This method requires large solution volumes, which means the smaller the buffer concentration and the larger the electric current, the larger the volumes required. (b) External recirculation of the buffers between the end compartments [9,17,18] allows the stabilization of the buffer composition in these compartments (Fig. 8), if the conditions given in the previous article [1] are fulfilled. Some problems of pH stabilization appear in the central compartments because the initial hypothesis that the transport numbers should remain constant between these compartments and the end compartments is not fulfilled experimentally.
Conclusions The installation of a multicompartment isoelectric focusing device has made possible the study of p H profiles stabilized with simple non-amphoteric buffers such as acetic acid and sodium acetate. The experiments show the importance of the constant ionic transport numbers, which requires accurate control of the concentration profiles while minimizing the variation of the relative ionic mobilities. Although other non-amphoteric buffers may be investigated in different pH ranges, the p H stability obtained with acetic acid/sodium acetate buffer over about 10 h is sufficient for the separation of amino acids in a multicompartment isoelectric focusing device. The next article [19] will describe experiments related to separations and will recommend parameters for the scale-up of multicompartment apparatus.
References 1 2 3 4
Wenger,P. and Javet, Ph. (1986) J, Biochem. Biophys. Methods 13, 259-273 Svensson,H. (1948) Adv. Protein Chem. 4, 251-295 Righetti, P.G. (1975) Sep. Purif. Methods 4, 23-72 Righetti, P.G. (1983) IsoelectricFocusing: Theory, Methodologyand Applications, Elsevier,Amsterdam
287 5 Rilbe, H. (1969) Prot. Biol. Fluids 17, 369 6 Rilbe, H., Forchheimer, A., Petterson, S. and Jonsson, M. (1975) in Progress in Isoelectric Focusing and Isotachophoresis (Righetti, P.G., ed.) pp. 51-63, North Holland Publ. Co., Amsterdam 7 Rilbe, H. (1978) J. Chromatogr. 159, 193-205 8 Jonsson, M. and Fredriksson, S. (1981) Electrophoresis 2, 193-203 9 Martin, A.J.P. and Hampson, F. (1978) J. Chromatogr. 159, 101-110 10 Martin, A.J.P. and Hampson, F. (1981) U.S. Patent No. 4'243'507' 11 Bier, M., Egen, N.B., Allgyer, T.T., Twitty, G.E. and Mosher, R.A. (1979) in Peptides: Structure and Biological Function (Gross, E. and Meienhofer, J., eds.) pp. 79-89, Pierce, Rockford, IL, USA 12 Bier, M. and Egen, N.B. (1979) in Electrofocus '78 (Haglund, H., Westerfeld, J.-G. and Ball, J.-T., eds.) pp. 35-48, Elsevier, Amsterdam 13 Wenger, P. (1985) Ph.D. Thesis No 604, EPFL, Lausanne, Switzerland 14 Kielland, J. (1937) J. Am. Chem. Soc. 59, 1675 15 Dobos, D. (1975) Electrochemical Data, Elsevier, Amsterdam 16 Martell, A.E. and Smith, R.M. (1974) Critical Stability Constants, Plenum Press, New York 17 Rilbe, H. (1981) Electrophoresis 2, 261-272 18 Rilbe, H. (1982) Electrophoresis 3, 332-336 19 Wenger, P. and Javet, Ph. (1986) J. Biochem. Biophys. Methods 13, 289-303
275
Elsevier BBM 00570
Isoelectric focusing using non-amphoteric buffers in free solution: II. Apparatus and measures of pH stability Pierre Wenger and Philippe Javet Swiss FederalInstitute of Technology, Institute of Chemical Engineering CH-1015 Lausanne, Switzerland (Received 3 June 1986) (Accepted 25 August 1986)
Summary Isoelectric focusing of amino acids or proteins requires a time-invariant pH gradient. The maintenance, in an electric field, of such profiles with simple non-amphoteric buffers is possible in multicompartment cells using well determined concentration profiles. The method for determining these concentration profiles has been proposed in a preceding paper [1]. A multicompartment cell has been built for the purpose of verifying the assumptions made. The technical characteristics of this cell will be described here. The time-stability of pH profiles obtained with sodium acetate/acetic acid buffers has been measured for concentration profiles determined so as to keep constant the transport number of each ion throughout the cell. Measured over a time span of 10 h and with a current density of 96 mA cm -2, the pH shift is, in the worst case, 0.009 pH unit/h. This corresponds to a much better stability than the one obtained with a constant concentration of sodium acetate in all compartments, and justifies the chosen concentration profiles based on the condition of constant transport numbers. The above mentioned method for the computation of the buffer concentration assumed a constant relative mobility of the ions. The experiments have shown the limits of this assumption, i.e. the variation of ionic mobilities with the ionic strength and the temperature diminishes the stability of the pH gradient. However, slight adjustments of the concentration in the end compartments allow some improvement of the stability. If the temperature could be precisely controlled in each compartment, a better pH stability than what has been achieved in these experiments should be reached. Two methods of compensation of the migration of ions in the end compartments (buffer renewal method and external recycling method [1]) have been tested and will be discussed. Key words: Isoelectric focusing: Non-amphoteric buffer; Multicompartment electrolyzer; Rheoelectrolysis
Correspondence address: Dr. P. Javet, Swiss Federal Institute of Technology, Institute of Chemical Engineering, CH-1015 Lausanne, Switzerland. List of symbols used: C, concentration (kmol/m 3); Ca, concentration of acid (kmol/m 3); C b, concentration of base (kmol/m3); F, Faraday constant (C/eq); I, electric current (A); J, ionic strength (kmol/m3); Ka, acidity constant (kraal/m3); Kw, ionic product of water (kmol2/m6); P, pressure (Pa); T, temperature (°C); Tk, transport number (-); t, time (s); 15", volumetric flow (ml/min); 1~'(~), volumetric flow in (~) (ml/min); l/C), volumetric flow in (~) (ml/min); U, electric tension (V); ~i, relative ionic mobility (-); fl, buffering capacity (mol/m3). Letters and numbers in a circle indicate a compartment of the cell. 0165-022X/86/$03.50 © 1986 Elsevier Science Publishers B.V. (Biomedical Division)
276
Introduction The first apparatus used for isoelectric focusing of proteins or amino acids were multicompartment cells. Svensson [2], in a review article, describes the apparatus developed until 1948. At that early time the pH was not stabilized, but was allowed to vary with time or the charge passed. During some decades, the interest in multicompartment cells was not pursued because of the success of the isoelectric focusing in gel layers using a mixture of carrier-ampholytes (such as Ampholine®, Pharmalyte*, and Servalyte® [3,4]). Recently Rilbe [5-7], Jonsson [8], Martin [9,10] or Bier [11,12] have suggested to replace these expensive ampholytes by simple non-amphoteric buffers. The apparatus using such simple buffers seems to open a very promising way for economical, continuous separations of large amounts of proteins. However, the separations seem to present problems associated with the stability of the pH gradient in an electric field, the choice of membranes, or the characterization of the separations. Many studies in progress are searching for the best stability of a pH gradient formed with simple non-amphoteric buffers. Many authors have mentioned as a stability criterion, the need to keep the transport number constant of all the ions present in the cell. In either acidic (pH < 5.0) or basic (pH > 8) media they note that this condition cannot be fulfilled with the buffers and concentrations used. In a preceding paper [1], the present authors have proposed a way of determining buffer profiles which maintain the constancy of the transport number of each ion, in all the compartments of the cell. The experimental results shown here have allowed to test the proposed concentration profiles and to verify the assumptions made in the theoretical development [1].
Apparatus An electrochemical cell based on the concept developed by Martin [9] has been built. The working diagram of this cell is given in Fig. 1. This cell consists of: Two electrode compartments ((~ and (~)) separated from the others by an ion exchange membrane. The anode is made out of Pt with 0.1 N H2SO4 as anolyte and the cathode of Fe with 0.1 N NaOH as catholyte. The anode is separated from the next compartment by a cationic exchange membrane (Ionics 61 CZL-386), and the cathode is separated by an anionic exchange membrane (Ionics 61 QZL-386). The electrolytes are recirculated from thermostatted reservoirs (L in Fig. 3) which allows the disengagement of the gas formed on the electrodes and the maintenance of a constant temperature. Five working compartments ( ( ~ to (~) in Fig. 3) separated by neutral flow-tight membranes. In these compartments five buffer solutions circulate with increasing pH from the anode to the cathode. An external heat exchanger (E in Fig. 3), which allows the maintenance of a constant temperature.
277
t
°1
Fig. 1. Working scheme of a multicompartment isoelectric focusing cell, with an external heat exchanger. fE = recirculation flow through the heat-exchanger; V = electrolyte flows for compensation of the migration in the end compartments.
Two working compartments are shown in Fig. 2. They are milled in three PVC-plates glued together and have a total thickness of 4 mm. They can be packed at will on the top of each other, to reach the desired number of separation chambers. A lateral distributor and collector allows a homogeneous flow of the buffer solution within the compartment. The electric field is perpendicular to the hydraulic flow. The cross-area of a compartment is thus: for the current 52 cm 2, for the electrolyte 3 cm 2. Membranes (W in Fig. 2) which separate these compartments are prepared by impregnation of two layers of cotton fabric with a hot gel of agarose 6.2% (w/v). This composite is pressed, still hot, between two metallic plates, and then cooled. It is hardened by a chemical treatment: 15 min in 2.5 M NaOH, then left overnight in a solution of epichlorhydrin (8%, v/v) in xylene. These membranes are tightened between compartments with soft silicone gaskets (0.5 mm thick) (V in Fig. 2). The working compartments are then placed between the electrode compartments. The isoelectric focusing cell, assembled in such a manner, is fed with a constant current of 5 A from a stabilized DC-power supply (500 V).
o~v
uoo•o ,./;
w o
V
U
0 0
Fig. 2. Working compartments: U = compartments; V = silicone gasket; W = membrane.
278
F
,ii
,J PC
I
Fig. 3. General view of the apparatus: B = magnetic stirrers; D = peristaltic pumps; E = heat-exchanger; F = isoelectric focusing cell; L = thermostatted electrode tanks; M = five diaphragm pump heads; N = cooling water; P = manometers; pH = pH probes; R = storage tanks; S = centrifugal pumps; T = temperature probes (Pt 100).
The heat exchanger (E in Fig. 3) is made out of 0.5 mm thick steel plates. The heat exchanger surface is 312 cm 2 for each circuit. With a mean temperature difference of 10°C, the dissipated power is approximately 160 W for each circuit. The buffer solutions are recirculated through this heat exchanger by a membrane pump (M in Fig. 3) with five synchron heads which allow a controlled flow through in each compartment of the isoelectric focusing cell (F in Fig. 3). The flow rate is chosen to have an effective cooling and identical pressure in each compartment. For example, with an electric current of 5 A, a throughflow of 30 1/h is required to maintain an increase of the temperature of 1-5 ° C between the inlet and the outlet of each compartment (depending on the conductivity of the buffer solution). Storage tanks of 200 ml (R in Fig. 3) are located in each recirculation loop (total volume of one loop: 500 ml). These reservoirs allow for the introduction of the buffer solution in each loop at the start-up of the system, and to withdraw samples for analysis. The content of the two tanks of the end compartments (~) and © is stirred (B in Fig. 3) to allow the addition or the withdrawal of solution to keep the buffer composition in the end compartments constant [1]. When the buffer renewal method is used, the additions are delivered by two peristaltic pumps (D in Fig. 3), and the withdrawals by overflow systems. The pH in the end compartments is measured and printed at regular time intervals. It is used to control the flow of the two peristaltic pumps that feeds basic solution to compartment (~) and acidic solution to compartment © to maintain the pH in the end compartments at a constant value [1].
279 When the method with external recycling of the buffers is used, there is no addition of acidic or basic solutions to the end compartments. A peristaltic pump is used to transfer a controlled part of the content of compartment (~) to compartment © while an overflow system between these compartments transfers an equal volume of the buffer solution © back to (~). The rate of recycling depends on the electric current and the composition of the buffer solutions in the end compartments [1]. Stability of preformed pH gradients prepared with acetic acid/sodium acetate buffers (pK a = 4.76 at infinite dilution J = 0 and T = 15°C [16]), has been measured in the cell described in section 2. As described earlier [1], the compositions of the buffer solutions are computed in such a way as to keep the transport number of each ion constant throughout the cell. This study was undertaken to verify that these calculated compositions allow a stabilization of pH gradient which maintains correct separation conditions in the isoelectric focusing cell over a sufficiently long period of time. As mentioned in our preceding paper [1], two operating modes are proposed: (i) buffer renewal method in the end compartments e.g. addition of sodium acetate solution in compartment (~) and of acetic acid solution in compartment © ; (ii) buffer recycling method from compartment (~) to © , and from © to ( ~ . These two cases will be examined separately.
Buffer renewal method Based on the theoretical expressions developed in our preceding paper [1], the compositions of the buffer solutions have been computed for constant transport numbers and the experimental conditions are given in Table 1. The following values for the relative ionic mobilities ( J = 0.05 M, T = 15°C) and the ionic product of water ( T = 15°C) have been used: ~a+=l.00;
]~A-I=0.69;
~ri+=8.15;
1~oH-1=4.45;
gw=0.6.10-~'
The values chosen for the protonic transport number (TH+ = 0.0024) and the pH in each compartment fix the concentration of sodium acetate to be added in the buffer solution. The concentration of acetic acid to be dissolved in the buffer solution to obtain the chosen pH was computed using the equation of acidity equilibrium (Ka) with the activity. The activity coefficients have been evaluated as described [14]. In this article, the transport numbers have been computed using the ionic concentrations. Since the pH values measured with glass pH electrodes give the protonic activity, the protonic and hydroxyl concentrations have been determined using computed hydroxyl and protonic activity coefficients. The obtained stability of the pH profile as well as measured conductivity are shown in Fig. 5. Fig. 5 shows the remarkable pH stability over a time span of 10 h, using the proposed concentration profiles. A small residual shift of 0.01 p H / h (0.05 pH/Faraday) is nevertheless still present in compartment (~), which is of course the most sensitive since its buffering capacity (43.2 mM) is the lowest. This residual pH shift is due to a difference in experimental transport numbers with regard to the computed values.
280
TABLE 1 BUFFER RENEWAL METHOD Circuits
0
(D
(~)
(~)
0
4.30 271.95 125.75 177.0
4.55 86.01 67.10 83.6
4.75 34.25 40.82 43.2
4.90 17.17 28.22 25.4
0.0024 3.10 - 9 0.4074 0.5092
0.0024 1.10 -8 0.4074 0.5092
0.0024 3.10 -8 0.4074 0.5092
0.0024 5.10 -8 0.4074 0.5092
Buffer compositions: pH Ca (mM) C b (mM) fl (mM)
4.15 542.57 183.96 268.5
Computed transport numbers; Tn + TOH TA TNa~
0.0024 2.10 - 9 0.4074 0.5092
~Na+=l.00; [ ~ A - 1 ~ 0 . 6 9 ; fiH +=8.15; [~on 1=4.45 Experimental conditions for the stabilization of the p H gradient illustrated in Fig. 5. Compositions and flow rates of the solutions added in the end compartments; compositions of the buffer solutions computed to keep constant transport numbers (TH* = 0.0024). I = 5 A; U = 145 V ---,130 V; t -- 10 h; I/¢'~,~ X" = 5.7 m l / m i n CCH3COONa = 0.505 M; ~-~l/t~= 65.0 m l / m i n CCH3COOn = 0.0649 M.
Fig. 4. Overview of the isoelectric focusing cell device: B = magnetic stirrers; D = peristaltic pumps; E = heat-exchanger; F = isoelectric focusing cell; L = electrode tanks; M = five diaphragm p u m p heads; N = cooling water; P = manometers; p H = p H glass electrodes; p r i m = p H meter; T = temperature probes; T m = digital thermometers; --, = loop of the buffer solutions.
281
5.0 pH 4.8
4.6
(1.72)
(1.71)
(1.72)
(2.4o)
(2.43)
<2.4 L
(3.63)
(4.09)
(4.40)
(6.52)
(6.99)
(7.19)
4.4
4.2
['" °"" °-'---4.0
0
Q~
(9~40) I
i
t20
240
o ,
o-(~)
o-
(8.96)
(~'I°I
360 Time
I
I
600
480
A
720
( min )
Fig. 5. Buffer renewal method: variations of pH and conductivity with time in the working compartments. The buffer compositions have been computed to keep a constant protonic transport number (Table 1: Tx+ = 0.0024). Electric current: 5 A. Conductivities are indicated by the numbers in parentheses and are given in # - 1. cm - I. 10- 3.
5.0 (2.81)
pH
A~
4.B
I
/="(2.54)
~--~-T----" ,-~-4.7
4.5 0
(2.40)
(2.56)
=
(2.6oi
~ O
(2.39)'~
0(2.43)
i
i
i
120
E40
360
4BO
Time ( min ) Fig. 6. Buffer renewal method: variations of pH and conductivities (I2-1. cm-}. 10S-3) with time in the compartment (~) for various buffer compositions in the compartment (~): ( ~ ) Cb('c] = 0.0282 M (value computed in agreement with our previous paper [1]); ( i ) CbtLE)= 0.0250 M; (A) Cb~ = 0.0233 M. Electric current: 5 A. The starting concentrations in compartment (~) are always the same, as in Table 1.
282 Three facts may explain this difference. (1) The small electric conductivity of the electrolyte in compartment (~) induces a heat evolution larger than in the other chambers. The physical constants, and thus the transport numbers are temperature dependent. At the temperature of this compartment, the transport numbers are not identical any more to the one prevailing in the others. (2) The numer, ical values for ionic mobilities as reported by various authors in the literature exhibit discrepancy. The values adopted here for the computation [15], are probably slightly different from those prevailing in the agarose membranes. (3) The absolute ionic mobilities are a function of the ionic strength of the solution. This last parameter is different in each compartment. The relative mobilities, assumed constant in the determination of the concentration profiles, could vary somewhat from one compartment to the other due to the different ionic strength in each of them. Fig. 6 shows the measured p H stability in compartment (~) as a function of time, after some adjustment of the computed concentration in compartment © has been made. It is to be noted that small concentration variations in compartment © result in the inversion of the slope of the residual p H shift in compartment (~) which ensures a much better p H stability. However, the remaining variations of conductivity show that the equality of transport numbers is still not completely reached. These p H stabilities have been compared with those obtained with a constant concentration of base in each compartment (namely C b = 40.8 raM). The results for these experimental conditions are shown in Fig. 7.
5.0 pH
(2.59) 4.8
(2.60)
i~2. 45 ) •
1"-1~_(2.18
4.6
, 45)
4.4
_
®
<~~ < ~
4.2
)
(2.28) ~ ~ <~...._ ~ )
(2°42) ~Z~--.... (2.40) A""A~A ~
O
...~,O ~ E ] ~ FI~E] ~ (2.49) (2,57) I
4.0
0
120
,
I
240
Q ,
I
,
I
360 480 Time ( min )
i
1
600
A,
720
Fig. 7. Buffer renewal method: variations of pH and conductivity (/2 1-cm-i-10-3) with time in the working compartments with a constant concentration of sodium acetate in the buffer solutions (Table 2: Cb = 40.82 raM). Electric current 5 A.
283 TABLE 2 BUFFER RENEWAL METHOD Circuits
@
(~)
(~)
@
4.30 96.60 40.82 61.2
4.55 54.29 40.82 52.0
4.75 34.25 40.82 43.2
4.90 24.24 40.82 36.4
0.0068 9.10 -9 0.4058 0.5874
0.0038 2.10 -8 0.4069 0.5893
0.0024 3.10 -8 0,4074 0,5902
0.0017 4.10 -8 0.4077 0.5906
(~)
Buffer compositions: pH Ca (mM) C b (mM) fl (mM)
4.15 136.50 40.82 65.5
Computed transport numbers: TH +
TOllT^ TNa +
0.0096 7-10 -9 0.4048 0.5856
fiN,+=1.00; IfiA 1=0.69; fi8+=8.15; IfioH-l=4.451 Experimental conditions for the stabilization of the pH gradient illustrated in Fig. 7. Compositions and flow rates of the solutions added in the end compartments; composition of the buffer solutions computed by keeping constant the concentration of sodium acetate (C b = 40.82 mM). I = 5 A; I/(~)= 22.56 m l / m i n CCH3COON, = 0.1215 M; I)(~) = 44.98 m l / m i n
CCH3COO H =
0.0932 M.
External recycling of the buffers between end compartments In the end compartments, the stabilization of the buffer compositions can be obtained by recycling the buffer from one compartment to the other at a flow rate
TABLE 3 B U F F E R RECYCLING METHOD Circuits
~)
(~
(~)
(~)
(~
4.40 105.7 57.45 79.8
4.80 16.76 21.34 21.9
5.20 2.66 8.09 4.8
5.65 50 500 127.2
0.003 7.10 -9 0.443 0.554
0.003 4.10 - s 0.443 0.554
0.003 3.10 -7 0.443 0.554
2.10 - 5 2.10 s 0.444 0.556
Buffer compositions: pH C. (mM) C b (mM) fl (mM)
4.00 667 158 241.1
Computed transport numbers: Tn + TollTA-
TNa+
0.003 1.10 -9 0.443 0.554
fiNa+ = 1.0; I fiA- I = 0.8; fill+ = 7.2; l fioH- I ~ 3.9 Experimental conditions for the stabilization of the pH gradient illustrated in Fig. 8. In compartments (~) to (~), the composition of the buffer solutions is computed to keep constant transport numbers (Trt+ = 0.003). In compartment (~), the buffer composition is computed to allow a stabilization of the p H gradient by the buffer recycling method [1]. I = 2 A; 1;"= 2 ml/min.
284 B.B
pH
6"2 f 5 B (21.8)
(21.0)
©
5 4 -
!-,4.1 .o/ s
o
_
I--i_
I
(0.72)
4 B -
i___.
(3.3) ~
(I .65) i _ (3.3)
A--A_A
~
A
42 _(7 9 0 o 3
B
,
o
,
@ @ @ @
/ n ~ l ~
,
I
t20
L ,
,
J
I
J
a4o Time ( min
i
1
360
I
I
I
d80
Fig. 8. External recyclingmethod: variations of pH and condactivity(J2-l. cm-1.10 3) with time. The compositions of the buffer solutions are given in Table 3. Electric current: 2 A. Recycling flow: 2.2 ml/min. established by a mass balance [1,13]. To verify the equations developed in ref. 13, the following experiment has been made: the pH of the five working compartments has been chosen and the composition of the solution of chambers (~) to (~) has been evaluated so as to keep the transport numbers constant. The composition of compartment C ) is thus fixed, but the transport numbers, as demonstrated in our previous paper [1], are not identical to those of the other compartments. The experimental conditions are given in Table 3, and the stability of pH profiles as well as the conductivities are illustrated in Fig. 8. According to the theory, the recycling flow should be 2 ml/min; experimentally, the stabilization has been reached with a slightly higher flow, 2.2 ml/min. However, even with this corrected value, the pH of compartment 3 shifts markedly and this for two reasons: (1) the buffering capacity of this compartment is the lowest (fl = 4.8 mM); (2) the transport numbers of ions in compartments (~) and ( ~ a r e not and cannot be identical [1]. However, if one reduces the pH range covered in one experiment, the stability can be increased.
Discussion The above reported results demonstrate the possibility of stabilizing a pH profile in acid medium with simple non-amphoteric buffers. In a multicompartment cell, a range of 0.75 pH unit can be stabilized for more than 10 h, with a current density of 96 rnA cm-2.
285 A criterion for the pH stability can be defined in each compartment as the inverse of the pH shift per Faraday and per unit of buffering capacity. With this criterion, the stabilities measured here are 4-20 times higher than the one given in the literature [8-10], where the condition of constant transport numbers is not satisfied. In acid medium, the pH stability of compartment (~) which is the most sensitive due to its low buffering capacity can be improved by a small adjustment of the composition of the next compartment ( ( ~ ) . This observation shows that the condition of equal ionic transport numbers is not entirely fulfilled. A possible explanation of this residual pH shift can be obtained from the experiment conducted with a constant concentration of base (Table 2). Since the ionic strength and the temperature in the cell were constant during this experiment (identical conductivity of the buffer solutions), the transport numbers can be computed for each compartment using equivalent ionic conductivities [15]. The values are reported in Table 2. With these calculated transport numbers, the mass balance over compartment (~) based on the migrational fluxes entering and leaving the compartment suggests that sodium ions are depleted, protons and acetate ions are accumulating and acetic acid is produced. Electroneutrality is, of course, maintained: the accumulation of protons is exactly balanced by the increase in sodium ions and the increase in acetate ions. This appears experimentally (Fig. 7) as a decrease in both pH and conductivity in compartments (1), (~) and (~) due to the formation of acetic acid and the depletion of sodium ions. The ionic transport numbers calculated using equivalent ionic conductivity allow thus to predict the variation of the buffer concentration in a compartment. The residual pH shift observed with concentration profiles which are supposed to keep a constant transport number (Fig. 5), may be explained by the same arguments. These concentration profiles imply profiles of both conductivity and ionic strength (0.01-0.1 M). Temperature profiles (10-20°C) between compartments (~) and © are then observed due to the heat exchanger which does not allow variable heat transfer rates from each compartment. These variations in ionic strength and temperature from one compartment to another cause slight differences of relative ionic mobilities between the compartments. Then ionic transport numbers (Table 1) are reevaluated: with these new values corresponding to the experimental conditions (ionic strength, temperature), the transport numbers of Na + and H + decrease from compartment (~) to compartment ( ~ by about 5%, while the transport numbers of acetate ion increase by 8%. The mass balance over the ionic species in the compartments allows the prediction of the increase in the concentration of sodium acetate and acetic acid. Effectively, an increase in the conductivity of the buffer solution contained in the central compartments is observed experimentally (Fig. 5). These observations indicate the necessity to limit the variations of the relative ionic mobilities to reach a better pH stability. It is possible to work only with small difference of buffer concentrations between adjacent compartments or using a buffer in which the ionic mobilities of the components vary in the same manner with the ionic strength. One could use the temperature of the buffers to adjust ionic mobilities.
286 It has been shown [13] that agarose membranes with cotton supports are not entirely neutral causing transport of water by electroosmosis from one compartment to another. The smaller the buffer concentration, the more important the water transport, which contributes also to the residual pH shift. Martin [9] proposes the utilization of amphoteric membranes to suppress electroosmosis. Such membranes have been prepared and their tests indicate that they can effectively reduce electroosmosis (inversion of water transport depending on the pH)~'However, they have not yet been tested in the isoelectric focusing cell. Finally, these experiments indicate the problems associated with the two methods of stabilizing buffer composition in the end compartments: (a) Using a fresh solution allows the stabilization of the composition and the p H of the buffer solutions as time progresses, even in the case of very dilute buffer concentrations (C b = 25 mM). The second experiment (Fig. 6) indicates how important it is to maintain the composition in the end compartments constant in order to avoid the shift in the p H and conductivity in the central compartments. This method requires large solution volumes, which means the smaller the buffer concentration and the larger the electric current, the larger the volumes required. (b) External recirculation of the buffers between the end compartments [9,17,18] allows the stabilization of the buffer composition in these compartments (Fig. 8), if the conditions given in the previous article [1] are fulfilled. Some problems of pH stabilization appear in the central compartments because the initial hypothesis that the transport numbers should remain constant between these compartments and the end compartments is not fulfilled experimentally.
Conclusions The installation of a multicompartment isoelectric focusing device has made possible the study of p H profiles stabilized with simple non-amphoteric buffers such as acetic acid and sodium acetate. The experiments show the importance of the constant ionic transport numbers, which requires accurate control of the concentration profiles while minimizing the variation of the relative ionic mobilities. Although other non-amphoteric buffers may be investigated in different pH ranges, the p H stability obtained with acetic acid/sodium acetate buffer over about 10 h is sufficient for the separation of amino acids in a multicompartment isoelectric focusing device. The next article [19] will describe experiments related to separations and will recommend parameters for the scale-up of multicompartment apparatus.
References 1 2 3 4
Wenger,P. and Javet, Ph. (1986) J, Biochem. Biophys. Methods 13, 259-273 Svensson,H. (1948) Adv. Protein Chem. 4, 251-295 Righetti, P.G. (1975) Sep. Purif. Methods 4, 23-72 Righetti, P.G. (1983) IsoelectricFocusing: Theory, Methodologyand Applications, Elsevier,Amsterdam
287 5 Rilbe, H. (1969) Prot. Biol. Fluids 17, 369 6 Rilbe, H., Forchheimer, A., Petterson, S. and Jonsson, M. (1975) in Progress in Isoelectric Focusing and Isotachophoresis (Righetti, P.G., ed.) pp. 51-63, North Holland Publ. Co., Amsterdam 7 Rilbe, H. (1978) J. Chromatogr. 159, 193-205 8 Jonsson, M. and Fredriksson, S. (1981) Electrophoresis 2, 193-203 9 Martin, A.J.P. and Hampson, F. (1978) J. Chromatogr. 159, 101-110 10 Martin, A.J.P. and Hampson, F. (1981) U.S. Patent No. 4'243'507' 11 Bier, M., Egen, N.B., Allgyer, T.T., Twitty, G.E. and Mosher, R.A. (1979) in Peptides: Structure and Biological Function (Gross, E. and Meienhofer, J., eds.) pp. 79-89, Pierce, Rockford, IL, USA 12 Bier, M. and Egen, N.B. (1979) in Electrofocus '78 (Haglund, H., Westerfeld, J.-G. and Ball, J.-T., eds.) pp. 35-48, Elsevier, Amsterdam 13 Wenger, P. (1985) Ph.D. Thesis No 604, EPFL, Lausanne, Switzerland 14 Kielland, J. (1937) J. Am. Chem. Soc. 59, 1675 15 Dobos, D. (1975) Electrochemical Data, Elsevier, Amsterdam 16 Martell, A.E. and Smith, R.M. (1974) Critical Stability Constants, Plenum Press, New York 17 Rilbe, H. (1981) Electrophoresis 2, 261-272 18 Rilbe, H. (1982) Electrophoresis 3, 332-336 19 Wenger, P. and Javet, Ph. (1986) J. Biochem. Biophys. Methods 13, 289-303