- Email: [email protected]
An effective method for defatting albumin using resin columns
376 BBA
SHORT COMMUNICATIONS
33245
An effective method for defatting albumin using resin columns Preparations of native serum albumin are known to contain varying amounts of bound f a t t y acid (refs. I 4; J. K. FULL~;R, unpublished). To obtain f a t t y acid-free preparations for the s t u d y of the physical properties of albumin, several methods have been developed for removing the bound lipids 5 s. It is important t h a t the process can be carried out under conditions which are not likely to denature the protein. We report here a m e t h o d for defatting albumin in the pH range between 3.0 and 4.o using buffered ion-exchange or surface adsorber resin columns. We have found this method at least comparable in effectiveness to tile method using charcoal as an adsorber 8, which is the only other m e t h o d which is carried out at these relatively mild p H values. The following albumin preparations were used: bovine mercaptalbuinin, batch I I I, and a mixture of batches I and I I (8 9 % dimer ; 0.72 SH group per mole) prepared by Dr. T. E. Thompson according to the method of HUGHES AND DINTZIS'a which involved the addition of I mole stearic acid per mole of protein ; H u m a n mercaptalbumin ( 5 - 6 % dimer) prepared by Dr. S. Allerton using the same method, except t h a t no fatty acid was added; and crystalline bovine plasma albumin lot E715o 3 (=-6 °//o dimer' o.65 SH group per mole) from the A r m o u r Pharmaceutical ('o. J The resins employed were: Amberlite CG-4oo, IOO 20o mesh, chromatographic grade, from Mallinekrodt Chemical Works, a strongly basic, quaternary a m m o n i u m anion-exchange polystyrene resin ; and an experimental sample of Amberlite XAI)-2, lot SW 66-o698, a nonexchanging polystyrene sorbent with surface area of 333 m~/g, from R o h m and Haas. Darco activated charcoal was obtained from Atlas Chemical Industries. All chemicals were reagent grade; water was de-ionized. Preparatio~t of io~-exchange resin. Fresh, dry resin was immersed overnight in o.I M NaOH. It was then washed in water, poured into a coluinn, and an extraction mixture containing isopropanol, heptane, and 2. 5 .~I H2SO 4 in the ratio of 4 o : I o : 1 was passed over the resin until the visible front reached the b o t t o m of the column. Three column volumes were collected at the rate of I column vol./h. This removed shedded resin material which would otherwise titrate as fatty acid in a subsequent assay ; it leaves the resin free of such material fi)r at least 6 months. The resin was then washed with isopropanol and water, in succession, and put on the acetate cycle by passing over it a 3-%1d excess of I.O M acetic acid at the rate of I column vol./h. Procedure for removal of fatt 3, acid from albumbz. A 5 % albumin solution was prepared in acetate buffer adiusted to contain o.I M acetate ion at the appropriate pH. This solution was equilibrated by dialysis against the same buffer for 24 h at 2 5 ':~. After this period some solutions showed turbidity which was removed bv filtration through a o.45-# Millipore filter or by centrifugation for 3 h at 34 ooo ~< g at Io °. The filtered sample was then applied to a resin column, which had previously been equilibrated with the same buffer as the protein solution, by passing the buffer down the column for several hours at the flow rate at which the colunm was to be operated. The sample was passed down the column at a flow rate of less than o, 3 column vol./h. Fat O, acid determination. Samples were assayed for long-chain f a t t y acid content by the m e t h o d of DOLF ~°, with two modifications: because the samples were buffered, Biochim. Biophys. Acta, 221 ( t 9 7 o) 376--37~
377
SHORT COMMUNICATIONS TABLE
I
E F F E C T I V E N E S S OF F A T T Y ACID REMOVAL FROM A L B U M I N BY F I L T R A T I O N , BY PASSING OVER A p H - C O N T R O L L E D RESIN COLUMN, AND B Y T H E CHARCOAL METHOD 10
(z) Protein sample
i
Bovine mercaptalbumin (Thompson III)
(2) Initial fatty acid bound*
(4) Fatty acid bound* after filtration
(5) Resin Final fatty acid bound* used in column (6) Re- (7) Charcoal sin method" method
4.8
1.72
--
--
--
2 3
4.4 4 .0
1.23 0.39
4 5
-CG-4°° XAD-2 ---
3.7 0-32 4 . ° ( P ) *** 1.72 0 . 8 8 (72 h ) 4.0 I.O 7 CG-4oo XAD-2 3.7 -CG-4oo
-°.°5 o. 13 ---
------
o.I3 0.29 0.08
--0 . 0 8 ( p H 3.0)
3-7 3.0
0.07 o.oi
o.31 o.oi
6
Human
7
Bovine mercaptalbumin (Thompson I a n d II) Bovine plasma albumin
8 9
mercaptalbumin
1.76
(3) p H of acetate buffer
I.I 1.9 o.66
0.34 0.38
* Moles fatty acid per mole protein. ,t Chen method performed at same pH as comparison "*~ P h o s p h a t e buffer used instead of acetate buffer.
CG-4oo CG-4oo
run, except
where
otherwise
noted.
the extraction mixture was made with 2.5 M instead of o.5 M H2S04; and, to remove acetic acid from the non-aqueous phase containing the assayable material after extraction, this phase was washed 3 times with 3 ml of 0.05 M phosphate buffer at p H 5.5. Since this reduced the volume of the heptane phase, 2 instead of 3 ml were titrated. The effectiveness of the two steps of the method in the removal of bound f a t t y acid from albumin is presented in Table I. Column 4 of Table I shows that a significant reduction in bound f a t t y acid can be achieved for some samples in the step before resin treatment, by filtering off the turbid material produced upon equilibration in the buffer. WILLIAMS AND FOSTERn have found that in o.I M HC1 the f a t t y acids liberated from the albumin coagulate and float on centrifugation of the albumin solutions. However, the aggregates observed in the present studies sedimented on centrifugation and contained about 20% (w/w) of protein, as determined by the method of LowRY et al. 12. The pellet from a sample of bovine mercaptalbumin (Thompson batch III) had been twice resuspended in KC1 and re-sedimented, prior to analysis. The f a t t y acid :protein ratio an the pellet would represent a mole ratio of i o o o : I assuming the protein to be whole albumin. Line 5 of Table I shows that the aggregation is not merely a factor of sample and pH, but also of the buffer anion. Line 6 shows that a sample in which there was virtually no turbidity was nevertheless defatted effectively by the resin. It is suggested that the formation of the filterable turbidity at pH above 3 is related to the presence of small amounts of protein impurity or partially denatured albumin, and, while providing a convenient pre-treatment, is not an essential element of the defatting Biochim. Biophys. Acta,
221 ( 1 9 7 o) 3 7 6 - 3 7 8
378
SHORT COMMUNI( ATI()NS
process described here. I t was n o t e d t h a t p H is an i m p o r t a n t variable in f a t t y acid removal, both in the formation of aggregate a n d in the removal of f a t t y acid by the resin t r e a t m e n t . Similar p H dependence had been found for removal of f a t t y acids b y charcoal t r e a t m e n t in the p H range from 3.o to 5.0 (ref. 8). Lines 7 () show a comparison of the effectiveness of the resin m e t h o d a n d the charcoal method at two p H values for bovine plasma a l b u m i n a n d bovine m e r c a p t a l b u m i n (Thonapson batch I a n d II). The n o n - e x c h a n g i n g resin XAD-2 was found to be only slightly le~,s effective t h a n the ion-exchange resin CG-4oo. On s t u d y i n g fatty acid-acetate exchange in acetate cycle CG-4oo resin in the range of p H 4.6 to pH 6.5, it was found t h a t at the lower of these p H values at least some of the long-chain f a t t y acid b o u n d to the resin w i t h o u t displacing acetate. This suggests t h a t at the p i t values used in the defatting of a l b u m i n , both resins act primarily as surface area sorbents for the long chain f a t t y acids, a n d t h a t for some reason other t h a n its exchange capacity, perhaps due to its smaller mesh size, the CG-4oo is more effective t h a n the XAD-2. No noticeable increase in dimer c o n t e n t (determined b y s e d i m e n t a t i o n velocity analysis on the Spinco model E ultracentrifuge) was observed following defatting of the a l b u m i n in either the resin or the charcoal method. However, some decrease in the sulfllydryl c o n t e n t (measured b y a modified E l l m a n reaction 13) was observed following defatting with resin t r e a t m e n t (approx. IO%), a n d to a greater e x t e n t following charcoal t r e a t m e n t (approx. 2o-25 %). The authors wish to t h a n k Miss Inger H a n s e n tbr her expert technical assistance, Dr. M. J. H u n t e r for her helpful advice, a n d the R o h m and Haas laboratories for the sample of Amberlite XAD-2 resin. Support for the work of J.K.I;. was provided by U.S. Public Health Service National I n s t i t u t e s of Health, Biophysics T r a i n i n g G r a n t GM-I355. Institute of Science and Technology, Biophysics Research Division, Universi(v of Michigan, A n n Arbor, Mich. 48zo5 ( U . S . A . )
W. SCHEIDER .1. K . 1;UI.LER
I }¢. E. KENDALL,J. Biol. Chem., 138 (I94I) 97. 2 E. J. COHN, V~'~. L. HUGHES, JR. AND J. H. WEARE, J. A~ll. Chem. Sot., 69 (1947) I753. 3 R. S. GORDON, JR. AND A. CHERKES,J. Clin. Invest., 35 (1956) 2o6. 4 A. SAIFERAND L. GOLDMAN,J. Lipid Res., 2 (1961) 268. 5 H. M. DINTZlS, Ph.D. Thesis, Harvard University, 1952. 6 D. S. GOODMAN,Science, 125 (1957) 1296. 7 E. J. WILLIAMSAND J. F. FOSTER,J. Am. Chem. Soc., St (19591 865. 8 R. F. CHEN, J. Biol. Chem., 242 (19671 173. 9 W. L. HUGHES, JR. AgO H. M. DINTZlS,J. Biol. Chem., 239 (1964) 845. IO V. P. DOLE, J. Clin. Invest., 35 0956 ) 15°II IL. J. ~VILLIAI~,ISAND J. F. FOSTER,J. Am. Chem. Soc., 82 (I96o) 3741. I 2 0 . H. LOWRY, N. J. ROSEBROUGH,A. L. FARR AND }~. J. RANDALL,.]. Biol. Chem., 193 (1951) 265 . 13 J. JANATOVA,J. K. I;ULLERAND M. J. HUNTER,.]. Biol. Chem., 243 (19681 3612. Received J u l y 28th, I97O Biochim. Biophys. Acta, 2zi (197 ° ) 376 378
SHORT COMMUNICATIONS
33245
An effective method for defatting albumin using resin columns Preparations of native serum albumin are known to contain varying amounts of bound f a t t y acid (refs. I 4; J. K. FULL~;R, unpublished). To obtain f a t t y acid-free preparations for the s t u d y of the physical properties of albumin, several methods have been developed for removing the bound lipids 5 s. It is important t h a t the process can be carried out under conditions which are not likely to denature the protein. We report here a m e t h o d for defatting albumin in the pH range between 3.0 and 4.o using buffered ion-exchange or surface adsorber resin columns. We have found this method at least comparable in effectiveness to tile method using charcoal as an adsorber 8, which is the only other m e t h o d which is carried out at these relatively mild p H values. The following albumin preparations were used: bovine mercaptalbuinin, batch I I I, and a mixture of batches I and I I (8 9 % dimer ; 0.72 SH group per mole) prepared by Dr. T. E. Thompson according to the method of HUGHES AND DINTZIS'a which involved the addition of I mole stearic acid per mole of protein ; H u m a n mercaptalbumin ( 5 - 6 % dimer) prepared by Dr. S. Allerton using the same method, except t h a t no fatty acid was added; and crystalline bovine plasma albumin lot E715o 3 (=-6 °//o dimer' o.65 SH group per mole) from the A r m o u r Pharmaceutical ('o. J The resins employed were: Amberlite CG-4oo, IOO 20o mesh, chromatographic grade, from Mallinekrodt Chemical Works, a strongly basic, quaternary a m m o n i u m anion-exchange polystyrene resin ; and an experimental sample of Amberlite XAI)-2, lot SW 66-o698, a nonexchanging polystyrene sorbent with surface area of 333 m~/g, from R o h m and Haas. Darco activated charcoal was obtained from Atlas Chemical Industries. All chemicals were reagent grade; water was de-ionized. Preparatio~t of io~-exchange resin. Fresh, dry resin was immersed overnight in o.I M NaOH. It was then washed in water, poured into a coluinn, and an extraction mixture containing isopropanol, heptane, and 2. 5 .~I H2SO 4 in the ratio of 4 o : I o : 1 was passed over the resin until the visible front reached the b o t t o m of the column. Three column volumes were collected at the rate of I column vol./h. This removed shedded resin material which would otherwise titrate as fatty acid in a subsequent assay ; it leaves the resin free of such material fi)r at least 6 months. The resin was then washed with isopropanol and water, in succession, and put on the acetate cycle by passing over it a 3-%1d excess of I.O M acetic acid at the rate of I column vol./h. Procedure for removal of fatt 3, acid from albumbz. A 5 % albumin solution was prepared in acetate buffer adiusted to contain o.I M acetate ion at the appropriate pH. This solution was equilibrated by dialysis against the same buffer for 24 h at 2 5 ':~. After this period some solutions showed turbidity which was removed bv filtration through a o.45-# Millipore filter or by centrifugation for 3 h at 34 ooo ~< g at Io °. The filtered sample was then applied to a resin column, which had previously been equilibrated with the same buffer as the protein solution, by passing the buffer down the column for several hours at the flow rate at which the colunm was to be operated. The sample was passed down the column at a flow rate of less than o, 3 column vol./h. Fat O, acid determination. Samples were assayed for long-chain f a t t y acid content by the m e t h o d of DOLF ~°, with two modifications: because the samples were buffered, Biochim. Biophys. Acta, 221 ( t 9 7 o) 376--37~
377
SHORT COMMUNICATIONS TABLE
I
E F F E C T I V E N E S S OF F A T T Y ACID REMOVAL FROM A L B U M I N BY F I L T R A T I O N , BY PASSING OVER A p H - C O N T R O L L E D RESIN COLUMN, AND B Y T H E CHARCOAL METHOD 10
(z) Protein sample
i
Bovine mercaptalbumin (Thompson III)
(2) Initial fatty acid bound*
(4) Fatty acid bound* after filtration
(5) Resin Final fatty acid bound* used in column (6) Re- (7) Charcoal sin method" method
4.8
1.72
--
--
--
2 3
4.4 4 .0
1.23 0.39
4 5
-CG-4°° XAD-2 ---
3.7 0-32 4 . ° ( P ) *** 1.72 0 . 8 8 (72 h ) 4.0 I.O 7 CG-4oo XAD-2 3.7 -CG-4oo
-°.°5 o. 13 ---
------
o.I3 0.29 0.08
--0 . 0 8 ( p H 3.0)
3-7 3.0
0.07 o.oi
o.31 o.oi
6
Human
7
Bovine mercaptalbumin (Thompson I a n d II) Bovine plasma albumin
8 9
mercaptalbumin
1.76
(3) p H of acetate buffer
I.I 1.9 o.66
0.34 0.38
* Moles fatty acid per mole protein. ,t Chen method performed at same pH as comparison "*~ P h o s p h a t e buffer used instead of acetate buffer.
CG-4oo CG-4oo
run, except
where
otherwise
noted.
the extraction mixture was made with 2.5 M instead of o.5 M H2S04; and, to remove acetic acid from the non-aqueous phase containing the assayable material after extraction, this phase was washed 3 times with 3 ml of 0.05 M phosphate buffer at p H 5.5. Since this reduced the volume of the heptane phase, 2 instead of 3 ml were titrated. The effectiveness of the two steps of the method in the removal of bound f a t t y acid from albumin is presented in Table I. Column 4 of Table I shows that a significant reduction in bound f a t t y acid can be achieved for some samples in the step before resin treatment, by filtering off the turbid material produced upon equilibration in the buffer. WILLIAMS AND FOSTERn have found that in o.I M HC1 the f a t t y acids liberated from the albumin coagulate and float on centrifugation of the albumin solutions. However, the aggregates observed in the present studies sedimented on centrifugation and contained about 20% (w/w) of protein, as determined by the method of LowRY et al. 12. The pellet from a sample of bovine mercaptalbumin (Thompson batch III) had been twice resuspended in KC1 and re-sedimented, prior to analysis. The f a t t y acid :protein ratio an the pellet would represent a mole ratio of i o o o : I assuming the protein to be whole albumin. Line 5 of Table I shows that the aggregation is not merely a factor of sample and pH, but also of the buffer anion. Line 6 shows that a sample in which there was virtually no turbidity was nevertheless defatted effectively by the resin. It is suggested that the formation of the filterable turbidity at pH above 3 is related to the presence of small amounts of protein impurity or partially denatured albumin, and, while providing a convenient pre-treatment, is not an essential element of the defatting Biochim. Biophys. Acta,
221 ( 1 9 7 o) 3 7 6 - 3 7 8
378
SHORT COMMUNI( ATI()NS
process described here. I t was n o t e d t h a t p H is an i m p o r t a n t variable in f a t t y acid removal, both in the formation of aggregate a n d in the removal of f a t t y acid by the resin t r e a t m e n t . Similar p H dependence had been found for removal of f a t t y acids b y charcoal t r e a t m e n t in the p H range from 3.o to 5.0 (ref. 8). Lines 7 () show a comparison of the effectiveness of the resin m e t h o d a n d the charcoal method at two p H values for bovine plasma a l b u m i n a n d bovine m e r c a p t a l b u m i n (Thonapson batch I a n d II). The n o n - e x c h a n g i n g resin XAD-2 was found to be only slightly le~,s effective t h a n the ion-exchange resin CG-4oo. On s t u d y i n g fatty acid-acetate exchange in acetate cycle CG-4oo resin in the range of p H 4.6 to pH 6.5, it was found t h a t at the lower of these p H values at least some of the long-chain f a t t y acid b o u n d to the resin w i t h o u t displacing acetate. This suggests t h a t at the p i t values used in the defatting of a l b u m i n , both resins act primarily as surface area sorbents for the long chain f a t t y acids, a n d t h a t for some reason other t h a n its exchange capacity, perhaps due to its smaller mesh size, the CG-4oo is more effective t h a n the XAD-2. No noticeable increase in dimer c o n t e n t (determined b y s e d i m e n t a t i o n velocity analysis on the Spinco model E ultracentrifuge) was observed following defatting of the a l b u m i n in either the resin or the charcoal method. However, some decrease in the sulfllydryl c o n t e n t (measured b y a modified E l l m a n reaction 13) was observed following defatting with resin t r e a t m e n t (approx. IO%), a n d to a greater e x t e n t following charcoal t r e a t m e n t (approx. 2o-25 %). The authors wish to t h a n k Miss Inger H a n s e n tbr her expert technical assistance, Dr. M. J. H u n t e r for her helpful advice, a n d the R o h m and Haas laboratories for the sample of Amberlite XAD-2 resin. Support for the work of J.K.I;. was provided by U.S. Public Health Service National I n s t i t u t e s of Health, Biophysics T r a i n i n g G r a n t GM-I355. Institute of Science and Technology, Biophysics Research Division, Universi(v of Michigan, A n n Arbor, Mich. 48zo5 ( U . S . A . )
W. SCHEIDER .1. K . 1;UI.LER
I }¢. E. KENDALL,J. Biol. Chem., 138 (I94I) 97. 2 E. J. COHN, V~'~. L. HUGHES, JR. AND J. H. WEARE, J. A~ll. Chem. Sot., 69 (1947) I753. 3 R. S. GORDON, JR. AND A. CHERKES,J. Clin. Invest., 35 (1956) 2o6. 4 A. SAIFERAND L. GOLDMAN,J. Lipid Res., 2 (1961) 268. 5 H. M. DINTZlS, Ph.D. Thesis, Harvard University, 1952. 6 D. S. GOODMAN,Science, 125 (1957) 1296. 7 E. J. WILLIAMSAND J. F. FOSTER,J. Am. Chem. Soc., St (19591 865. 8 R. F. CHEN, J. Biol. Chem., 242 (19671 173. 9 W. L. HUGHES, JR. AgO H. M. DINTZlS,J. Biol. Chem., 239 (1964) 845. IO V. P. DOLE, J. Clin. Invest., 35 0956 ) 15°II IL. J. ~VILLIAI~,ISAND J. F. FOSTER,J. Am. Chem. Soc., 82 (I96o) 3741. I 2 0 . H. LOWRY, N. J. ROSEBROUGH,A. L. FARR AND }~. J. RANDALL,.]. Biol. Chem., 193 (1951) 265 . 13 J. JANATOVA,J. K. I;ULLERAND M. J. HUNTER,.]. Biol. Chem., 243 (19681 3612. Received J u l y 28th, I97O Biochim. Biophys. Acta, 2zi (197 ° ) 376 378