Journal of Ncurochumrrir). Vol. 31. pp. I143 to 1 I50 Pergarnon Press Ltd 1979 Printed in Great Britain 0 International Society for Neurochernistry Ltd
00?2-304?~79~1?01I 14290?.00/0
PURIFICATION OF 2’,3’-CYCLIC NUCLEOTIDE 3’-PHOSPHOHYDROLASE FROM BOVINE BRAIN BY IMMUNOAFFINITY CHROMATOGRAPHY: FURTHER BIOCHEMICAL CHARACTERIZATION OF THE PROTEIN ROBERTJ. DRUMMOND’ With the technical assistance of ELIZABETH B. HAMILL The Rockefeller University, I230 York Avenue, New York, NY 10021, U S A . (Received 22 February 1979. Accepted 16 April 1979)
Abstract-The purification of small amounts of 2’,3‘-cyclic nucleotide 3’-phosphohydrolase from bovine white matter by ion-exchange techniques (DRUMMOND et al., 1978) has been used to provide antigen for the production of specific rabbit antibodies to this enzyme. Specific antibody has been purified from immune serum by affinity chromatography on a column of Sepharose to which the enzyme has been attached, and the purified antibody has been coupled to cyanogen bromide-activated Sepharose. Affinity chromatography on the immunoadsorbent effectively purifies 2‘,3‘-cyclicnucleotide 3 -phosphohydrolase in one step from an extract of an acetone powder made from bovine white matter. This modified purification procedure has reduced the time required for purification and increased the yield of the enzyme to 57%. In SDS-gel electrophoresis in phosphate buffer the enzyme migrates as an aggregate of about 98,000MW. When the buffer is Tris-glycine, the apparent MW is about 44,000 and under specific conditions two proteins of only slightly different mobilities can be discerned. Within experimental error the amino acid compositions of the proteins in the two bands are indistinguishable. Peptide patterns obtained by polyacrylamide gel electrophoresis following proteolytic digestion with Straphylococcus aureus V8 protease or papain show extensive structural homology between the two proteins, but detectable differences are apparent.
CYCLIC nucleotide 3’-phosphohydrolase (EC 3.1.4.37) is a membrane-bound enzyme present in highest concentration in white matter from brain and et al., 1962; R. J. in spinal cord (G. I. DRUMMOND DRUMMOND et al., 1978). Early studies demonstrated the presence of this enzymic activity in myelin preparations (KURIHARA& TSUKADA, 1967), parallel in-
in myelin preparations may account for a major portion of the phosphohydrolase activity found therein. We have undertaken the preparation of antibodies to the enzyme (purified according to DRUMMOND et a/., 1978) for the preparation of an immunoadsorbent suitable for the preparation of large quantities of 2’,3’-CN3’-ase and to facilitate the study of the cellular and subcellular localization of the enzyme.
creases in enzymic activity and myelin development in the CNS (KURIHARA& TSUKADA, 1968; BRAUN & BARCHI, 1972), and low levels of enzymic activity MATERIALS AND METHODS in mutant strains of mice whose brains are partially et al., 1969; KURIHARA Materials. Substances and sources were as follows: deficient in myelin (OLAFSON 2’-3’-cyclic cytidylate (sodium salt), 2(N-morpho1ino)-ethet al., 1970). More recent investigations have included the histo- ane sulfonic acid, and papain (2 x crystallized). ribonucchemical work of ENGEL & WOOD(1976), the studies lease (bovine pancreas, Type XII-A), (Sigma, St. Louis, of ZANETTAet al. (1972) and SUNDARRAJ et al. (1975) MO); benzamidine hydrochloride (Aldrich, Milwaukee, WI); bovine serum albumin (protein standard solution, with myelin-free cultured cells, the work of PODUSLO Armour Pharmaceutical Co., Phoenix, AZ); Fast Green & NORTON (1972) and PODUSLO (1975) with purified (Fischer Scientific, Fair Lawn, NJ); Staphylococcus aureus glial membrane preparations, and the localization of V8 protease (Miles Laboratories, Elkhart, IN); molecular enzymic activity in rabbit brain by the use of con- weight markers (product No. 990515) and ultrapure guanitinuous sucrose gradient ultracentrifugation by dinium chloride (Schwarz/Mann, Orangeburg, NY); acrylSHAPIRA et al. (1978). From these experiments it now amide. bis-acrylamide, ammonium persulfate, sodium appears that 2’,3‘-CN 3’-ase may be a glial membrane dodecyl sulfate, and Coomassie Brilliant Blue R-250 (Bioprotein; the occurrence of glial membrane fragments Rad, Richmond, CA); Freud’s adjuvants (Difco. Detroit, MI); immunodiffusion plates (Hyland product No. Abbreviations used: SDS, sodium dodecyl sulfate; MES, 085-073, Costa Mesa, CA); cyanogen bromide activated 2(N-morpholino)-ethane sulfonic acid; 2’,3’-CN 3’-ase, Sepharose 4B (Pharmacia, Uppsala, Sweden); glycine (Calbiochem, San Diego, CA); and male New Zealand rabbits 2’,3’-cyclic nucleotide 3’-phosphohydrolase. 1143
1 I44
ROBERTJ. DRUMMOND
et al., 1978) or the crude (Dutchland, Denver, PA). All other reagents were of ana- phosphohydrolase (DRUMMOND extract (Fig. I). lytical grade. Serum titration. Hemagglutination tests were performed Assay for 2‘,3’-cyclic nucleotide 3’-phosphohydrolase. following the procedure of AVRAMEAS et al. (1969). Human Assays were routinely performed spectrophotometrically by the procedure of HUGLIet al. (1973) (cf. DRUMMONDerythrocytes were coated with antigen by coupling the two with glutaraldehyde (0.5 mg antigen/100 pl packed erythroet al., 1978). One unit of activity is defined as 1 pmol of 2’-CMP formed per min in 1 ml of 1 mht-substrate solution cytes). The reaction was carried out in 0.15 M-phosphate buffer (pH 7.2). 0.028 M in glutaraldehyde. under the conditions of the assay. Preparation of afinity adsorbents. Phosphohydrolase Protein estimation. Protein was determined by the et al., 1978) or phosphohydromethod of LQWRY et al. (1951) with bovine serum albumin (preparation of DRUMMOND lase-specific antibody was coupled to commercially availas standard. Polyacrylamide gel electrophoresis. Disc polyacrylamide able cyanogen bromide-activated Sepharose 4B as follows: gel electrophoresis in the presence of SDS and mercap protein (5-10 mg/ml gel) was dialyzed against NaHCO, toethanol in phosphate buffer was performed according to buffer (0.1 M, pH 8.3), 0.5 M in NaC1. The dialyzed protein, the procedure of WEBER& OSBORN (1969). Slab polyacryl- plus any precipitate which formed during dialysis, was amide gel electrophoresis in the presence of SDS and mer- added to reswollen cyanogen bromide-activated Sepharose captoethanol in Tris-glycine buffer was performed follow- 4B and the coupling was allowed to proceed for 12 h at ing the method of LAEMMLI (1970). Gels were stained for 4°C with gentle mixing. The Sepharose was filtered on sinprotein with Coomassie Brilliant Blue. For estimates of tered glass and washed with the bicarbonate buffer. Unmolecular weight by SDS-gel electrophoresis, bovine serum reacted active groups on the resin were blocked by mixing albumin, chymotrypsinogen, ovalbumin, cytochrome c and the washed gel with 1 M-redistilled ethanolamine at pH 9.0 myoglobin were used as standards. For determination of and 25°C for 2h. The gel was next washed with protein in polyacrylamide gels, gels were stained with Fast 0.1 M-sodium acetate buffer (PH 4.0), 0.5 M in NaC1, folGreen and scanned with a Gilford Model 2520 gel scanner lowed by the NaHCO, buffer, and finally with 0.05 M-MES at 600nm. buffer (pH 6.4), 5% in glycerol. Peptide patterns by gel electrophoresis. Gel electroPreparation of specific antibody. Immune serum phoretic peptide patterns were generated following the pro- (1G30ml) was pumped at a flow rate of 1 ml/min through cedure of CLEVELAND et al. (1977). The proteins were first a column (0.9 x 1Ocm) of the 2.3’-CN 3’-ase-Sepharose resolved on 1G20% acrylamide gradient SDS slab gels. affinity adsorbent. The column was washed with The two bands were excised and the gel slices containing O.O~M-MESbuffer (pH 6.4), 0 . 1 5 ~in NaCl and 5% in the proteins were transferred to a second slab-gel for pro- glycerol until the adsorbtion of the effluent at 254 nm had teolytic digestion and separation of peptides. The large dropped to a stable value, as observed with an in-stream peptide fragments resulting from enzymic cleavage by U.V.monitor. At this point the phosphohydrolase-specific either papain or S. aureus V8 protease during the second antibody was eluted in a single asymmetric peak (Fig. 2) electrophoretic run were visualized by staining with Coo- with 0.2 M-glycine-HC1 buffer (pH 2.8), 0.5 M in NaCI. massie Brilliant Blue. Immediately after elution of antibody from the phosphoAmino acid analyses. For the analysis of proteins in poly- hydrolaseSepharose, the column was re-equilibrated with acrylamide gel slices, a procedure similar to that described starting buffer. If the column was not to be reused within by HOUSTON (1971) was employed. Protein bands visua- a few hours it was washed with starting buffer, 0.01% in lized by staining with Coomassie Brilliant Blue were cut sodium azide. All affinity column operations were perfrom the slabs, placed in hydrolysis tubes and lyophilized. formed at room temperature. The binding capacity of the After lyophilization. 1 ml of 6 N-HCl, 0.2% in phenol and phosphohydrolase-Sepharose for antibody was normally 0.25% in mercaptoacetic acid was added to each tube. 2-3 times greater than the amount of phosphohydrolase Tubes were evacuated, sealed and hydrolyzed for 24 h at originally coupled to the Sepharose. In order to ascertain 110°C. Subsequent preparation of the samples and amino whether all of the antibody in a given sample of serum acid analyses were handled as previously described (DRUMMOND f?f d., 1978). E , b 4 Production of antibody, The procedure for preparation of the purified enzyme by ion exchange chromatography 2.0 R and related techniques has been described previously et al., 1978). Male New Zealand rabbits were (DRUMMOND ”E 1.0. o ~ , injected with 2.0mg of the purified enzyme in Freund’s e complete adjuvant. Initially, one intraperitoneal, two sub5: 0 ~ n 4 15 30 245 60 75 90 . cutaneous, two intramuscular, and two foot-pad injections 15 30 45 60 75 90 4 0 were made, the antigen suspension being divided equally Effluent volume (ml) among the sites. One month later each rabbit was given a second set of similar injections containing a total of FIG. 2. Purification of 2,3’-CN 3’-ase-specific antibodies 2.0 mg of antigen suspended in 1 part Freund’s complete on 2’,3’-CN 3’-aseSepharose. Five milliliters of antiserum adjuvant and 8 parts of Freund’s incomplete adjuvant. was applied to a column (0.9 x IOcm) of immunoadsorWeekly bleedings were taken from the ear vein and the bent. The column was washed (at A) with 0.05wMES serum was titrated by the hemagglutination test. Serum buffer (pH 6.4), 0.15 M in NaCl and 5% in glycerol. Phostitres of 1 :3,000 to 1 : 10,000 were observed 4-8 weeks after phohydrolase-specific antibody was eluted from the adsorthe boosting injections. The specificity of the antiserum bent by washing (at B) with 0.2 M-glycine-HC1 (pH 2.8), in NaCI. Eluted protein was collected (w) and im(about 100ml, stored at -20°C) was checked by the 0 . 5 ~ OUCHTERLONY (1953) immunodiffusion technique. The mediately dialyzed against 0.05 M-MES buffer (pH 6.4), 5% in glycerol. serum yielded a single precipitin arc with either purified
z
:,)\I
FIG. 1. Immunodiffusion test. The center well was filled with antiserum against purified 2‘.3’-CN 3’-ase. Well 1 contained 6 p g of purified enzyme. Wells 2 and 3 contained 7 pl each (21 p g of protein per well) of a crude extract (0.02 M in guanidinium chloride) of an acetone powder from bovine brain white matter.
FIG.4. Gradient (10--200/, acrylamide) slab gel electrophoresis of protein samples from different stages of phosphohydrolase purification. Slab gel electrophoresis was performed following the procedure of LAEMMLI (1970) as described in Materials and Methods. Prior to application the samples were heated for 10min at 100°C in the presence of SDS and mercaptoethanol to dissociate aggregates. Lane 1 contained molecular weight markers. Lane 2 contained 100,ug of protein extracted from an acetone powder of bovinc brain white matter with 1% SDS. Lane 3 contained 1OOpg or protein from the 1 M-guanidinium chloride crude extract. Lane 4 contained 100 p g of protein from the 0.2 M-guanidinium chloride crude extract. Lane 5 contained 50 p g of purified phosphohydrolase from the antibody-Sepharose column. 1145
1146
2‘,3‘-Cyclicnucleotide 3’-phosphohydrolase had been adsorbed, the initial sample was run a second or a third time through the adsorbent until no protein was eluted with the glycine-HCI wash. Antibody was pooled. dialyzed against NaHCO, buffer (0.1 M, pH 8.3), 0.5 M in NaCI, and coupled to cyanogen bromide-activated Sepharose in the same manner as that described for the enzyme. Immunoafinity chromatographic purijication of 2’,3‘-CN 3‘-ase. The phosphohydrolase was solubilized from an acetone powder of bovine white matter by guanidinium chloride extraction as previously described (DRUMMOND et a/., 1978). The only modification of that procedure was the addition of 1 mwbenzamidine hydrochloride to all dialysis solutions to act as an inhibitor of proteolysis. The crude extract, dialyzed free of most of the guanidinium chloride was the starting material for the purification outlined below. Crude extract was pumped through a column (0.9 x 10 cm) of the immunoadsorbent (2’,3’-CN 3I-ase-specific antibody coupled to Sepharose) at a flow rate of 0.5-1 ml/min. Phosphohydrolase activity in the effluent was periodically assayed to monitor the amount of enzyme being bound. When the enzyme activity of the effluent reached 10% of the activity of the influent the bed was washed with 0.05 M-MES (pH 6.4), 0.15 M in NaCl and 5% in glycerol. Once the U.V.absorption of the effluent had fallen to a stable value the bound protein was eluted with glycine-HCI buffer (0.2 M, pH 2.8) 0.5 M in NaCl and 5% in glycerol. The purified phosphohydrolase was eluted in a single symmetrical peak (Fig. 3). Effluent containing the phosphohydrolase was immediately dialyzed against 0.05 M-MES @H 6.4), 25% in glycerol. After regeneration of the adsorbent by washing with 25-50 ml of 0.05 M-MES (pH 6.4), 5% in glycerol, the column was again ready for use. The capacity of a column of this size is about 3000 units of enzyme; it takes four to five runs to process the extract obtained from 20 g of acetone powder. All chromatographic steps were performed at room temperature. Both the enzyme column and the antibody column were used 30 times without evidence of deterioration. Phosphate determination. All phosphohydrolase preparations to be analyzed for phosphate were first checked for homogeneity by SDS-gel electrophoresis under reducing conditions. Samples in which 95% of more of the protein was present in the doublet, as determined by densitometry on Fast Green-stained gels, were dialyzed against 1 1. of 0.05 M-potassium phosphate buffer (pH 6.8) for 12 h, 1 1. of 0.5 M-NaCI for 12 h and 4 x 1 1. of 0.1 M-acetic acid for 48 h. Dialyzed samples (5 mg each) were lyophilized
1147
E
C
0
ft
2In
0
15
75
90
I05
120
135
Effluent volume (ml)
FIG.3. Purification of 2’,3’-CN 3’-ase on antibody-Sepharose. A crude extract of bovine brain acetone powder (0.02 M in guanidinium chloride) was pumped through a column (0.9 x IOcm) of immunoadsorbent at a flow rate of 0.5 ml/min. The column was washed (at A) with 0.05 M-MES buffer (pH 6.4), 0.15 M in NaCl and 57, in glycerol. Purified phosphohydrolase was eluted from the adsorbent by washing the column (at B) with 0.2 M-glyCineHCI (PH 2.8), 0.5 M in NaCl and 5% in glycerol. Eluted protein was collected (t.) and immediately dialyzed against 0.05 M-MES buffer (pH 6.4), 25% in glycerol. in acid washed hydrolysis tubes. One milliliter of 6 N-HCI, 0.2% in phenol, was added to each tube. The tubes were evacuated, sealed and hydrolyzed at 110°C for 24 h. Aliquots of the hydrolysate were assayed for phosphate following the procedure of ITAYA& UI (1966). Protein concentration was determined by amino acid analysis. Ribonuclease was carried through the same dialysis procedure to serve as a control. RESULTS
Purification b y immunoafinity chromatography The purification of 2’,3’-CN 3’-ase by immunoadsorption is summarized in Table 1. The purity of the enzyme a t each step of the procedure, as determined by SDS-gel electrophoresis, is shown in Fig. 4; the significance of the doublet observable in the figure is discussed below. Essentially n o contaminating protein could be observed on SDS gels when 5 0 p g samples of the purified enzyme were analyzed. By this criterion, as well as by densitometric scans of Fast Green-stained gels, it was estimated that the preparations of phosphohydrolase were of greater than 95% purity. The overall recovery of enzymic activity was 57%. Specific activities of different preparations
FIG.5. SDS gradient (1&20% acrylamide) slab gel electrophoresis of proteins in gel slices cut from a previously run gradient slab gel which resolved the slower and the faster components of the enzyme preparation. These slices were loaded onto a second gradient SDS slab gel and run according to the procedure of CLEVELAND et al. (1977). Lanes 1 and 5 contained molecular weight markers. Lane 2 contained a single gel slice containing the initially slower moving protein. Lane 3 contained two gel slices containing the faster moving protein. Lane 4 contained two gel slices containing the faster moving component and a third gel slice containing the slower moving one. FIG.6. SDS slab gel electrophoresis of peptides from proteins in the upper and lower bands (Fig. 4). Phosphohydrolase components were previously resolved on SDS slab gels run under reducing conditions. Upper and lower protein bands were excised, transferred to a second slab gel, and peptide maps were generated by the procedure of CLEVELAND et af. (1977) as outlined in Materials and Methods. Lane 1 contained molecular weight markers. As a control lane 2 contained the peptides formed when 1Opg of bovine serum albumin in solution and 0.2pg of papain were submitted to electrophoresis. Lanes 3 and 4 contained upper and lower phosphohydrolase bands, respectively, plus 0.2pg of S . aureus V8 protease (-+ marks the band resulting from the protease). Lanes 5 and 6 contained upper and lower phosphohydrolase bands, respectively, plus 0.2 pg of papain. Lane 7 contained 0.2 pg of papain, a nondetectable amount of protein.
1 I48
ROBERTJ. DRUMMOND
TABLE1. PURlFlCATlON OF
2,3’-CYCLlC NUCLEOTIDE
Total activity Step 1. Crude extract
(I M in GuClt) 2. Supernatant after dialysis of extract to 0.2~-GuC1 3. Supernatant after dialysis of extract to 0.02 M-GuCI 4. Immunoadsorption
3’-PHOSPHOHYDROLASE BY
IMMUNOAFFINITY CHROMATOGRAPHY
(4
Total protein (ms)
Specific activity* (u/mg protein)
Recovery of activity
15,290
23 10
6.6
100
21,100
1620
13.0
> 100
13.680
840
16.3
89
8755
38
(%)
233
57
* Assay of HUGLIet al. (1973). t Soluble protein was first extracted from 20 g of acetone powder (approximately equivalent to
100 g of white matter) by aqueous buffer at pH 7.5. The insoluble pellet was extracted with 1 M-guanidinium chloride (see DRUMMOND et of., 1978).
of enzyme varied from a low value of approx 200u/mg of protein to a high value of nearly 300 u/mg of protein depending on the age of the acetone powder. It was observed that the total enzymic activity that could be extracted from 20g of acetone powder gradually declined from approx 20,000 units (freshly prepared acetone powder) to 10,OOO units after 3 months storage at 4°C. The phosphohydrolase preparations purified from freshly prepared acetone powder had the highest specific activities. Gel analyses under reducing conditions on preparations of enzyme with low or high specific activities revealed both preparations to consist of more than 95% of the doublet shown in lane 5, Fig. 4. When the immunoadsorbent-Sepharose is handled as previously described (See Materials and Methods) it has been found to be stable for up to 6 months with only minimal loss in binding capacity or specificity. However, we have encountered difficulty when attempting to employ the immunoadsorbents prepared from sera of bleedings taken 6 1 2 months after the initial inoculation. These sera have titres of from 1 :10,OOO to 1:2O,OOO. Immunoadsorbents prepared from such sera yield phosphohydrolase of approx 80% purity. Contaminating protein consists almost solely of the major low molecular weight protein seen with the crude extract sample in Fig. 4, lane 2. We believe that this contaminant is myelin basic protein because of its apparent molecular weight of approx 18,000 (EYLAR,1970) and the fact that it is the predominant protein when samples of purified myelin are run on SDS gels. Gel electrophoresis
In the course of these studies, gel electrophoresis under different conditions has established some special properties of the enzyme. As reported earlier (DRUMMOND et al., 1978), when the purified enzyme (1969) techis examined by the WEBER& OSBORN nique under non-reducing conditions, the enzyme migrates as a protein of about 98,000MW. The buffer in this instance is phosphate. However, when we conduct the electrophoresis without reducing agent using
the LAEMMLI (1970) system in Tris-glycine buffer, we obtain 75% of the protein in a band in the 44,000 MW range. With either system, when reduction and heat are used to carry the disaggregation to completion, all of the protein is in the lower molecular weight form. Thus this phosphohydrolase tends to remain aggregated in phosphate buffer, even with SDS present. A second observation concerns the nature of the two barely separable proteins observable in Fig. 4. The ratio of the amount of protein in the slower band to that in the faster one was generally about 2 : 1 when Fast Green-stained gels were scanned densitometrically. In order to characterize further the proteins in these two zones, the bands were excised from the slab TABLE2. AMINO ACID ANALYSES OF UPPER AND LOWER BANDS OF 2’,3’-CYCLIC NUCLEOTIDE 3’-PHOSPHOHYDROLASE FROM SDS-GELS* Amino acid Asx
Thr Ser Glx Pro GIY Ala Val Met Ile Leu TYr Phe His LYs ‘4%
Mole yo Upper band Lower band 8.24 k 0.58 5.40 f 0.13 6.57 & 0.31 13.7 f 0.52 4.59 k 0.35 8.94 f 0.28 7.58 k 0.49 5.13 0.01 1.80 f 0.31 2.16 f 0.19 12.3 -t 0.57 3.32 f 0.15 5.42 0.21 1.36 f 0.05 9.87 k 0.55 6.29 It: 0.54
+
8.51 f 0.16 5.05 f 0.43 5.95 f 0.37 13.9 0.35 4.61 + 0.32 8.58 k 0.20 7.54 + 0.02 5.14 It: 0.41 1.81 k 0.10 2.10 k 0.07 12.4 0.42 3.57 & 0.20 5.09 0.18 1.47 k0.15 10.0 f 0.43 6.22 _+ 0.09
*
* Twenty-four hour acid hydrolyses were performed on the upper and lower protein bands of 2‘,3’-cyclicnucleotide 3’-phosphohydrolase (prepared according to the method presented in this paper) after separation on SDS polyacrylamide slab gels run under reducing conditions (See Materials and Methods).Values are expressed as the mean f S.D. (three determinations each on upper and lower bands) and have not been corrected for the loss of labile amino acids. Half-cystine and tryptophan were not determined.
2’,3‘-Cyclic nucleotide 3’-phosphohydrolase
gel and resubmitted to electrophoresis (Fig. 5 ) ; each retained its characteristic mobility. To determine whether the smaller molecular weight component of the doublet might be a product of proteolysis taking place during purification, a modified purification protocol was adopted. The only difference between this method and the standard procedure was that benzamidine hydrochloride was not added to the dialysis solutions and all crude extracts (steps 1-3, Table 1) were brought to room temperature for at least 1 h. Gel analysis, under reducing conditions, of the enzyme thus purified showed the same relative amounts of the two proteins as were found for the enzyme prepared by the standard procedure. Amino acid analysis The next step was to see whether the two proteins differed in amino acid composition. They were resolved by gel electrophoresis under reducing conditions and amino acid analyses were performed on the excised gel slices. The amino acid compositions are given in Table 2. Because of relatively large blank gel corrections (10-30% of the uncorrected values) for certain amino acids, this procedure yielded corrected figures having standard deviations of 5-10%. The data permit the conclusion that the two proteins clearly have nearly the same amino acid composition. Peptide patterns
To compare structural features in further detail peptide patterns of each form were compared by SDS-polyacrylamide gel electrophoresis (Fig. 6). Control experiments were run with bovine serum albumin to show that specific enzymic clevages were occurring during the gel electrophoresis. As demonstrated by CLEVELAND et al. (1977), specific peptide patterns were observed when bovine serum albumin was subjected to electrophoresis in the presence of either papain or S. aureus V8 protease. Since we were not working with radioactively labeled protein we found it difficult to load sufficient amounts of protein in gel slices to yield observable amounts of peptides and still maintain the resolution required for this particular application. Hence, the gel analyses presented in Fig. 6 contained barely discernible peptides when stained with Coomassie Brilliant Blue. Although sample loads were not optimal it was found that a number of peptides of identical SDS-gel electrophoretic mobility were generated for the two phosphohydrolase components when either papain or S. aureus V8 protease were present. Some of the peptides produced by papain digestion were clearly distinct from those produced by S. aureus protease. Although the majority of the papain or S. aureus V8 peptides from each protein had the same mobilities (Nos. 1, 2, 3 and 4 for papain; Nos. 1, 2, 3, 4, 5, 6 and 9 for S. aureus V8) there were two unique peptides for each subunit (Nos. 5 and 6 for papain and Nos. 7 and 8 for S. aureus V8). Thus the two proteins are homologous, but differences exist.
1149
Phosphate determination
In order to determine whether we might be dealing with phosphoproteins, samples of the enzyme and ribonuclease (as a control) were exhaustively dialyzed against 0.5 M-NaCl and 0.1 M-acetic acid. Phosphate determinations on these samples showed less than 0.06 pmol of phosphate per pmol of ribonuclease and 0.294.48pnol of phosphate per pmol of phosphohydrolase subunits (three individual preparations of enzyme). The major protein of the phosphohydrolase doublet is clearly not a phosphoprotein and further work will be required to determine whether the small amount of phosphate detectable represents residual non-covalent binding of phosphate or organically bound phosphate in the minor protein. DISCUSSION
By employing the technique of immunoaffinity chromatography, 2’,3’-cyclic nucleotide 3’-phosphohydrolase can now be purified in 40mg amounts in half the time that it previously took to purify 4mg. The overall yield of protein has been increased from 7 to 57%. The initial investment of time and effort for the preparation of pure antigen by ion-exchange chromatography and related techniques (DRUMMOND et al., 1978) and for producing the antibody in rabbits is repaid many-fold when the enzyme is finally prepared from crude extract by the immunoaffinity technique. Antigen preparations employed in this study did not contain amounts of myelin basic protein detectable by gel-electrophoresis, but we conclude that very small amounts of this protein or a protein of identical SDSgel electrophoretic properties were present and gave rise to significant antibody titres after extended lengths of time, If early bleedings are employed for the preparation of affinity adsorbents. antibody against contaminating protein is not present in sufficient amounts to cause interference in the immunoaffinity purification. The recent experiments of DE JONG et a/. (1978) on SDS-gel electrophoresis bear upon the interpretation of the different mobilities of the proteins seen in the doublet in Fig. 4. DE JONG et al. show that changes as small as the substitution of one residue in a sequence can result in significant differences in mobility, probably as a result of changes in SDS binding. They recommend caution in the interpretation of differences in mobility in terms of molecular weight alone. Our own experience substantiates this view; relative to the same markers in Tris-glycine buffer we observe a 10% higher ‘apparent molecular weight’ for the phosphohydrolase in the Weber & Osborn system than in the Laemmli system. The two separable proteins seen in Fig. 4 may not differ as much in molecular weight as might be indicated by a conventional interpretation of the gel pattern. The agreement between the amino acid compositions does not leave much room for the deletion of a peptide fragment of appreciable size.
ROBERTJ. DRUMMOND
1I50
Further research will be required to establish the nature of the chemical differences in the proteins as evidenced by the peptide maps in Fig. 6. Sequence differences are a possibility; a covalently bound prosthetic group is another. The presence of measurable phosphate led us t o consider the possible presence of ADP-ribose; when the U.V.spectra of the enzyme a nd serum albumin were compared the phosphohydrolase was found t o have n o unique absorption a t 260 nm, which eliminates the possibility of the presence of a nucleotide. A unique crosslink in one of the proteins could also result in a slight difference in mobility in SDS-gels between two proteins which have essentially identical amino acid sequences. An isopeptide crosslink between the €-amino group of lysine and a glycine carboxyl group results in a branched structure linking nonhistone and histone 2A polypeptides (GOLDKNOPF & BUSCH, 1977). E-(y-glutamy1)-Lysine crosslinks between fibrin subunits have also been characterized (PISANO et al., 1971). Although these are less probable explanations, crosslinks of this nature could exist and give rise to proteins and peptides of different SDS-gel mobility. The increased availability of the purified protein should make possible more detailed studies of its chemistry and the use of immunological methods for its subcellular localization. Acknowledgements-This research was supported in part by USPHS grants GM 07256 and GM 25323 and an R. J. Reynolds Post-Doctoral Fellowship. The author is indebted to GEORGE KUZMYCZ of EDWARD REICH’S laboratory for assistance in the preparation of antisera, to STANFORD MOOREand WILLIAM H. STEINfor cooperation in CURthe organization of this manuscript, and to BARBARA TOPELLE for the preparation of the typescript. REFERENCES AVRAMEASS., TAUDOU B. & CHUILON S. (1969) Glutaraldehyde, cyanuric chloride and tetraazotized o-dianisidine as coupling reagents in the passive hemagglutination test. Immunochemistry 6, 67-76. BRAUN P. E. & BARCHI R. L. (1972) 2’,3’-Cyclicnucleotide 3’-phosphodiesterase in the nervous system. Electrophoretic properties and developmental studies. Brain Res. 40,437444. S. G., KIRSCHNER M. W. & CLEVELAND D. W., FISCHER LAEMMLI U. K. (1977) Peptide mapping by limited proteolysis in sodium dodecyl sulfate and analysis by gel electrophoresis. J . b i d . Chem. 252, 1102-1 106. DE JONC W. W., ZWEERSA. & COHENL. H. (1978) Influence of single amino acid substitutions on electrophoretic mobility of sodium dodecyl sulfate-protein complexes. Biochem. biophys. Res. Commun. 82, 532-539. DRUMMOND G. I., IYERN. T . & KEITHJ . (1962) Hydrolysis of ribonucleoside 2’,3’-cyclicphosphates by a diesterase from brain. J . biol. Chem. 237, 3535-3539. E. B. & GUHAA. (1978) PurificaDRUMMOND R. J., HAMILL tion and comparison of 2‘,3’-cyclic nucleotide 3‘-phosphohydrolases from bovine brain and spinal cord. J . Neurochem. 31, 871-878. ENCELE. L. & WOODJ. G. (1976) Cytochemical localiza-
tion of 2’,3‘-cyclic nucleotide 3‘-phosphohydrolase in CNS white matter. Neurosci. Abstr. 2, 410. EYLARE. H. (1970) Amino acid sequence of the basic protein of the myelin membrane. Proc. natn. Arad. Sci., U.S.A. 67, 1425-1431. GOLDKNOPF I. L. & BUSCHH. (1977) Isopeptide linkage between nonhistone and histone 2A polypeptides of chromosomal conjugate-protein A24. Proc. natn. Acad. Sci., U.S.A. 74, 864-868. HOUSTON L. L. (1971)Amino acid analysis of stained bands from polyacrylamide gels. Analyt. Biocltem. 44, 81-88. HUGLIT. E., BUSTINM. & MCORES. (1973) Spectrophotometric assay of 2’.3’-cyclicnucleotide 3’-phosphohydrolase: application to enzyme in bovine brain. Brain Res. 58, 191-203. ITAYAK. & UI M. (1966) A new micromethod for the colorimetric determination of inorganic phosphate. Clinica chim. Acta 14, 361-366. Y. (1967) The regional and subKURIHARA T. & TSUKADA cellular distribution of 2’,3‘-cyclicnucleotide 3’-phosphohydrolase in the central nervous system. J . Netrrochem. 14, 1167-1174. KURIHARA T. & TSUKADA Y. (1968) 2‘,3’-Cyclicnucleotide 3‘-phosphohydrolase in the developing chick brain and spinal cord. J . Neurochem. 15, 827-832. KURIHARA T., NUSSBAUM J. L. & MANDELP. (1970) 2’,3’-Cyclicnucleotide 3’-phosphohydrolase in brains of mutant mice with deficient myelination. J . Neurochem. 17, 993-997. LAEMMLI U. K. (1970) Cleavage of structural proteins during the assembly of the head or bacteriophage T4. Nature, Lond. 227, 680-6235, LOWRY0. H., ROSEBROUCHN. J., FARRA. L. & RANDALL R. J. (1951) Protein measurement with the Fohn phenol reagent. J . biol. Cheni. 193, 265-215. OLAFSON R. W., DRUMMOND G. I. & LEE J. F. (1969) Studies on 2’,3’-cyclic nucleotide-3’-phosphohydrolase from brain. Can. J . Biocheni. 47, 961-966. OUCHTERLONY D. (1953) Antigen-antibody reactions in gels. Acta path. scand. 32. 231-240. PISANO J. J., FINLAYSON J. S., PEYTON M. P. & NAGAI Y.(1971)e-(y-glutamyl) lysine in fibrin: lack of crosslink formation in Factor XI11 deficiency. Proc. narn. Acad. Sci., U S A . 68, 77e712. PODUSLO S. E. (1975) The isolation and characterization of a plasma membrane and a myelin fraction derived from oligodendroglia of calf brain. J . Neurochem. 24, 647-654. PODUSLO S. E. & NORTONW. T. (1972) Isolation and some chemical properties of oligodendroglia from calf brain. J . Neurochem. 19, 727-736. M. R., SHAPIRA R., MOBLEY W. C., THIELES. B., WILHELMI WALLACEA. & KIBLERR. F. (1978) Localization of 2,3’-cyclic nucleotide-3’-phosphohydrolase of rabbit brain by sedimentation in a continuous sucrose gradient. J . Neurochem. 30, 735-744. S. E. (1975) SUNDARRAJ N., SCHACHNER M. & PFEIFF~R Biochemically differentiated mouse glial lines carrying a nervous system specific cell surface antigen (NS-I). Proc. natn. Acad. Sci., U.S.A. 12, 1927-1931. WEBERK. & OSBORNM. (1969) Reliability of molecular weight determinations by dodecyl sulfate polyacrylamide gel electrophoresis. J . biol. Chem. 244, 4406-4412. ZANETTA J. P., BENDAP., GOMBOSG. & MORGANI. G. (1972) The presence of 2‘,3’-cyclicAMP 3’-phosphohydrolase in glial cells in tissue culture. J . Neurochem. 19, 881-883.
00?2-304?~79~1?01I 14290?.00/0
PURIFICATION OF 2’,3’-CYCLIC NUCLEOTIDE 3’-PHOSPHOHYDROLASE FROM BOVINE BRAIN BY IMMUNOAFFINITY CHROMATOGRAPHY: FURTHER BIOCHEMICAL CHARACTERIZATION OF THE PROTEIN ROBERTJ. DRUMMOND’ With the technical assistance of ELIZABETH B. HAMILL The Rockefeller University, I230 York Avenue, New York, NY 10021, U S A . (Received 22 February 1979. Accepted 16 April 1979)
Abstract-The purification of small amounts of 2’,3‘-cyclic nucleotide 3’-phosphohydrolase from bovine white matter by ion-exchange techniques (DRUMMOND et al., 1978) has been used to provide antigen for the production of specific rabbit antibodies to this enzyme. Specific antibody has been purified from immune serum by affinity chromatography on a column of Sepharose to which the enzyme has been attached, and the purified antibody has been coupled to cyanogen bromide-activated Sepharose. Affinity chromatography on the immunoadsorbent effectively purifies 2‘,3‘-cyclicnucleotide 3 -phosphohydrolase in one step from an extract of an acetone powder made from bovine white matter. This modified purification procedure has reduced the time required for purification and increased the yield of the enzyme to 57%. In SDS-gel electrophoresis in phosphate buffer the enzyme migrates as an aggregate of about 98,000MW. When the buffer is Tris-glycine, the apparent MW is about 44,000 and under specific conditions two proteins of only slightly different mobilities can be discerned. Within experimental error the amino acid compositions of the proteins in the two bands are indistinguishable. Peptide patterns obtained by polyacrylamide gel electrophoresis following proteolytic digestion with Straphylococcus aureus V8 protease or papain show extensive structural homology between the two proteins, but detectable differences are apparent.
CYCLIC nucleotide 3’-phosphohydrolase (EC 3.1.4.37) is a membrane-bound enzyme present in highest concentration in white matter from brain and et al., 1962; R. J. in spinal cord (G. I. DRUMMOND DRUMMOND et al., 1978). Early studies demonstrated the presence of this enzymic activity in myelin preparations (KURIHARA& TSUKADA, 1967), parallel in-
in myelin preparations may account for a major portion of the phosphohydrolase activity found therein. We have undertaken the preparation of antibodies to the enzyme (purified according to DRUMMOND et a/., 1978) for the preparation of an immunoadsorbent suitable for the preparation of large quantities of 2’,3’-CN3’-ase and to facilitate the study of the cellular and subcellular localization of the enzyme.
creases in enzymic activity and myelin development in the CNS (KURIHARA& TSUKADA, 1968; BRAUN & BARCHI, 1972), and low levels of enzymic activity MATERIALS AND METHODS in mutant strains of mice whose brains are partially et al., 1969; KURIHARA Materials. Substances and sources were as follows: deficient in myelin (OLAFSON 2’-3’-cyclic cytidylate (sodium salt), 2(N-morpho1ino)-ethet al., 1970). More recent investigations have included the histo- ane sulfonic acid, and papain (2 x crystallized). ribonucchemical work of ENGEL & WOOD(1976), the studies lease (bovine pancreas, Type XII-A), (Sigma, St. Louis, of ZANETTAet al. (1972) and SUNDARRAJ et al. (1975) MO); benzamidine hydrochloride (Aldrich, Milwaukee, WI); bovine serum albumin (protein standard solution, with myelin-free cultured cells, the work of PODUSLO Armour Pharmaceutical Co., Phoenix, AZ); Fast Green & NORTON (1972) and PODUSLO (1975) with purified (Fischer Scientific, Fair Lawn, NJ); Staphylococcus aureus glial membrane preparations, and the localization of V8 protease (Miles Laboratories, Elkhart, IN); molecular enzymic activity in rabbit brain by the use of con- weight markers (product No. 990515) and ultrapure guanitinuous sucrose gradient ultracentrifugation by dinium chloride (Schwarz/Mann, Orangeburg, NY); acrylSHAPIRA et al. (1978). From these experiments it now amide. bis-acrylamide, ammonium persulfate, sodium appears that 2’,3‘-CN 3’-ase may be a glial membrane dodecyl sulfate, and Coomassie Brilliant Blue R-250 (Bioprotein; the occurrence of glial membrane fragments Rad, Richmond, CA); Freud’s adjuvants (Difco. Detroit, MI); immunodiffusion plates (Hyland product No. Abbreviations used: SDS, sodium dodecyl sulfate; MES, 085-073, Costa Mesa, CA); cyanogen bromide activated 2(N-morpholino)-ethane sulfonic acid; 2’,3’-CN 3’-ase, Sepharose 4B (Pharmacia, Uppsala, Sweden); glycine (Calbiochem, San Diego, CA); and male New Zealand rabbits 2’,3’-cyclic nucleotide 3’-phosphohydrolase. 1143
1 I44
ROBERTJ. DRUMMOND
et al., 1978) or the crude (Dutchland, Denver, PA). All other reagents were of ana- phosphohydrolase (DRUMMOND extract (Fig. I). lytical grade. Serum titration. Hemagglutination tests were performed Assay for 2‘,3’-cyclic nucleotide 3’-phosphohydrolase. following the procedure of AVRAMEAS et al. (1969). Human Assays were routinely performed spectrophotometrically by the procedure of HUGLIet al. (1973) (cf. DRUMMONDerythrocytes were coated with antigen by coupling the two with glutaraldehyde (0.5 mg antigen/100 pl packed erythroet al., 1978). One unit of activity is defined as 1 pmol of 2’-CMP formed per min in 1 ml of 1 mht-substrate solution cytes). The reaction was carried out in 0.15 M-phosphate buffer (pH 7.2). 0.028 M in glutaraldehyde. under the conditions of the assay. Preparation of afinity adsorbents. Phosphohydrolase Protein estimation. Protein was determined by the et al., 1978) or phosphohydromethod of LQWRY et al. (1951) with bovine serum albumin (preparation of DRUMMOND lase-specific antibody was coupled to commercially availas standard. Polyacrylamide gel electrophoresis. Disc polyacrylamide able cyanogen bromide-activated Sepharose 4B as follows: gel electrophoresis in the presence of SDS and mercap protein (5-10 mg/ml gel) was dialyzed against NaHCO, toethanol in phosphate buffer was performed according to buffer (0.1 M, pH 8.3), 0.5 M in NaC1. The dialyzed protein, the procedure of WEBER& OSBORN (1969). Slab polyacryl- plus any precipitate which formed during dialysis, was amide gel electrophoresis in the presence of SDS and mer- added to reswollen cyanogen bromide-activated Sepharose captoethanol in Tris-glycine buffer was performed follow- 4B and the coupling was allowed to proceed for 12 h at ing the method of LAEMMLI (1970). Gels were stained for 4°C with gentle mixing. The Sepharose was filtered on sinprotein with Coomassie Brilliant Blue. For estimates of tered glass and washed with the bicarbonate buffer. Unmolecular weight by SDS-gel electrophoresis, bovine serum reacted active groups on the resin were blocked by mixing albumin, chymotrypsinogen, ovalbumin, cytochrome c and the washed gel with 1 M-redistilled ethanolamine at pH 9.0 myoglobin were used as standards. For determination of and 25°C for 2h. The gel was next washed with protein in polyacrylamide gels, gels were stained with Fast 0.1 M-sodium acetate buffer (PH 4.0), 0.5 M in NaC1, folGreen and scanned with a Gilford Model 2520 gel scanner lowed by the NaHCO, buffer, and finally with 0.05 M-MES at 600nm. buffer (pH 6.4), 5% in glycerol. Peptide patterns by gel electrophoresis. Gel electroPreparation of specific antibody. Immune serum phoretic peptide patterns were generated following the pro- (1G30ml) was pumped at a flow rate of 1 ml/min through cedure of CLEVELAND et al. (1977). The proteins were first a column (0.9 x 1Ocm) of the 2.3’-CN 3’-ase-Sepharose resolved on 1G20% acrylamide gradient SDS slab gels. affinity adsorbent. The column was washed with The two bands were excised and the gel slices containing O.O~M-MESbuffer (pH 6.4), 0 . 1 5 ~in NaCl and 5% in the proteins were transferred to a second slab-gel for pro- glycerol until the adsorbtion of the effluent at 254 nm had teolytic digestion and separation of peptides. The large dropped to a stable value, as observed with an in-stream peptide fragments resulting from enzymic cleavage by U.V.monitor. At this point the phosphohydrolase-specific either papain or S. aureus V8 protease during the second antibody was eluted in a single asymmetric peak (Fig. 2) electrophoretic run were visualized by staining with Coo- with 0.2 M-glycine-HC1 buffer (pH 2.8), 0.5 M in NaCI. massie Brilliant Blue. Immediately after elution of antibody from the phosphoAmino acid analyses. For the analysis of proteins in poly- hydrolaseSepharose, the column was re-equilibrated with acrylamide gel slices, a procedure similar to that described starting buffer. If the column was not to be reused within by HOUSTON (1971) was employed. Protein bands visua- a few hours it was washed with starting buffer, 0.01% in lized by staining with Coomassie Brilliant Blue were cut sodium azide. All affinity column operations were perfrom the slabs, placed in hydrolysis tubes and lyophilized. formed at room temperature. The binding capacity of the After lyophilization. 1 ml of 6 N-HCl, 0.2% in phenol and phosphohydrolase-Sepharose for antibody was normally 0.25% in mercaptoacetic acid was added to each tube. 2-3 times greater than the amount of phosphohydrolase Tubes were evacuated, sealed and hydrolyzed for 24 h at originally coupled to the Sepharose. In order to ascertain 110°C. Subsequent preparation of the samples and amino whether all of the antibody in a given sample of serum acid analyses were handled as previously described (DRUMMOND f?f d., 1978). E , b 4 Production of antibody, The procedure for preparation of the purified enzyme by ion exchange chromatography 2.0 R and related techniques has been described previously et al., 1978). Male New Zealand rabbits were (DRUMMOND ”E 1.0. o ~ , injected with 2.0mg of the purified enzyme in Freund’s e complete adjuvant. Initially, one intraperitoneal, two sub5: 0 ~ n 4 15 30 245 60 75 90 . cutaneous, two intramuscular, and two foot-pad injections 15 30 45 60 75 90 4 0 were made, the antigen suspension being divided equally Effluent volume (ml) among the sites. One month later each rabbit was given a second set of similar injections containing a total of FIG. 2. Purification of 2,3’-CN 3’-ase-specific antibodies 2.0 mg of antigen suspended in 1 part Freund’s complete on 2’,3’-CN 3’-aseSepharose. Five milliliters of antiserum adjuvant and 8 parts of Freund’s incomplete adjuvant. was applied to a column (0.9 x IOcm) of immunoadsorWeekly bleedings were taken from the ear vein and the bent. The column was washed (at A) with 0.05wMES serum was titrated by the hemagglutination test. Serum buffer (pH 6.4), 0.15 M in NaCl and 5% in glycerol. Phostitres of 1 :3,000 to 1 : 10,000 were observed 4-8 weeks after phohydrolase-specific antibody was eluted from the adsorthe boosting injections. The specificity of the antiserum bent by washing (at B) with 0.2 M-glycine-HC1 (pH 2.8), in NaCI. Eluted protein was collected (w) and im(about 100ml, stored at -20°C) was checked by the 0 . 5 ~ OUCHTERLONY (1953) immunodiffusion technique. The mediately dialyzed against 0.05 M-MES buffer (pH 6.4), 5% in glycerol. serum yielded a single precipitin arc with either purified
z
:,)\I
FIG. 1. Immunodiffusion test. The center well was filled with antiserum against purified 2‘.3’-CN 3’-ase. Well 1 contained 6 p g of purified enzyme. Wells 2 and 3 contained 7 pl each (21 p g of protein per well) of a crude extract (0.02 M in guanidinium chloride) of an acetone powder from bovine brain white matter.
FIG.4. Gradient (10--200/, acrylamide) slab gel electrophoresis of protein samples from different stages of phosphohydrolase purification. Slab gel electrophoresis was performed following the procedure of LAEMMLI (1970) as described in Materials and Methods. Prior to application the samples were heated for 10min at 100°C in the presence of SDS and mercaptoethanol to dissociate aggregates. Lane 1 contained molecular weight markers. Lane 2 contained 100,ug of protein extracted from an acetone powder of bovinc brain white matter with 1% SDS. Lane 3 contained 1OOpg or protein from the 1 M-guanidinium chloride crude extract. Lane 4 contained 100 p g of protein from the 0.2 M-guanidinium chloride crude extract. Lane 5 contained 50 p g of purified phosphohydrolase from the antibody-Sepharose column. 1145
1146
2‘,3‘-Cyclicnucleotide 3’-phosphohydrolase had been adsorbed, the initial sample was run a second or a third time through the adsorbent until no protein was eluted with the glycine-HCI wash. Antibody was pooled. dialyzed against NaHCO, buffer (0.1 M, pH 8.3), 0.5 M in NaCI, and coupled to cyanogen bromide-activated Sepharose in the same manner as that described for the enzyme. Immunoafinity chromatographic purijication of 2’,3‘-CN 3‘-ase. The phosphohydrolase was solubilized from an acetone powder of bovine white matter by guanidinium chloride extraction as previously described (DRUMMOND et a/., 1978). The only modification of that procedure was the addition of 1 mwbenzamidine hydrochloride to all dialysis solutions to act as an inhibitor of proteolysis. The crude extract, dialyzed free of most of the guanidinium chloride was the starting material for the purification outlined below. Crude extract was pumped through a column (0.9 x 10 cm) of the immunoadsorbent (2’,3’-CN 3I-ase-specific antibody coupled to Sepharose) at a flow rate of 0.5-1 ml/min. Phosphohydrolase activity in the effluent was periodically assayed to monitor the amount of enzyme being bound. When the enzyme activity of the effluent reached 10% of the activity of the influent the bed was washed with 0.05 M-MES (pH 6.4), 0.15 M in NaCl and 5% in glycerol. Once the U.V.absorption of the effluent had fallen to a stable value the bound protein was eluted with glycine-HCI buffer (0.2 M, pH 2.8) 0.5 M in NaCl and 5% in glycerol. The purified phosphohydrolase was eluted in a single symmetrical peak (Fig. 3). Effluent containing the phosphohydrolase was immediately dialyzed against 0.05 M-MES @H 6.4), 25% in glycerol. After regeneration of the adsorbent by washing with 25-50 ml of 0.05 M-MES (pH 6.4), 5% in glycerol, the column was again ready for use. The capacity of a column of this size is about 3000 units of enzyme; it takes four to five runs to process the extract obtained from 20 g of acetone powder. All chromatographic steps were performed at room temperature. Both the enzyme column and the antibody column were used 30 times without evidence of deterioration. Phosphate determination. All phosphohydrolase preparations to be analyzed for phosphate were first checked for homogeneity by SDS-gel electrophoresis under reducing conditions. Samples in which 95% of more of the protein was present in the doublet, as determined by densitometry on Fast Green-stained gels, were dialyzed against 1 1. of 0.05 M-potassium phosphate buffer (pH 6.8) for 12 h, 1 1. of 0.5 M-NaCI for 12 h and 4 x 1 1. of 0.1 M-acetic acid for 48 h. Dialyzed samples (5 mg each) were lyophilized
1147
E
C
0
ft
2In
0
15
75
90
I05
120
135
Effluent volume (ml)
FIG.3. Purification of 2’,3’-CN 3’-ase on antibody-Sepharose. A crude extract of bovine brain acetone powder (0.02 M in guanidinium chloride) was pumped through a column (0.9 x IOcm) of immunoadsorbent at a flow rate of 0.5 ml/min. The column was washed (at A) with 0.05 M-MES buffer (pH 6.4), 0.15 M in NaCl and 57, in glycerol. Purified phosphohydrolase was eluted from the adsorbent by washing the column (at B) with 0.2 M-glyCineHCI (PH 2.8), 0.5 M in NaCl and 5% in glycerol. Eluted protein was collected (t.) and immediately dialyzed against 0.05 M-MES buffer (pH 6.4), 25% in glycerol. in acid washed hydrolysis tubes. One milliliter of 6 N-HCI, 0.2% in phenol, was added to each tube. The tubes were evacuated, sealed and hydrolyzed at 110°C for 24 h. Aliquots of the hydrolysate were assayed for phosphate following the procedure of ITAYA& UI (1966). Protein concentration was determined by amino acid analysis. Ribonuclease was carried through the same dialysis procedure to serve as a control. RESULTS
Purification b y immunoafinity chromatography The purification of 2’,3’-CN 3’-ase by immunoadsorption is summarized in Table 1. The purity of the enzyme a t each step of the procedure, as determined by SDS-gel electrophoresis, is shown in Fig. 4; the significance of the doublet observable in the figure is discussed below. Essentially n o contaminating protein could be observed on SDS gels when 5 0 p g samples of the purified enzyme were analyzed. By this criterion, as well as by densitometric scans of Fast Green-stained gels, it was estimated that the preparations of phosphohydrolase were of greater than 95% purity. The overall recovery of enzymic activity was 57%. Specific activities of different preparations
FIG.5. SDS gradient (1&20% acrylamide) slab gel electrophoresis of proteins in gel slices cut from a previously run gradient slab gel which resolved the slower and the faster components of the enzyme preparation. These slices were loaded onto a second gradient SDS slab gel and run according to the procedure of CLEVELAND et al. (1977). Lanes 1 and 5 contained molecular weight markers. Lane 2 contained a single gel slice containing the initially slower moving protein. Lane 3 contained two gel slices containing the faster moving protein. Lane 4 contained two gel slices containing the faster moving component and a third gel slice containing the slower moving one. FIG.6. SDS slab gel electrophoresis of peptides from proteins in the upper and lower bands (Fig. 4). Phosphohydrolase components were previously resolved on SDS slab gels run under reducing conditions. Upper and lower protein bands were excised, transferred to a second slab gel, and peptide maps were generated by the procedure of CLEVELAND et af. (1977) as outlined in Materials and Methods. Lane 1 contained molecular weight markers. As a control lane 2 contained the peptides formed when 1Opg of bovine serum albumin in solution and 0.2pg of papain were submitted to electrophoresis. Lanes 3 and 4 contained upper and lower phosphohydrolase bands, respectively, plus 0.2pg of S . aureus V8 protease (-+ marks the band resulting from the protease). Lanes 5 and 6 contained upper and lower phosphohydrolase bands, respectively, plus 0.2 pg of papain. Lane 7 contained 0.2 pg of papain, a nondetectable amount of protein.
1 I48
ROBERTJ. DRUMMOND
TABLE1. PURlFlCATlON OF
2,3’-CYCLlC NUCLEOTIDE
Total activity Step 1. Crude extract
(I M in GuClt) 2. Supernatant after dialysis of extract to 0.2~-GuC1 3. Supernatant after dialysis of extract to 0.02 M-GuCI 4. Immunoadsorption
3’-PHOSPHOHYDROLASE BY
IMMUNOAFFINITY CHROMATOGRAPHY
(4
Total protein (ms)
Specific activity* (u/mg protein)
Recovery of activity
15,290
23 10
6.6
100
21,100
1620
13.0
> 100
13.680
840
16.3
89
8755
38
(%)
233
57
* Assay of HUGLIet al. (1973). t Soluble protein was first extracted from 20 g of acetone powder (approximately equivalent to
100 g of white matter) by aqueous buffer at pH 7.5. The insoluble pellet was extracted with 1 M-guanidinium chloride (see DRUMMOND et of., 1978).
of enzyme varied from a low value of approx 200u/mg of protein to a high value of nearly 300 u/mg of protein depending on the age of the acetone powder. It was observed that the total enzymic activity that could be extracted from 20g of acetone powder gradually declined from approx 20,000 units (freshly prepared acetone powder) to 10,OOO units after 3 months storage at 4°C. The phosphohydrolase preparations purified from freshly prepared acetone powder had the highest specific activities. Gel analyses under reducing conditions on preparations of enzyme with low or high specific activities revealed both preparations to consist of more than 95% of the doublet shown in lane 5, Fig. 4. When the immunoadsorbent-Sepharose is handled as previously described (See Materials and Methods) it has been found to be stable for up to 6 months with only minimal loss in binding capacity or specificity. However, we have encountered difficulty when attempting to employ the immunoadsorbents prepared from sera of bleedings taken 6 1 2 months after the initial inoculation. These sera have titres of from 1 :10,OOO to 1:2O,OOO. Immunoadsorbents prepared from such sera yield phosphohydrolase of approx 80% purity. Contaminating protein consists almost solely of the major low molecular weight protein seen with the crude extract sample in Fig. 4, lane 2. We believe that this contaminant is myelin basic protein because of its apparent molecular weight of approx 18,000 (EYLAR,1970) and the fact that it is the predominant protein when samples of purified myelin are run on SDS gels. Gel electrophoresis
In the course of these studies, gel electrophoresis under different conditions has established some special properties of the enzyme. As reported earlier (DRUMMOND et al., 1978), when the purified enzyme (1969) techis examined by the WEBER& OSBORN nique under non-reducing conditions, the enzyme migrates as a protein of about 98,000MW. The buffer in this instance is phosphate. However, when we conduct the electrophoresis without reducing agent using
the LAEMMLI (1970) system in Tris-glycine buffer, we obtain 75% of the protein in a band in the 44,000 MW range. With either system, when reduction and heat are used to carry the disaggregation to completion, all of the protein is in the lower molecular weight form. Thus this phosphohydrolase tends to remain aggregated in phosphate buffer, even with SDS present. A second observation concerns the nature of the two barely separable proteins observable in Fig. 4. The ratio of the amount of protein in the slower band to that in the faster one was generally about 2 : 1 when Fast Green-stained gels were scanned densitometrically. In order to characterize further the proteins in these two zones, the bands were excised from the slab TABLE2. AMINO ACID ANALYSES OF UPPER AND LOWER BANDS OF 2’,3’-CYCLIC NUCLEOTIDE 3’-PHOSPHOHYDROLASE FROM SDS-GELS* Amino acid Asx
Thr Ser Glx Pro GIY Ala Val Met Ile Leu TYr Phe His LYs ‘4%
Mole yo Upper band Lower band 8.24 k 0.58 5.40 f 0.13 6.57 & 0.31 13.7 f 0.52 4.59 k 0.35 8.94 f 0.28 7.58 k 0.49 5.13 0.01 1.80 f 0.31 2.16 f 0.19 12.3 -t 0.57 3.32 f 0.15 5.42 0.21 1.36 f 0.05 9.87 k 0.55 6.29 It: 0.54
+
8.51 f 0.16 5.05 f 0.43 5.95 f 0.37 13.9 0.35 4.61 + 0.32 8.58 k 0.20 7.54 + 0.02 5.14 It: 0.41 1.81 k 0.10 2.10 k 0.07 12.4 0.42 3.57 & 0.20 5.09 0.18 1.47 k0.15 10.0 f 0.43 6.22 _+ 0.09
*
* Twenty-four hour acid hydrolyses were performed on the upper and lower protein bands of 2‘,3’-cyclicnucleotide 3’-phosphohydrolase (prepared according to the method presented in this paper) after separation on SDS polyacrylamide slab gels run under reducing conditions (See Materials and Methods).Values are expressed as the mean f S.D. (three determinations each on upper and lower bands) and have not been corrected for the loss of labile amino acids. Half-cystine and tryptophan were not determined.
2’,3‘-Cyclic nucleotide 3’-phosphohydrolase
gel and resubmitted to electrophoresis (Fig. 5 ) ; each retained its characteristic mobility. To determine whether the smaller molecular weight component of the doublet might be a product of proteolysis taking place during purification, a modified purification protocol was adopted. The only difference between this method and the standard procedure was that benzamidine hydrochloride was not added to the dialysis solutions and all crude extracts (steps 1-3, Table 1) were brought to room temperature for at least 1 h. Gel analysis, under reducing conditions, of the enzyme thus purified showed the same relative amounts of the two proteins as were found for the enzyme prepared by the standard procedure. Amino acid analysis The next step was to see whether the two proteins differed in amino acid composition. They were resolved by gel electrophoresis under reducing conditions and amino acid analyses were performed on the excised gel slices. The amino acid compositions are given in Table 2. Because of relatively large blank gel corrections (10-30% of the uncorrected values) for certain amino acids, this procedure yielded corrected figures having standard deviations of 5-10%. The data permit the conclusion that the two proteins clearly have nearly the same amino acid composition. Peptide patterns
To compare structural features in further detail peptide patterns of each form were compared by SDS-polyacrylamide gel electrophoresis (Fig. 6). Control experiments were run with bovine serum albumin to show that specific enzymic clevages were occurring during the gel electrophoresis. As demonstrated by CLEVELAND et al. (1977), specific peptide patterns were observed when bovine serum albumin was subjected to electrophoresis in the presence of either papain or S. aureus V8 protease. Since we were not working with radioactively labeled protein we found it difficult to load sufficient amounts of protein in gel slices to yield observable amounts of peptides and still maintain the resolution required for this particular application. Hence, the gel analyses presented in Fig. 6 contained barely discernible peptides when stained with Coomassie Brilliant Blue. Although sample loads were not optimal it was found that a number of peptides of identical SDS-gel electrophoretic mobility were generated for the two phosphohydrolase components when either papain or S. aureus V8 protease were present. Some of the peptides produced by papain digestion were clearly distinct from those produced by S. aureus protease. Although the majority of the papain or S. aureus V8 peptides from each protein had the same mobilities (Nos. 1, 2, 3 and 4 for papain; Nos. 1, 2, 3, 4, 5, 6 and 9 for S. aureus V8) there were two unique peptides for each subunit (Nos. 5 and 6 for papain and Nos. 7 and 8 for S. aureus V8). Thus the two proteins are homologous, but differences exist.
1149
Phosphate determination
In order to determine whether we might be dealing with phosphoproteins, samples of the enzyme and ribonuclease (as a control) were exhaustively dialyzed against 0.5 M-NaCl and 0.1 M-acetic acid. Phosphate determinations on these samples showed less than 0.06 pmol of phosphate per pmol of ribonuclease and 0.294.48pnol of phosphate per pmol of phosphohydrolase subunits (three individual preparations of enzyme). The major protein of the phosphohydrolase doublet is clearly not a phosphoprotein and further work will be required to determine whether the small amount of phosphate detectable represents residual non-covalent binding of phosphate or organically bound phosphate in the minor protein. DISCUSSION
By employing the technique of immunoaffinity chromatography, 2’,3’-cyclic nucleotide 3’-phosphohydrolase can now be purified in 40mg amounts in half the time that it previously took to purify 4mg. The overall yield of protein has been increased from 7 to 57%. The initial investment of time and effort for the preparation of pure antigen by ion-exchange chromatography and related techniques (DRUMMOND et al., 1978) and for producing the antibody in rabbits is repaid many-fold when the enzyme is finally prepared from crude extract by the immunoaffinity technique. Antigen preparations employed in this study did not contain amounts of myelin basic protein detectable by gel-electrophoresis, but we conclude that very small amounts of this protein or a protein of identical SDSgel electrophoretic properties were present and gave rise to significant antibody titres after extended lengths of time, If early bleedings are employed for the preparation of affinity adsorbents. antibody against contaminating protein is not present in sufficient amounts to cause interference in the immunoaffinity purification. The recent experiments of DE JONG et a/. (1978) on SDS-gel electrophoresis bear upon the interpretation of the different mobilities of the proteins seen in the doublet in Fig. 4. DE JONG et al. show that changes as small as the substitution of one residue in a sequence can result in significant differences in mobility, probably as a result of changes in SDS binding. They recommend caution in the interpretation of differences in mobility in terms of molecular weight alone. Our own experience substantiates this view; relative to the same markers in Tris-glycine buffer we observe a 10% higher ‘apparent molecular weight’ for the phosphohydrolase in the Weber & Osborn system than in the Laemmli system. The two separable proteins seen in Fig. 4 may not differ as much in molecular weight as might be indicated by a conventional interpretation of the gel pattern. The agreement between the amino acid compositions does not leave much room for the deletion of a peptide fragment of appreciable size.
ROBERTJ. DRUMMOND
1I50
Further research will be required to establish the nature of the chemical differences in the proteins as evidenced by the peptide maps in Fig. 6. Sequence differences are a possibility; a covalently bound prosthetic group is another. The presence of measurable phosphate led us t o consider the possible presence of ADP-ribose; when the U.V.spectra of the enzyme a nd serum albumin were compared the phosphohydrolase was found t o have n o unique absorption a t 260 nm, which eliminates the possibility of the presence of a nucleotide. A unique crosslink in one of the proteins could also result in a slight difference in mobility in SDS-gels between two proteins which have essentially identical amino acid sequences. An isopeptide crosslink between the €-amino group of lysine and a glycine carboxyl group results in a branched structure linking nonhistone and histone 2A polypeptides (GOLDKNOPF & BUSCH, 1977). E-(y-glutamy1)-Lysine crosslinks between fibrin subunits have also been characterized (PISANO et al., 1971). Although these are less probable explanations, crosslinks of this nature could exist and give rise to proteins and peptides of different SDS-gel mobility. The increased availability of the purified protein should make possible more detailed studies of its chemistry and the use of immunological methods for its subcellular localization. Acknowledgements-This research was supported in part by USPHS grants GM 07256 and GM 25323 and an R. J. Reynolds Post-Doctoral Fellowship. The author is indebted to GEORGE KUZMYCZ of EDWARD REICH’S laboratory for assistance in the preparation of antisera, to STANFORD MOOREand WILLIAM H. STEINfor cooperation in CURthe organization of this manuscript, and to BARBARA TOPELLE for the preparation of the typescript. REFERENCES AVRAMEASS., TAUDOU B. & CHUILON S. (1969) Glutaraldehyde, cyanuric chloride and tetraazotized o-dianisidine as coupling reagents in the passive hemagglutination test. Immunochemistry 6, 67-76. BRAUN P. E. & BARCHI R. L. (1972) 2’,3’-Cyclicnucleotide 3’-phosphodiesterase in the nervous system. Electrophoretic properties and developmental studies. Brain Res. 40,437444. S. G., KIRSCHNER M. W. & CLEVELAND D. W., FISCHER LAEMMLI U. K. (1977) Peptide mapping by limited proteolysis in sodium dodecyl sulfate and analysis by gel electrophoresis. J . b i d . Chem. 252, 1102-1 106. DE JONC W. W., ZWEERSA. & COHENL. H. (1978) Influence of single amino acid substitutions on electrophoretic mobility of sodium dodecyl sulfate-protein complexes. Biochem. biophys. Res. Commun. 82, 532-539. DRUMMOND G. I., IYERN. T . & KEITHJ . (1962) Hydrolysis of ribonucleoside 2’,3’-cyclicphosphates by a diesterase from brain. J . biol. Chem. 237, 3535-3539. E. B. & GUHAA. (1978) PurificaDRUMMOND R. J., HAMILL tion and comparison of 2‘,3’-cyclic nucleotide 3‘-phosphohydrolases from bovine brain and spinal cord. J . Neurochem. 31, 871-878. ENCELE. L. & WOODJ. G. (1976) Cytochemical localiza-
tion of 2’,3‘-cyclic nucleotide 3‘-phosphohydrolase in CNS white matter. Neurosci. Abstr. 2, 410. EYLARE. H. (1970) Amino acid sequence of the basic protein of the myelin membrane. Proc. natn. Arad. Sci., U.S.A. 67, 1425-1431. GOLDKNOPF I. L. & BUSCHH. (1977) Isopeptide linkage between nonhistone and histone 2A polypeptides of chromosomal conjugate-protein A24. Proc. natn. Acad. Sci., U.S.A. 74, 864-868. HOUSTON L. L. (1971)Amino acid analysis of stained bands from polyacrylamide gels. Analyt. Biocltem. 44, 81-88. HUGLIT. E., BUSTINM. & MCORES. (1973) Spectrophotometric assay of 2’.3’-cyclicnucleotide 3’-phosphohydrolase: application to enzyme in bovine brain. Brain Res. 58, 191-203. ITAYAK. & UI M. (1966) A new micromethod for the colorimetric determination of inorganic phosphate. Clinica chim. Acta 14, 361-366. Y. (1967) The regional and subKURIHARA T. & TSUKADA cellular distribution of 2’,3‘-cyclicnucleotide 3’-phosphohydrolase in the central nervous system. J . Netrrochem. 14, 1167-1174. KURIHARA T. & TSUKADA Y. (1968) 2‘,3’-Cyclicnucleotide 3‘-phosphohydrolase in the developing chick brain and spinal cord. J . Neurochem. 15, 827-832. KURIHARA T., NUSSBAUM J. L. & MANDELP. (1970) 2’,3’-Cyclicnucleotide 3’-phosphohydrolase in brains of mutant mice with deficient myelination. J . Neurochem. 17, 993-997. LAEMMLI U. K. (1970) Cleavage of structural proteins during the assembly of the head or bacteriophage T4. Nature, Lond. 227, 680-6235, LOWRY0. H., ROSEBROUCHN. J., FARRA. L. & RANDALL R. J. (1951) Protein measurement with the Fohn phenol reagent. J . biol. Cheni. 193, 265-215. OLAFSON R. W., DRUMMOND G. I. & LEE J. F. (1969) Studies on 2’,3’-cyclic nucleotide-3’-phosphohydrolase from brain. Can. J . Biocheni. 47, 961-966. OUCHTERLONY D. (1953) Antigen-antibody reactions in gels. Acta path. scand. 32. 231-240. PISANO J. J., FINLAYSON J. S., PEYTON M. P. & NAGAI Y.(1971)e-(y-glutamyl) lysine in fibrin: lack of crosslink formation in Factor XI11 deficiency. Proc. narn. Acad. Sci., U S A . 68, 77e712. PODUSLO S. E. (1975) The isolation and characterization of a plasma membrane and a myelin fraction derived from oligodendroglia of calf brain. J . Neurochem. 24, 647-654. PODUSLO S. E. & NORTONW. T. (1972) Isolation and some chemical properties of oligodendroglia from calf brain. J . Neurochem. 19, 727-736. M. R., SHAPIRA R., MOBLEY W. C., THIELES. B., WILHELMI WALLACEA. & KIBLERR. F. (1978) Localization of 2,3’-cyclic nucleotide-3’-phosphohydrolase of rabbit brain by sedimentation in a continuous sucrose gradient. J . Neurochem. 30, 735-744. S. E. (1975) SUNDARRAJ N., SCHACHNER M. & PFEIFF~R Biochemically differentiated mouse glial lines carrying a nervous system specific cell surface antigen (NS-I). Proc. natn. Acad. Sci., U.S.A. 12, 1927-1931. WEBERK. & OSBORNM. (1969) Reliability of molecular weight determinations by dodecyl sulfate polyacrylamide gel electrophoresis. J . biol. Chem. 244, 4406-4412. ZANETTA J. P., BENDAP., GOMBOSG. & MORGANI. G. (1972) The presence of 2‘,3’-cyclicAMP 3’-phosphohydrolase in glial cells in tissue culture. J . Neurochem. 19, 881-883.
