Treatment of dialysis membranes for simultaneous dialysis and concentration

ANALYTICAL

BIOCHEMISTRY

145, 343-350 (1985)

Treatment

of Dialysis Membranes for Simultaneous Dialysis and Concentration’

VIRGINIA L. RICHMOND, ROSANNA ST. DENIS, AND EDITH COHEN Pac$c Northwest Research Foundation. 1102 Columbia Street, Seattle, Washington 98104 Received August 30, 1984 Dialysis membranes used for simultaneous dialysis-concentration required pretreatment to remove uv-absorbing compounds leached from the membranes and to reduce the absorption of protein to the membranes. This was accomplished with sodium carbonate and ethanol or with “sulfur-removal solutions.” Protein determinations were made with a micro-Bradford protein reaction and with uv absorbance at 280 nm. Soluble membrane components contributed to aberrant uv spectra and altered the ratio of 280/260-nm absorbance. Simultaneous dialysis and concentration in the micro protein dialyzer-concentrator apparatus, combining aspects of thin-layer dialysis and ultrafiltration, resulted in rapid removal of salts from the protein solutions. Prior treatment of membranes reduced uncertainties in retentate recoveries, eliminated uv-absorbing components of membranes, and improved recoveries of protein. o 1985 Academic Press, Inc. KEY WORDS: dialysis: protein/enzyme purification: protein determination; protein concentration; ultrafiltration.

Dialysis of protein samples in conventional membranes has been an essential procedure for many biochemical preparations in transferring proteins or other macromolecules to different buffers and in removing small-molecular-weight components (1). Conventional dialysis usually results in the same as or in a more dilute concentration of macromolecules than in the original solution, requiring lyophilization or pervaporation to obtain adequate solute concentration for subsequent analyses, which inactivates some proteins (2). An apparatus which simultaneously dialyzes (1,3) and concentrates (4) macromolecules in solution has the advantages of reducing the handling and time required to accomplish this two-step procedure while retaining the macromolecules in solution (5). In preliminary experiments with both manually assembled and preassembled membranes we experienced loss of protein from dilute solutions, increased absorbance in the ’ The authors thank the Council for Tobacco Research, Inc., USA, Grant 145 1. for partial support of this work. 343

uv range, and skewing of the uv-absorbance spectrum of the protein solution. Prior treatment of the membranes was necessary to quantitatively recover the protein retentate and to eliminate soluble membrane components from the retentate. The paper describes membrane treatment procedures to improve protein recovery using both the Bradford reaction (6) and absorbance at 280 nm for measurements, recovery of ‘251-bovine serum albumin, and uv spectra of retentates to verify decreased contribution of membrane soluble components. MATERIALS

AND

METHODS

Bovine serum albumin (BSA),* Cohn Fraction V (96%) was obtained from Sigma Chemical Company. Sodium carbonate (AR), sodium azide (reagent grade), and ammonium acetate (reagent crystals) were obtained from Matheson, Coleman and Bell; sodium bicarbonate was Arm and Hammer brand, and gelatin was from Knox Gelatin, Inc. “Sulfur’ Abbreviation used: BSA, bovine serum albumin. 0003-2697185 $3.00 Copyright Q 1985 by Academic Press. Inc. All rights of reprcduction in any form reserved.

344

RICHMOND,

ST. DENIS, AND COHEN

removal solutions” A and B were obtained as a kit from Spectrum Medical Industries, Los Angeles; potassium sorbate (2,4-hexadienoic acid, potassium salt) MCB, No. SX 0845-3, was a gift from M. W. Cook, BioMolecular Dynamics, Beaverton, Oregon. Visking dialyzer tubing, Size 27, was from VWR Scientific, Inc.; disodium ethylenediaminetetracetic acid (EDTA) reagent was obtained from J. T. Baker Chemical Company; and TSK-HW-55S, superfine fractogel, was obtained from E. Merck, Darmstadt, West Germany. All glassware was treated with siliconizing fluid (Surfasil), Pierce Chemical Company, as a 2% solution in acetone, air dried for 24 h, and washed gently before use. Following its use for the Bradford reaction, glassware subsequently rinsed with 95% ethanol and then water gave consistent results for at least 6 months. Vials were siliconized to obtain reliable results when concentrations down to 0.625 pg/ml BSA were made for BSA standards. Spectra determinations were made with lml cells with a l-cm light path in a Beckman Model 35 spectrophotometer versus the dialyzing buffer or appropriate blank. Bradford protein determinations (6) were made using either a 20% solution of Bio-Rad Bradford reagent or a 50% solution of Pierce protein assay reagent in water. The unknown sample, 50 ~1, was pipetted into a siliconized test tube, 13 X 100 mm, and 1.O ml of diluted Bradford reagent was added and swirled gently to mix. The absorbance was determined versus a blank containing buffer at exactly 5 min postmixing at 595 nm (6). A standard curve with bovine serum albumin was made for each Bradford reagent and standards run with each set. Slopes obtained for each reagent were essentially the same. Standard BSA solution was diluted in dialysis buffer to give concentrations from 0.625 to 25.0 @g/50 ~1 by adjusting the absorbance of each solution to the proper quantity based on A’% = 6.9 at 279 nm (7).

Standard BSA solutions for dialyses were either 0.4 or 4.0 mg/ml in 5 mM ammonium acetate, pH 8.0, and were adjusted as described above to have an absorbance at 280 nm of 0.276 or 2.76/ml, respectively (7). The dialysis buffer (5 mM ammonium acetate) and 0.01% sodium azide, pH 8.0, used for filling the sample reservoir cups or for BSA dilution was filtered through 0.22 pm Millipore filters (Millipore Corp.). The buffer solution for the 5-liter container was the same composition and pH but was not filtered. The ratio of absorbance at 280/260 nm was determined (8) to detect some of the increased absorbance below 280 nm, which was contributed by soluble components from the membranes. The Model 320 Micro-Pro-Di-Concentrator, Bio-Molecular Dynamics, Beaverton, Oregon, with the three-membrane-place lid, had a 5.5-liter chamber capacity. Dialysis membranes (Bio-Molecular Dynamics, Inc.) had a lO,OOO-MW cutoff and were used with either the lOOO- or the 50-~1 sample collection reservoir and appropriate sample collection rod, for either the manual assembly or the preassembled membranes. Membranes were shipped in a solution containing 1.5% potassium sorbate, 1% glycerine, 0.1% sulfur compounds, and 0.05% sodium azide. Membranes were transferred to the laboratory storage solution of 1% glycerine and 0.05% sodium azide, or left in the shipping solution at 5°C. Treatment

of Membranes

Membranes were treated with the following procedures, rinsed well in distilled water, and either used immediately or immersed in the storage solution and maintained at 5°C until they were rinsed well and used. “Membranes” indicates the preassembled dialysis membranes with collar and sample collection reservoir attached as obtained from the manufacturer. “Manual assembly” indicates the membranes obtained from the

TREATMENT

OF

DIALYSIS

manufacturer but requiring assembly in the laboratory with the collar and sample collection reservoir. “Conventional” dialysis was with Visking No. 27 dialysis tubing. Membranes were partially stabilized by insertion of the Pi rod, filled with the indicated solution and placed in 20 X 200-mm test tubes with either a grommet or an adapter at the top to hold the membrane upright. Procedure A. Membranes were soaked in two changes of glass-distilled water or 5 IIIM ammonium acetate buffer, pH 8.0, at room temperature for 15 min each. Procedure B. Membranes were heated in distilled water at 100°C for 30 min, then 0.05 mg/ml gelatin was dialyzed through the membrane overnight. The membranes were rinsed well with ammonium acetate buffer. Procedure C. Membranes were boiled twice in 0.1 M sodium bicarbonate, pH 8.3, containing 0.5% gelatin for 15 min, then in distilled water, 100°C for 15 min, and held in two changes of 50°C water for 15 min each. Procedure D. Conventional No. 27 Visking dialysis tubing was treated with 0.05% sodium carbonate, pH 1 I, four times for 30 min at 60°C rinsed with distilled water, washed with 95% ethanol for 30 min at 50°C and rinsed well with distilled water. Procedure E. Manual assembly membranes were washed in sodium carbonate-ethanol as for Procedure D. Procedure F. Membranes were washed with sulfur-removal solutions A and B (pH 3.0) at 80 and 60°C respectively, as described in the kit directions. Procedure G. Membranes, supplied by the manufacturer, were packed in a shipping solution of 1% glycerine and 0.05% sodium azide, with no potassium sorbate or other additives. These were soaked in two changes of ammonium acetate buffer for 15 min at room temperature. “‘1-BSA was prepared from a 1 mg/ml solution of BSA and lz51using Iodogen (Pierce Chemical Co.) according to the manufactur-

MEMBRANES

345

er’s directions. The ‘251-BSA was freed of reactants by chromatography through a disposable column of Sephadex G-25, 0.5 X 5.0 cm, in PBS. ‘*‘I-BSA was further purified by gel filtration through a 0.9 X 30-cm column of TSK-HW-55s fractogel in 5 mrvt ammonium acetate buffer to remove BSA components smaller than 68,000 M,. This stock of 12’I-BSA, approximately 22,800 cpm, was diluted with chromatographed cold BSA and buffer to give 2000 cpm corrected for background and an absorbance of 0.276/ml at 280 nm (0.4 mg/ml of BSA). Gamma counting was made in Pica “hangin” vials (Packard Instrument Co.) for 5 min with the preset program No. 1, for ‘25-I, in a Beckman 7000 gamma counter. The counts per minute were corrected for background and then corrected for decay back to the date of initial dilution. Dialysis

Procedures

Membranes were set up as described by the manufacturer using approximately 5 liters of 5 mM ammonium acetate buffer with 0.01% sodium azide, pH 8.0. The MPDC apparatus was attached to a water aspirator with reduction in pressure taking place over a 15-min period. After 45 min aspiration, or until bubbles no longer formed, with the solution stirred by a magnetic bar, the aspirator outlet was clamped off. The sample to be dialyzed was added to the sample reservoir cup with approximately 40 ml of buffer. The open sample cup was covered with a small plastic bag and dialysis-concentration was allowed to proceed overnight at 5°C or until all of the solution was below the top of the sample collection reservoir at the bottom of the membrane. After dialysis, the rod positioning band, sample reservoir cup, and sample rod were removed. The now concentrated solution was removed from the sample collection reservoir using a 9-in Pasteur pippet with 0.5 cm of Tygon tubing at the tip. After several tests of

346

RICHMOND,

ST. DENIS, AND COHEN

1.0

‘.O1

0.9

0.9

0.8

0.8

0.7

0" 0.7

0.6

f ; 0.6

it? 2 0.5

2< 0.5

8 f

z

0.4

DMysl* Number 10 20 30

.A 5. 6*

0.4

0.3 0.2 0.1

2

I

240

250

260

270 nm

280

290

:

D

0.7 6 0.6 2 m a 0.5 0 2 < 0.4

240

250

260

270 "ill

280

290

3

FIG. 1. Spectra of retentates dialyzed overnight, 5”C, in 5 liters of 5 IIIM ammonium acetate and 0.01% sodium azide, pH 8.0. The membranes were MPDC, either 50 or 1000 ~1 reservoir cup size, with lO,OOOMW cutoff, or, as indicated, conventional dialysis membrane No. 27, with lO,OOO-MW cutoff. The spectra determinations for 4.0 mg BSA were made after 10X dilutions in dialysate buffer. A-G represent the treatments as described in Table 1. The sequence of dialysis and spectrum of undialyzed BSA are shown in the inset.

removing only 1.O ml of sample and washings of the membrane, from a lOOO-~1 reservoir, improved recoveries were obtained by wash-

ing the sides of the membrane with buffer and collecting the retentate in 2.0 ml final volume. Another l.O-ml buffer wash was

TREATMENT

OF

DIALYSIS

347

MEMBRANES

0.9 0.6

0.7 go.6 f LF 0.5 0 k? < 0.4

zO.6 2 2 0.5 5 3 0.4

0.3

0.3

0.2 0.' 240

250

260

270

260

290

300

2

250

260

270 280 "n-l

290

:

G

,

240

250

260 FIG.

made to verify complete retentate.

removal

270 nm

290

3

l-Continued.

of the

RESULTS

Spectra and recoveries of BSA were improved by pretreatment of membranes before

dialysis versus nontreatment. The results are given in Fig. 1 and in Table 1. Treatment of conventional dialysis tubing with 0.05% sodium carbonate and water, and then with 95% ethanol, has been consistently used in this laboratory to remove previously observed uv-absorbing material and to obtain

348

RICHMOND,

ST. DENIS, AND COHEN TABLE 1 Apparent percentage recovery ’

Dialysis No.

Retentate

Bradford reaction

Absorbance (280 nm)

Ratio AU b 280/260 nm

A. Preassembled, water, 15 min at 23°C

1 2 3

5 mM NH,OAcC 0.4 mg BSA’ 0.4 mg BSA

3 94 97

132 112 100

0.41 0.64 1.05

B. Preassembled, water, 30 min at IOO’C, then 0.05% gelatin dialyzed overnight

1 2 3 4 5 6

0.4 mg BSA 0.4 mg BSA 0.4 mg BSA 0.4 mg BSA 5 mM NH40Ac 5 mM NH,,OAc

84 90 93 79 32 14

106 115 151 201 127 95

1.26 1.30 1.14 0.97 0.85 0.87

C. Preassembled, 0.1 M NaHC03 + 0.5% gelatin 2 X 15 min at lOO”C, 1 X 15 min in water, 2 X 15 min at’5O”C in water

1 2 3

0.4 mg BSA 0.4 mg BSA 5 mM NH40Ac

59 97 4.5

100 91 17

0.98 1.18 0.73

D. Conventional membrane, 0.05% Na*CO, 2 X 30 min at 6O”C, 95% ethanol, 1 X 30 min at 60°C

1 2

0.4 mg BSA 5 mM NH,OAc

103 7.2

100 10

1.32 0.64

E. Manually assembled, 0.05% Na2C03, 2 X 30 min at 6O”C, 95% ethanol, 1 X 15 min at 50°C

1 2 3 4

0.4 0.4 0.4 0.4

mg mg mg mg

BSA BSA BSA BSA

86 104 86

82 114 81 43

1.53 1.44 1.60 1.43

F. Preassembled, sulfur-removal solutions

1 2 3 4

0.4 0.4 0.4 0.4

mg mg mg mg

BSA BSA BSA BSA

101 94 84

95 70 79 91

1.49 1.40 1.20 1.68

1 2 3 4 5 6 7 8

0.4 0.4 0.4 0.4 0.4 0.4 0.4 0.4

mg mg mg mg mg mg mg mg

“‘I-BSAe BSA “‘1-BSA “‘1-BSA lZ51-BSA BSA BSA BSA

98 100 121 95 111 103 108 107

117 133 160 76 115 114 102 131

1.30 1.25 1.22 1.38 1.32 1.27 1.21 1.22

Membrane and treatment

G. Preassembled,“ washed 5X in 5 NH40Ac buffer, pH 8.0

mM

‘Apparent percentage recovery of protein was calculated from the standard slope obtained for the Bradford reaction and for absorbance of BSA before dialysis. b The ratio of absorbances at 280 and 260 nm for the standard BSA was 1.56. c The standard BSA was adjusted to an absorbance of 0.276/m] at 280 nm using the value of A% om = 6.9 (7). Standard BSA containing 4.0 mg/ml was determined from a 1:lO dilution. NH,OAc is the abbreviation for ammonium acetate. d Preassembled membranes for Procedure G were sent by the vendor in solution containing only 1% glycerol and 0.05% sodium azide. e “‘1-BSA was made and diluted as described under Materials and Methods.

TREATMENT

OF

DIALYSIS

complete recoveries of proteins which absorb to unwashed membranes (Richmond, unpublished results). Application of the sodium carbonateethanol washing Procedure E (Table 1, E and Fig. 1E) to membranes which were manually assembled resulted in improved agreement of the Bradford reaction and absorbance at 280 nm for percentage protein recovery, in improved spectra, and in near theoretical ratios for 280/260 nm. These data were obtained using 4.0 mg BSA for each dialysis; results using 0.4 mg BSA were similar (data not shown). Membranes washed with sulfur-removal solutions (Procedure F) resulted in good agreement for recoveries of BSA as determined by the Bradford and 280 nm absorbances (Table 1, F and Fig. IF). The ratio 280/260-nm absorbance was about 1.56 for predialysis BSA. Multiple washing, up to four times with sulfur-removal solutions, improved the coincidence of the spectra, protein recoveries, and the ratio of 280/260 nm. Combination of membrane treatment procedures and sodium carbonate-ethanol (Procedures D and E) with the sulfur-removal solutions (Procedure F) resulted in good recoveries and improved coincidence of spectra for BSA (data not shown). Spectra for some of the compounds contained in the membrane-shipping solution were obtained. The absorbance for 0.01 mM potassium sorbate was 0.265 at the maximum of 256 nm. This solution was ten thousand times less concentrated than the 0.1 M ( 1.5%) potassium sorbate in the shipping solution. Soaking membranes in 0.1 M EDTA solution, pH 8.0, (3) did not improve protein recoveries, as shown by the spectrum in the region between 280 and 240 nm (data not shown). Using careful control of the reduction in pressure and sample removal procedures, together with appropriate membrane pretreatment, some membranes have been used up to 25 times with no apparent loss of retention

349

MEMBRANES

of BSA and with good coincidence of spectra with initial sample. DISCUSSION

Rapid dialysis-concentration of protein samples was obtained using the MPDC apparatus (5), which combines features of both thin-layer dialysis ( 1,3) and ultrafiltration (4) with constant stirring of a large volume of buffer. Improved recovery of BSA was obtained by pretreatment of membranes with either, or both, sodium carbonate-ethanol or sulfur-removal solutions. These treatments removed soluble membrane components which absorbed in the uv from 240 to 290 nm and resulted in membranes which apparently did not absorb the dialyzing proteins ( 1,3). To estimate recovery of protein in retentates, two procedures for measurements should be used; for example, Bradford protein reaction and uv absorbance. Inaccurate results may result from the use of absorbance only and unaccountable absorbance in the spectrum will not be detected by the Bradford reaction. Adsorption of protein to membranes has been a problem with various batches of conventional Visking dialysis membranes ( 1,3), which has largely been overcome by washing procedures or by modification of carboxyl groups (3). Treatment of membranes to alter pore size or to derivatize carboxyl or other groups ( I ,3) serves an inadvertent purpose of removing soluble components from membranes which contribute to uv absorbance of the retentate. Slightly alkaline washing treatment may remove residual xanthates, a result of the manufacture, from the membranes since xanthates can be brought into solution with alkali (9). Xanthates have a maximum absorbance at 270 nm, in the range of soluble membrane components which we found slowly leached into the retentate at pH 8.0.

350

RICHMOND,

ST. DENIS, AND COHEN

Unidentified sulfur compounds may also contribute to the uv absorbance and are probably removed by the alkaline washing or by treatment of the slightly acidic sulfurremoval solutions. Procedures described here for membrane washing improved both the recovery of protein and the uv spectrum of the retentate, giving assurance of reduced contribution of the membranes to the retentate in at least two ways. This dialyzer-concentrator apparatus enabled us to quantitatively recover acidic proteins after membranes had been treated as described.

REFERENCES 1. Craig, L. C. (1967) in Methods in Enzymology (Hirs, C. H. W., ed.), Vol. I I, pp. 870-905, Academic Press, New York. 2. Richmond, V. L. (198 1) Biochim. Biophys. Acta 669, 193-205.

Chen, H.-C., Craig, L. C., and Stoner, E. (1972) Biochemistry 11, 3559-3564. 4. Berggard, 1. (1962) Ark. Kemi 18, 291-313. 5. Cook, M. W. (1976) Amer. Lab. (Fairfield. Corm.) 8, 3.

61-69.

6. Bradford, M. M. (I 976) Anal. Biochem. 72, 248-254. 7. Koberstein, R., Weber, B., and Jaenicke, R. (1968) Z. Naturforsch. 23b, 474-483. 8. Warburg, O., and Christian, W. (1941) Biochemische Z. 310, 384-421. 9.

Kalckar, H. M. (1947) J. Biol. Chem. 167, 429-443.