Focusing of pepsin in strongly acidic immunobilized pH gradients

Journal of Biochemical and Biophysical Methods, 16 (1988) 185-192 Elsevier

185

BBM 00667

Focusing of pepsin in strongly acidic immobilized pH gradients Pier Giorgio Righetti, Marcella Chiari, Pranav K. Sinha 1 and Enzo Santaniello 2 Department of Biomedical Sciences and Technologies, University of Milan, Via Celoria 2, Milano 20133, Italy, 1 Institute of Clinical Chemistry and Biochemistry, Freie Universitiit Berlin, Klinikura Virchow, Spandauer Datum 130, D-I O00 Berlin 19, F.IL G. and 2 Dipartimento Chimica e Biochimica Medica, Universit& di Milano, Via Saldini 50, Milano 20133, Italy

(Received 7 March 1988) (Accepted 21 March 1988)

Summary A new acrylamido buffer has been synthesized, for use in isoeleetric focusing in immobilized pH gradients. This compound (2-acrylamido glycolic acid) has a pK = 3.1 (at 25 o C, 20 mM concentration during titration) and is used, by titration with the pK 9.3 Immob'tline, to produce a linear pH gradient in the pH 2.5-3.5 interval. Pepsin (from pig stomach) focused in this acidic pH gradient is resolved into four components, two major (with pI values 2.76 and 2.78) and two minor (having p I values 2.89 and 2.90). This is the first time that such strongly acidic proteins could be focused in an immobilized pH gradient. Even in conventional isoelectric focusing in amphoterie buffers it has been impossible to focus reproducibly very-low-p/macromoleeules. Key words: Immob'flized pH gradient; Isoelectric focusing; Acidic pH range; Pepsin

Introduction Isoelectric focusing (IEF) in immobilized pH gradients (IPGs) has proven to be the most versatile electrokinetic separation technique today available, definitely exhibiting an extremely high resolving power and able to support large protein loads in preparative separations [1-3]. Originally, it was believed that IPGs would be operative only in the pH 4-1.0 range, as it was supposed that outside these limits IPGs would simply fail, due to substantial differences in conductivity between the Correspondence address: Professor P.G. Righetti, University of Milan, Via Celoria 2, Milano 20133, Italy. 0165-022X/88/$03.50 © 1988 Elsevier Science Publishers B.V. (Biomedical Division)

186 Immobiline matrix and the bulk liquid [4]. In reality, in 1985, we were able to focus dansylated amino acids in a pH 3-4 IPG gradient by resorting to the use of 'conductivity quenchers', i.e. by incorporating in the IPG matrix a density gradient with the dense region located at the anodic side, so as to prevent the gel from drying around pH 3 (by electrosmosis due to its net negative charge) and to smooth the voltage gradient across the gel for sharper band focusing [5,6]. At that time it was understood that a new 'Immobiline' had to be added to our equations for pH gradient calculations: water, which would behave as a basic species with pK = 0.00 (when taken at unit molarity) or with pK = -1.74 (when expressed in its standard concentration of 55.56 M). Symmetrically, at the opposite extreme of the pH scale, water would behave as a basic species with pK = 14 (or 15.74 when expressed in its standard 55.56 molarity value) [7]. Concomitant with the higher conductivity of bulk water around pH 3 (and, in alkaline regions, around pH 11) its buffering capacity also becomes appreciable [8] and the strategy used to counteract it was to increase the Immobiline concentration in the gel by a factor of 2 or 3 as compared with their standard 10 mM level. More recently, we were able to extend also the fractionation range at the alkaline extreme, by creating IPG pH 10-11 intervals with the aid of a new Immobiline having pK = 10.3 (at 10 o C) [9]. Notwithstanding the hydrophobicity of alkaline pH ranges [10], non-hydrophobic proteins could be resolved into sharp zones, with retention of enzyme activity [11]. In the present report, we describe the synthesis of a new acidic acrylamido buffer, having pK = 3.1, and its use for creating strongly acidic (pH 2.5-3.5) IPG gradients. For the first time since the existence of IEF, good focusing patterns of pepsin were obtained.

Materials and Methods

Crystalline pepsin from porcine stomach mucosa was from Sigma, St. Louis, MO. Glyoxylic acid was from Aldrich-Chemie, Steinheim, FGR. Acrylamide, N,N'methylene bisacrylamide (Bis), TEMED, persulphate and Coomassie brilliant blue R-250 were from Bio-Rad Labs., Richmond, CA. pK 9.3 Immobiline, the Multiphor 2 chamber , Multitemp thermostat and Macrodrive power pack were from LKB Produkter AB, Bromma, Sweden.

Synthesis of the pK 3.1 acrylamido buffer The new 'Immobiline'-type buffer, 2-acrylamido glycolic acid, was synthesized from acrylamide and glyoxylic acid essentially as described by Schouteeten et al. [12]. The synthetic product was crystallized four times from water and, by elemental analysis and thin-layer chromatography, was found to be 98% pure. The compound crystallized as monohydrate and its molecular mass is 163 Da. All chemico-physical characteristics were in accordance with its structure. Melting point: 95-97°C for the monohydrate. Analysis: calculated for CsHTNO4 • H20: C, 36.8; H, 5.5; N, 8.6%; found: C, 37.0; H, 5.7; N, 8.8%.

187

Potentiometric titration The pK of the carboxyl group was determined with a Radiometer automatic titrimeter at 25"C in a 20 mM solution titrated with 50 mM NaOH. The data obtained were also simulated by a new computer program able to calculate pH courses, ionic strength and buffering power of any mixture of monoprotic or polyprotic buffers and titrants [13,14]. IPG p H 2.5-3.5 gradient The recipe for this gradient is given in Table 1. The gel contains a matrix of T% = 4, C% = 4 and an average of 20 mM pK 3.1 'Immobiline' titrated in the pH 2.5-3.5 interval with the pK 9.3 commercial Immobiline (this latter compound being dissolved in anhydrous propanol, in order to prevent hydrolysis and autopolymerization). The separation gel (17.5 cm long) was cast along the long side of the LKB polymerization cassette, with or without a 3 cm long pH 5 plateau for sample loading [15] and was 0.5 mm thick. After addition of TEMED to the pH 2.5 and pH 3.5 limiting solutions, they were titrated around neutrality by addition of solid Tris, so as to allow for correct polymerization conditions and prevent formation of strongly alkaline boundaries which could be generated if NaOH were added instead of Tris [16]. The gel formulation also contained 1% Ampholine in the pH 2.5-4 range. During gel casting, the pH gradient was stabilized by a 0-30% (cathode to anode) sorbitol density gradient. After gel polymerization (1 h at 50 ° C) [17], the matrix was used as such, without washing, and the pepsin sample (30 /~g/track) applied at the cathode end in pockets precast in the gel. In this way, the density gradient present in the gel acts as 'conductivity quencher' and prevents burning at the anode end. Running conditions: 4 h at 300 V followed by overnight at 500 V, 10 ° C. Staining was with Coomassie blue R-250 in the presence of copper sulphate [18].

Results

Titration of the new acrylamido buffer A 20 mM solution of 2-acrylamido glycolic acid (AGA) was titrated with a Radiometer (Copenhagen, Denmark) titrimeter with 20 mM NaOH (Fig. 1). As the starting pH of the solution was 2.5, it was difficult to measure directly the pK from the curve, since only a portion of the titration curve, around the presumptive pK value, was available. The pK value was inferred from the following data: (a) by accurate measurement of the pH of the solution when ½ equivalents of titrants were added. According to the Henderson-Hasselbach equation, when the solution is 20 mM in AGA and 10 mM in NaOH, AGA will be 50% dissociated and thus the pH prevailing in solution should be the pK value. This value (over 8 determinations) was found to be 3.1 + 0.05; (b) by accurately measuring the inflection point corresponding to the end of the titration (see arrow in Fig. 1). These two values experimentally derived were used in our computer program [13,14] to generate a theoretical titration curve (Fig. 2). In this last case, we could also simulate the

188

pH 2- acrytomidoglycolJc acid

11

/

-

~

/

/ ml 20 mM NoOH

Fig. 1. Titration of the carboxyl group of AGA. 5 ml of 20 mM 2-acrylamido glycolic acid were titrated with a 20 mM solution of NaOH in a thermostated vessel (25 ° C) with a Radiometer titrimeter. The pK was determined as the pH of a solution containing a 2 : 1 molar ratio buffer/titrant and was found to be 3.1 ±0.05 (average of 8 determinations). The last arrow on the fight marks the position of the inflection point at the end of the titration.

#

pH ~

DoDD

0

oOoOonoOO%o

B Do

DO o

DD

DDo

oDD Do DD

7

-F"F

DD,Djr~F,Jr'F$

]DDDDDDDD

$$

6;

.F.FF'~ '°o 'o,

.b.F.F

J6~~

0

•

,i

5

....

i ....

10

I

~F,'F"FF*

i ....

15

pH

i ....

20

i . . . .

25

i ....

30 Tube

i , , , , i

35

~]

50

_

....

,40

L ....

45

i ....

50

[,,,,~1

55

~TO

Fig. 2. Computer simulation of the titration of 2-acrylamido glycolic acid. A 20 mM solution of AGA was titrated in the pH 2.0-8.0 interval by taking 60 fractions and computing the pH course, buffering power (fl) and ionic strength (I). Note that a fl maximum is found at the presumptive pK value (3.1), notwithstanding the contribution of water to the fl power. The matching of the experimental (Fig. 1) and theoretical pH curves is judged by the position of the inflection point at the end of the titration, which is pH 4 . 2 i n both cases.

189 TABLE 1 RECIPE FOR A pH 2.5-3.5 IPG G R A D I E N T * Additive

Acidic chamber

Basic chambel"

p K 3.1 p K 9.3 T% = 30, C $ = 4 Sorbitol Ampholyte 2.5-4.5 Pharmalyte 2.5-5 TEMED Persulphate (40%)

800 #1 145 pl 1.3 ml 2.4 g 200 #1 200 #1 5 #1 7pI

800 pl 350 #1 1.3 ml 200 #1 200 pl 5 #1 7/t I

* All Immobilines, 0.2 M solutions; final volume in each chamber: 7 ml.

courses of accompanying buffeting power (fl) and ionic strength ( I ) values. The matching of theoretical and experimental curves was judged by the overlapping of the end point of the titration. Considering the very good matching of the two curves (inflection point of the end point of titration at pH 4.2 in both cases) and considering that our computer program automatically calculates the contribution of the hydrolytic products of water (see the fl power curve) we feel that the value of 3.1 for the pK of the carboxyl of AGA is an accurate measurement. Generation of a p H 2.5-3.5 IPG interval Table 1 gives the molarities of buffering ion (AGA) and titrant (pK 9.3 Immobiline)for generating a pH 2.5-3.5 IPG range. Fig. 3 shows the simulated pH 4 pN

I ×

B

x 13.

"'0~

x" 0-_

0-- ~ C---

~

.×

pH

X

[]~~~°--°'-O''O-.O. o~ .X' 3

~o." O."'0, i

X"

""I::1, \\O O

X'"

""0

X" . X""

", # 0

• X"

O'

'

•

10

,

'~

x '

20

/

X" X"'"

..X X"

15

'

'

5'

'

'

'

'

10

J

~

~

1A5

J

t

J

~

20

5

I

Tube ::14:

Fig. 3. Computer simulation of the physico-chemical parameters of a p H 2.5-3.5 gradient utilizing AOA as buffering group. A 20 mM solution of A G A was titrated in the p H 2.5-3.5 interval with the p K 9.3 Immobiline as titrant. The computer automatically calculates the pH, fl power and ionic strength of the 20 fractions eluted. Experimentally, a gradient was prepared with a 2-vessel gradient mixer and 11, 1 ml fractions collected. The pH of the experimental gradient is reported with solid triangles.

A

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

- pH 3.5

B pl 2.90 pl 2.89 2.90 2.89

pl 2.78 pl 2.76

:~,pe~n~il .......................

2.78 2.78

.... pH 2.5

~ ~

~. . . . . . . . . . . pepsin

iiiiii!iii!!!i!!!i!!!ii!iiiii!! ¸I

Fig. 4. IEF of pepsin. A T ~ = 4, C ~ = 4 I P G matrix in the p H 2.5-3.5 interval was made, containing 20 m M A G A as buffering ion and the p K 9.3 Immobiline as titrant. The gel was polymerized in the presence of 1~ Ampholine p H 2.5-4 and was 0.5 m m thick and 18 cm long. 30 # g of pepsin from pig stomach mucosa were loaded in pockets precast at the cathode. The gel was not pre-focused. R u n n i n g conditions: 4 h at 300 V followed by overnight at 500 V, 10 o C. Staining with Coomassie brilliant blue R-250 in copper sulphate. The p I values were derived by linear interpolation from the gradient slope of Fig. 3. (A) Picture of the entire gel over the 17.5 cm separation axis; (B) close-up of the 4 focused pepsin bands.

191 range obtained with the values given in Table 1. The gradient linearity is very good, with a deviation only in the last four points at the alkaline end. In order to check experimentally the shape of the gradient, the two limiting solutions were eluted from the two-vessel gradient mixer (in the absence of catalysts and avoiding neutraliTation) and collected in fractions: the pH gradient thus obtained was superimposable to the theoretical profile (Fig. 3, triangles).

Focusing of pepsin Pepsin is one of the most acidic protein known. Its pI value [19] is generally given as < pH 3. In the IPG pH 2.5-3.5 interval obtained as above, pepsin was focused by cathodic application in a pH 5 plateau. In reality, as shown in Fig. 4, identical results were obtained when pepsin was loaded directly at the extreme cathode end of the gel (pH ca. 3.3), in the absence of pH plateaus. Pepsin is resolved into four zones: two major ones, having p l values (at 10°C) of 2.76 (the most abundant) and of 2.78, and two minor bands, of about equal intensity, exhibiting p I values of 2.89 and 2.90. These bands, in control runs, were cut out, eluted and were found to possess proteolytic activity on a casein substrate [20] (not shown), while the heavy zone migrating upwards and focusing around p l 3.4 was found to be devoided of activity, suggesting the presence of either degradation products of pepsin or impurities in the preparation. Discussion

While there is a general agreement that pepsin is a strongly acidic protein, reports on its pI value vary widely in the literature. Thus, while De Koning and Draaisma [21] attribute a p ! of ca. 1 to bovine and swine pepsin, two isoforms are reported from hog gastric juice [19], a p I 2.86 (major) and a p l 2.94 (minor) species. In the case of bovine pepsin, Righetti et al. [20] report 3 major isoforms, with pI values 2.80, 2.90 and 2.99 and a minor component, with p I 3.09. The same authors [22] report for pig pepsin a major component at p l 3.02 and a minor one at p l 3.2. Also for human pepsin, p I values around pH 3 are generally given [23]. Thus, there is substantial agreement that pepsin, from different sources, should have p l values in the pH 2.5-3.0 range and should be polymorphic, exhibiting a spectrum of 2-4 enzyme-active bands. Our results fully confirm and validate previous literature reports on the range of p l values and on pepsin polymorphism. In particular, the set of four isoforms of pepsin from porcine stomach mucosa, here reported, closely resembles the p! spectrum and band distribution of bovine pepsin. A major breakthrough of the present research is the demonstration of the possibility of creating, in IPGs, such acidic pH gradients as a pH 2.5-3.5 interval. As stated in the introduction, at the beginning it was believed that IPGs would be restricted to pH 4-10 intervals. Subsequently, it was shown that it was possible to use IPGs also in the pH 3-4 interval, but no components with p I values below pH 3.2 were ever focused. The synthesis of this new Immobiline with a pK of 3.1 has opened the possibility of reaching such very low pH values as pH 2.5. Thus, today the know-how and the potential exist for creating IPG ranges spanning a maximum

192

interval of pH 2.5-11, which should permit separation and analysis of more than 99.5% of all potential phenotypes in any cell type. In particular, we are planning to extend the use of this technique to the analysis and fractionation of lysosomal membrane proteins, reported to have very acidic pI values.

Simplified description of the method A new method is reported for focusing strongly acidic proteins (e.g. acidic proteases, like pepsin). The method utilizes immobilized pH gradients in the pH 2.5-3.5 interval and has been made possible by the synthesis of a new acidic 'Immobiline'-type buffer, having a pK = 3.1. For successful fractionation, the sample is loaded in a midly acidic pH plateau (pH 5) at the cathode gel extremity and the gel contains a sorbitol density gradient (0-30%, dense part at the anode) so as the prevent drying and burning of the acidic gel extremity due to electrosmotic flow towards the cathode. To our knowledge, this is the first time that the synthesis of an 'Immobfline'-type buffer is reported in the scientific literature. Immobilized pH gradients have thus extended their fraetionation capability over the entire pH 2.5-11 range.

Acknowledgements Supported in part by Progetto Finalizzato Biotecnologie e Biosensori, CNR, Roma, and by Ministero della Pubblica Istruzione. We thank Mr. C. Tonani for help with the computer simulations.

References 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Righetti, P.G. (1984) J. Chromatogr. 300, 165-223 Righetti, P.G. and Gianazza, E. (1987) Methods Biochem. Anal. 32, 215-278 Righetti, P.G. (1986) Trends Anal. Chem. 5, 16-20 G-ianazza, E. Celentano, F., Dossi, G., Bjellqvist, B. and Righetti, P.G. (1986) Eleetrophoresis 7, 128-133 Bianehi-Bosisio, A., Righetti, P.G., Egen, N.B. and Bier, M. (1986) Electrophoresis 7, 128-133 Mosher, R.A., Bier, M. and Righetti, P.G. (1986) Eleetrophoresis 7, 59-66 Righetti, P.G., Gianazza, E. and Celentano, F. (1986) J. Chromatogr. 356, 9-14 Righetti, P.G., Gianazza, E., Gelfi, C. and Siaha, P.K. (1986) in: M.J. Dunn (Ed.), Electrophoresis '86, VCH, Weinheim, pp. 419-434 Gelfi, C., Bossi, M.L., Bjellqvist, B. and Righetti, (1987) J. Biochem. Biophys. Methods 15, 41-48 Rabilloud, T., Gelfi, C., Bossi, M.L. and Righetti, P.G. (1987) Electrophoresis 8, 305-312 Sinha, P.K. and Righetti, P.G. (1987) J. Biochem. Biophys. Methods 15, 199-206 Schouteeten, A., Christidis, Y. and Mattioda, G. (1978) Bull. Soc. China. France II, 248-254 Celentano, F.C., Tonani, C., Fazio, M., Gianazza, E. and Righetti, P.G. (1988) J. Bioehem. Biophys. Methods 16, 109-128 Righetti, P.G., Fazio, M., Tonani, C., Gianazza, E. and Celentano, F.C. (1988) J. Biochem. Biophys. Methods 16, 129-140 Ek, K., Bjellqvist, B. and Righetti, P.G. (1983) J. Biochem. Biophys. Methods 8, 134-155 Righetti, P.G., Chiari M. and Gelfi, C. (1988) Electrophoresis 9, 66-76 Righetti, P.G., Ek, K. and Bjellqvist, B. (1984) J. Chromatogr. 291, 31-42 Righetti, P.G. and Drysdale, J.W. (1974) J. Chromatogr. 98, 271-321 Righetti, P.G. and Caravaggio (1976) J. Chromatogr. 127, 1-28 Righetti, P.G., Molinari, B.M. and Molinari, G. (1977) J. Dairy Res. 44, 69-72 De Koning, P.J. and Draaisma, J.T.M. (1973) Neth. Milk Dairy Res. 27, 368-378 Righetti, P.G., Tudor, G. and Ek, K. (1981) J. Chromatogr 220, 115-194 Vesterberg, O. (1973) Acta Chem. Scand 27, 2415-2417