427 (1976) 428 442 ,:~} Elsevier Scientific Publishing Company, Amsterdam Biochimica et Biophysica Acre,
Printedin The Netherlands,
BBA 37293 T H E ISOLATION AND C H A R A C T E R I Z A T I O N OF THE GLYCOPEPTII.)ES FROM HORSERADISH PEROXIDASE ISOENZYME (7"
JOYCE CLARKE and LELAND M. SHANNON Department q / Biochemistw, University ~*1'Cal~/brnkl, Riverside, Cali[~ 92502 i U.S.A. J
(Received July 25th, 1975)
SUMMARY The purity of horseradish peroxidase isoenzyme C was demonstrated using isoelectric focusing, polyacrylamide gel electrophoresis at two pH values and cellulose acetate electrophoresis at two pH values. The glycopeptides obtained upon trypsin digestion were isolated using the plant lectin, concanavalin A, and were resolved using paper electrophoresis. The carbohydrate content of the native peroxidase was 86~o accounted for by the carbohydrate content of the glycopeptides thus suggesting little loss of carbohydrate during glycopeptide isolation and purilication In each of the seven glycopeptides isolated glucosamine was associated with asparagine, thus suggesting the carbohydrate chains are covalently bound to the peptide chain through N-glycosidic linkages. The purity of each glycopeptide was demonstrated by the sequential release of single amino acid residues by Edman degradation. As six glycopeptides had unique amino acid sequences, it was concluded that the carbohydrate prosthetic group was distributed in at least six units along the protein backbone. Five glycopeptides possessed the amino acid sequence about the point of carbohydrate attachment of Asn-X (Set, Thr) where X is any amino acid. The size of the carbohydrate units ranged from 1600 to 3000 daltons. The predominant carbohydrate residues in each glycopeptide were mannose and glucosamine with lesser and varying amounts ol" fucose, xylosc, and arabinose. There was no apparent correlation of the carbohydrate composition with the amino acid ~;equence.
INTRODUCTION Horseradish peroxidase is a haem-containing glycoprotein of molecular weight 40 000. The crude extract from the horseradish plant contains seven major peroxidase isoenzymes [1] and up to 13 additional minor isoenzymes [2]. The seven major isoenzymes have been highly purified and characterized by Shannon's laboratory with respect to chromatographic behavior, electrophoretic mobility, haemin characterization, and spectrophotometric properties [1] catalytic properties [3] circular dichroism [4], amino acid composition, tryptic peptide maps, and disulfide linkages [5], carbohydrate composition [6], temperal relationship between the synthesis of the polypeptide unit and the assembly of the carbohydrate unit [7], and the properties of the N-acetylglucosaminyltransferase [8]. On the bases of these studies the seven
429 major peroxidase isoenzymes may be categorized into three groups: one group A-1 and A-2; one group A-3; and one group B, C, D and E. The present investigation was undertaken to characterize the carbohydrate prosthetic groups of the most abundant isoenzyme, isoenzyme C, with respect to the number of sites of carbohydrate attachment, the amino acid sequence about the sites of attachment, and the complement of sugar residues in each carbohydrate chain. Characterization of the carbohydrate prosthetic group is a necessary prerequisite to elucidate the structural relationships amongst the isoenzymes. An understanding of the structural relationships may provide an insight into the mode of peroxidase isoenzyme biogenesis and thereby assist in establishing the significance of multiple forms of peroxidase. While the present work was in progress Welinder et al. [9-11 ] reported sequence information and carbohydrate attachment sites in tryptic and thermolytic peptides of commercial horseradish peroxidase. The present results confirm the amino acid sequence in 3 of the 8 glycopeptides reported by Welinder. Possible reasons for the discrepancies are discussed. MATERIALS AND METHODS
Isolation ofperoxidase isoenzyme C. Peroxidase isoenzymes were isolated from horseradish roots according to the method of Shannon et al. [1]. Isoenzyme C was separated from isoenzyme B using a modified elution gradient from the CM-cellulose chromatographic column, Fig. 1. The final purification was achieved by isoelectric focusing using an LKB 8100 electrofocusing column (LKB-Produkter AB, S-161, 25 Bromma, Sweden) according to procedures outlined by Vesterberg [12]. A pH 7-10 gradient was used and 20 mg of isoenzyme C were focused at 400 V at room temperature for 8 days. Fractions of approx. 1 ml were collected. After determining the absorption of each fraction at 402 nm and measuring the pH, the appropriate fractions were pooled and extensively dialysed against deionized water. Polyaerylamide electrophoresis. Electrophoresis at pH 4.5 was carried out according to Reisfeld et al. [13]. Electrophoresis at pH 10.6 employed glycine/NaOH buffer, 0.06 M. 4 ~ gels were prepared and polymerized with (NH4)28208. Gels stained for protein were immersed in 10 ~ trichloroacetic acid for 20 min, transferred to 0.05 ~ Coomassie brilliant blue in l0 ~ trichloroacetic acid for 12 h, and destained in 10~ trichloroacetic acid. Peroxidase activity was detected with benzidine/H202 according to the method of Brewer [14]. Cellulose acetate electrophoresis. The sample was applied to cellulose acetate strips equilibrated with electrode buffer and electrophoresis was run for 2 h at 4 °C with an applied voltage of 300 V. The electrode buffers used were 0.06 M sodium acetate, pH 5.4, or 0.06 M Tris, pH 8.5. After electrophoresis peroxidase was detected by immersing the strips in a solution of O-dianisidine. The O-dianisidine was prepared by dissolving 97.6 mg O-dianisidine in 70 ml 95 ~ ethanol. To this solution was added 15.5 ml 0.3 M sodium acetate, pH 5.4, and immediately prior to staining 4.4 ml 0.1 M HzOz were added. Tryptic digestion. The haem was removed from isoenzyme C by slowly adding 10 volumes of --20 °C acidic acetone (0.5 ~ HC1, v/v). The white apoprotein pellet was resuspended in --20 °C acidic acetone and the procedure repeated. The pellet
430 was then washed twice with 20 ' C acetone to remove traces of HCI. Tile apoprotein was dissolved in I ml 0. I M NH4HCO3, pH 8.0. Trypsin which had been treated with L-(l-tosyl-amide-2-phenyl) ethyl chloromethyl ketone was dissolved in I mM HCI and the incubation carried out for 6 h at 37 C , trypsin being added at intervals such that the final ratio of trypsin:apoprotein was I:1 (w/w). Isolation of tryptic glycopeptides, Concanavalin A, twice crystaltised from Miles-Yeda Ltd., was used to isolate the glycopeptides. Alter tryptic digestion of apoenzyme C, and inactivation of the trypsin by heating, the digest was adjusted to pH 6.0 and incubated at room temperature with concanavalin A such li'_,at tile ratio concanavalin A:glycopeptide was 100:1 (w/w). Alter 12 h the incubation mixture was transferred to an Amicon Diaflo ultrafittration apparatus fitted with a PM-10 membrane, which retains material with molecular weights exceeding t0 000. The mixture was subjected to ultrafiltration and was washed with three volumes ot" water. To the material retained by the membrane was added 1 ml 0.01 M methyl~t-mannopyranoside. After 15 min the material was washed with three \olumes of water and the material passing through the membrane was lyophilised. The lyophilisate was dissolved in 1 ml water and applied to an 18 cm 1.7 cm Biogel P-2 column on which the glycopeptides were separated from excess methyl-~-mannopyranoside. The fractions containing the glycopeptides were pooled, lyophilised, and analyzed for carbohydrate by gas-liquid chromatography. Paper electrophoresis. The lyophilised glycopeptide fraction was dissolved in water and applied to Whatman 3 MM paper. The paper was thoroughly wetted by spraying with pyridine/acetic acid buffer (pyridine/acetic acid/water, 1:10:289 v/v), pH 3.6. The electrophoresis was carried out in a Savant electrophoresis tank at 2000 V, 150 mA, 5 ':'C for 75 rain. The paper was dried and the peptides were located by means of marker strips sprayed with ninhydrin. The glycopeptides were quantitatively eluted from the paper and lyophilized. Pronase digestion. The sample to be digested was dissolved m [ mJ 0.1 M Tris. HCI, pH 7.5, containing 0.03 M CaCI,, To this solution was added pronasc from Streptomyces griseus (Sigma Chemical Co., St. Louis, Mo.), such that the ratio of pronase:sample was 1:1 (w/w). The incubation was carried out at room temperature for 72 h in an atmosphere of toluene to prevent bacterial growth. Reduction and carboxymethylation. The sample was dissolved in I ml 0.1 M NH4HCOa, pH 8.0. Reduction and carboxymethylation was carried out as described by Shih et al. [5]. E&nan degradation. Semi micro-degradation was carried out according to the procedure of Peterson et al. [15]. Reagents and solvents were purchased from Pierce Chemical Co. The degradation was carried out using the manual method in an atmosphere of N> Gas-liquid chromatography was employed t\)r identification or" the phenylthiocarbamoyl amino acids. The support system was 10'},~ SP-400 Chromosorb W-H-P, 100/120 mesh packed into a 180 cm :, 2 m m glass column. The carrier gas was N2 and detection was by flame ionisation. The temperature program was: initial temperature 185 ' C for 4 min and a rise of 4 'C/rain to a maximum temperature of 285 'C. The Ptc amino acids were sitylated in the injection port with N,O-bis(trimethylsilyl)-acetamide. With the exception of glycine and threonine all the Pie amino acids could be satisfactorily resolved. Glycine and threonine were separated by thin-layer chromatography using plates of Eastman 6060 silica gel. The plates were
431 developed using 2-propanol/xylene (10:35, v/v) and the Ptc amino acid visualised under ultraviolet light and the Re compared with authentic Ptc amino acids run simultaneously. Amino acid and glucosamine analysis. The samples for amino acid analysis were sealed under Nz and hydrolyzed in 5.7 M HCI at 110 °C for 20 h. After hydrolysis the HCI was removed by lyophilisation. Amino acid analysis was performed on a Beckman 120B amino acid analyser adapted to extend the sensitivity range to 2 nmol. The site of carbohydrate attachment was determined by analysing for glucosamine those samples of Ptc amino acids which could provide sites for carbohydrate attachment, i.e. aspartic acid, asparagine, threonine and serine. The samples for glucosamine analysis were sealed under Nz, hydrolyzed in 4 M HC1 at 100 °C for 4 h and analysed using the amino acid analyser. Carbohydrate analysis. Quantitative analysis was achieved by gas-liquid chromatography of the alditol acetate derivatives of the carbohydrate. The method of Kim et al. [16] was used for the preparation of the alditol acetate derivatives. Gasliquid chromatography was carried out as described by Partridge et al. [17]. A modification of the Molisch test as described by Dubois et al. [18] was employed to monitor carbohydrate in the effluent from chromatography columns. To an aliquot was added 0.5 ml 5% phenol and 3 ml concentrated H2SO4. The solution was vigorously agitated and allowed to stand for 15 min before reading the absorbance at 490 nm. Mannose was employed as the standard for sugar estimation. RESULTS AND DISCUSSION
Purification of peroxidase isoenzyme C Separation of peroxidase isoenzyme C from B is difficult due to the close similarity of their properties [1-8]. In the present paper an improved separation was achieved using a modified elution gradient from the CM-cellulose column (Fig. 1). B
B/C
08 O075
o.o65
0.6
o.o 5 0.2
~:
I
~, 0
667
687
, 707
, 727
747
"
]0.045
767
Fraction number
Fig. 1. Elution profile of peroxidase isoenzymes B and C from a CM-cellulose chromatographic column. The column, 40 cm × 2.5 cm, was equilibrated with 0.005 M sodium acetate, pH 4.5. A gentle elution gradient was developed using the Varigrad, Buchler Instruments Co., Fort Lee, N.J. The first chamber contained 500 ml 0.01 M sodium acetate, pH 4.5; the middle chamber contained 500 ml 0.06 M sodium acetate, pH 4.5; the terminal chamber contained 500 ml 0.1 M sodium acetate, pH 4.5. Flow rate was 10 ml/h and 1.5-ml fractions were collected and monitored for peroxidase by measuring the absorbance at 402 nm. Molarity of the effluent solution was determined using a conductivity meter.
432 pH 8 8
1
10
I
Q8
r
=e 0.6
I
'l
,, ,i
04
/ ~,~
),
!
/
Q2 F
/
0
20
pH 72
'
i
~
40
60
80
Fraction
100
~, ° 120
number
Fig. 2. Elution profile of peroxJdase isoenzyme C from an isoelectric focusing column.
lsoenzyme C obtained from the CM-cellulose column was further purified using isoelectric focusing. The elution profile alter 8 days, when no further migration of the peroxidase visible bands was observed, is shown in Fig. 2. The isoelectric point of the major peak corresponds with that reported by Welinder et al. [9] and is consistent with the high lysine and arginine content of isoenzyme C and with the absence of sialic acid [1]. Subsequent work was done on protein collected from the major peak of the isoelectric t\)cusing column. Criteria
of homogeneity
The homogeneity of peroxidase isoenzyme C was demonstrated by electrophoresis on polyacrylamide gels at pH 4.5 and 10.6 and on cellulose acetate strips at pH 5.6 and 8.3 (Fig. 3). Approx. 60 fig of peroxidase isoenzyme C was applied to each polyacrylamide gel. Electrophoresis at pH 4.5 or 10.6 showed one protein band when stained with Coomassie brilliant blue and one activity band when stained for peroxidase. The protein band and peroxidase activity band were coincident. Cellulose A
Top
B
~
m
m
B C
- -
DIM 4.5
7
DIM 10,6
.
DH 5.4
B C
Origin
÷
pH 8 3
Fig. 3. (A) Acrylamide gel electrophoresis of peroxidase isoenzyme C at pH 4.5 and 10.6. (B) Cellulose acetate electrophoresis of peroxidase isoenzyme B and C at pH 5.4 and 8.3.
433 acetate strip electrophoresis at pH 5.4 and 8.3 showed peroxidase isoenzyme C was devoid of the closely related isoenzyme B, as well as other anodic and cathodic peroxidase isoenzymes. The above observations indicate that peroxidase isoenzyme C used in these studies is highly purified and appears to be devoid of other peroxidase isoenzymes as well as contaminating proteins. The R.Z. (ratio of haem absorbance at 402 nm: protein absorbance at 280 nm) of this preparation was 3.5. The carbohydrate composition of horseradish peroxidase C used in these studies is listed in Table V.
Isolation of glycopeptides In the present study trypsin was employed to obtain the glycopeptides because of two advantages: the problem of microheterogeneity of the glycopeptides with respect to their amino acid composition is avoided since trypsin is specific in hydrolyzing peptide bonds on the carboxyl side of lysine or arginine; as each glycopeptide should contain only one residue of arginine or lysine, these residues can act as internal standards to correct for preparative losses. Concurrent work in this laboratory showed that concanavalin A, a plant lectin from jack bean, would precipitate peroxidase isoenzyme C from solution. Peroxidase could be solubilized from the precipitate by the addition of 0.01 M methyl-a-mannopyranoside which competes with peroxidase for the carbohydrate binding sites on concanavalin A. These observations suggested that concanavalin A might be useful to exclusively separate glycopeptides from non-glycopeptides. To test this hypothesis a tryptic digest of isoenzyme C was incubated with concanavalin A and then subjected to ultrafiltration using a PM-10 membrane which retains material with molecular weight above 10 000. Gas-liquid chromatographic analysis of the alditol acetate derivatives of sugar residues in material retained by the PM-10 membrane showed approx. 80 ~ of the sugar initially present on the native enzyme was retained by the membrane (Table I). When concanavalin A was omitted from the mixture all the carbohydrate passed through the membrane. These observations indicate that the glycopeptides are bound to concanavalin A and can be effectively separated from peptides which, having molecular weights of less than 10 000, pass through the memTABLEI S U G A R ANALYSIS OF STARTING MATERIAL, CONCANAVALIN A GLYCOPEPTIDE RETAINED BY PM-10 MEMBRANE, AND OF GLYCOPEPTIDE WHICH PASSED T H R O U G H PM-10 M E M B R A N E AFTER TREATMENT OF RETAINED C O N C A N A V A L I N A GLYCOPEPTIDE LIGAND WITH 0.01 M METHYL-ct-MANNOPYRANOSIDE 1 mg horseradish peroxidase C was employed. The values represent the/~g carbohydrate present in the different fractions.
Starting material* PM- 10 retained PM-10 Pass after methyl-a-mannopyranoslde
Fucose
Arabinose
Xylose
Mannose
32 -- **
4 3
29 23
130 109
29
2
25
110
* Peroxidase isoenzyme C before tryptic digestion. ** Not determined as the fucose peak was obscured by an unidentified peak from concanavalin A.
434
0 o o
0 o
0 0
o 0
0
0
0
04 0
0 0
03 02
0 0 0
01
7o
Gp
A
i
B
o 0 o
~-C
Fig. 4. Paper electrophoresis of a tryptic digest of peroxidase isoenzyrne C (A), of the tryptic glycopeptides (B), and of the tryptic digest after removal of the glycopeptides (C),
brane. Having established that the glycopeptides bind to concanavalin A, an experiment was designed to determine the extent of release of the glycopeptides by methyl~-mannopyranoside. To the material retained by the membrane was added 0.01 M methyl-ct-mannopyranoside. Following ultrafiltration, the material passing through the PM-10 membrane was analyzed for carbohydrate. The results, shown in Table l, indicated that methyl-rz-mannopyranoside quantitatively released the glycopeptides from concanavalin A. Furtherlnore, over 85 o{, of the carbohydrate in native peroxidase isoenzyme C was recovered in the isolated tryptic glycopeptides. The tryptic glycopeptides were fractionated using electrophoresis. The complete tryptic digest and peptides remaining after removal of the glycopeptides were also electrophoresed for comparison. The results are shown in Fig. 4. From the electrophoretic patterns it was concluded that not only were the glycopeptides ,~:eparated from each other, but also that the glycopeptides were effectively separated front the non-carbohydrate-containing peptides. The purity of each glycopeptide was determined by N-terminal analysis and Edman degradation.
Characterization o! the glycopeptides Amino acid analysis of the tryptic glycopeptides (data not presented) showed glycopeptides Nos. 1-5 were large and contained at least 20 amino acid residues per residue of arginine. Tryptic glycopeptide No. 6 contained only five amino acid residues per arginine residue. In order to obtain peptides of a size suitable to manual
435
Q9 Q8 Q7 E c
g
q
Q6 0.5 Q4 03 Q2 0.1 O
~
~'
40
60
80
Effluent
100
120
(ml)
Fig. 5. Elution profile of pronase glycopeptides from a Sephadex 0-25 column. Elution from a 1.5 c m x 100 cm Sephadex G-25 column equilibrated with 0.2 M acetic acid. Fractions of 2 ml were collected and the glycopeptides detected by the Molisch test as modified by Dubois et al. [18] and measured at 490 nm (O---O). Pronase and non glycopeptides, 280 nm (0--(-3).
Edman degradation, the tryptic glycopeptides, with the exception of No. 6, were separately digested with pronase. The pronase glycopeptides were separated from pronase and low molecular weight material on a previously calibrated Sephadex G-25 column. The elution pattern is shown in Fig. 5. No glycopeptides were detected when TABLE II AMINO ACID COMPOSITION OF PRONASE GLYCOPEPTIDES 1, 2, 3, 3A, 3B, 4, 4R, 5 AND TRYPTIC GLYCOPEPTIDE 6 Values are normalised to that amino acid marked 1 or 2 which was shown to be present as one or two residues in the tryptic glycopeptide from which the pronase glycopeptide was derived. The results are expressed as mol amino acid/mol glycopeptide. Peptide
Asp Thr Ser Pro Gly Ala Val Cys Leu Ile Phe Arg
1
2
3*
3A
3B
4**
4R
5
6
2.0 0.7 0.8 0.8 1.0 1 0.0 0.0 0.0 0.9 0.0 0.0
1.9 0.8 0.7 0.0 1 0.0 0.0 0.0 0.0 0.0 1.1 0.0
3.1 2.6 1.1 0.7 0.9 1 0.0 0.0 1.8 0.0 0.0 0.0
2.1 0.7 0.0 0.0 0.0 0.0 0.0 0.0 0.9 0.0 0.0 0.0
0.9 1.9 0.9 1.1 0.9 1 0.0 0.0 0.8 0.0 0.0 0.0
2.0 0.0 1.9 2.1 0.0 0.8 0.0 2 0.0 0.0 0.0 1.1
2.0 0.0 1.1 0.9 0.0 0.0 0.0 1 0,0 0.0 0.0 0.0
1.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1 0.0 0.0 0.0
1.8 0.0 0.0 0.0 1.0 0.0 0.9 0.0 1.0 0.0 0.0 1
* Glycopeptide 3 was resolved into two glycopeptides, 3A and 3B, by repeating the electrophoresis step. ** Since pronase glycopeptide 4 contained half cystine, it was reduced and carboxymethylated and reapplied to the Sephadex G-25 column. The resulting glycopeptide was designated 4R.
436 pronase was incubated in the absence of tryptic glycopeptides, indicating that pronase did not contribute to the glycopeptide pool. Pronase treatment yielded glycopeptides of a size suitable to the manual Edman degradation as indicated by, tile amino acid analysis of the glycopeptides shown in Table II. All glycopeptides contained amino acids commonly associated with the N-glycosidic carbohydrate protein linkage, i.e. aspartic acid, threonine and serine. As glycopeptide No. 4 contained half cystine it was reduced and carboxymethylated and reapplied to tile Sephadex G-25 column. Pronase glycopeptide No. 4 after reduction and carboxymethylation, is designated 4R. Gas-liquid chromatography was chosen to identify the Ptc amino acids because the sensitivity of the method permitted quantities as small as I nmol to be detected. All the Ptc amino acids were adequately resolved with tile exception of threonine and glycine which could be separated by thin-layer chromatography. Upon gas-liquid chromatographic analysis threonine and serine usually appeared as two peaks, ,~serine (phenylthiocarbamoyl of
Printedin The Netherlands,
BBA 37293 T H E ISOLATION AND C H A R A C T E R I Z A T I O N OF THE GLYCOPEPTII.)ES FROM HORSERADISH PEROXIDASE ISOENZYME (7"
JOYCE CLARKE and LELAND M. SHANNON Department q / Biochemistw, University ~*1'Cal~/brnkl, Riverside, Cali[~ 92502 i U.S.A. J
(Received July 25th, 1975)
SUMMARY The purity of horseradish peroxidase isoenzyme C was demonstrated using isoelectric focusing, polyacrylamide gel electrophoresis at two pH values and cellulose acetate electrophoresis at two pH values. The glycopeptides obtained upon trypsin digestion were isolated using the plant lectin, concanavalin A, and were resolved using paper electrophoresis. The carbohydrate content of the native peroxidase was 86~o accounted for by the carbohydrate content of the glycopeptides thus suggesting little loss of carbohydrate during glycopeptide isolation and purilication In each of the seven glycopeptides isolated glucosamine was associated with asparagine, thus suggesting the carbohydrate chains are covalently bound to the peptide chain through N-glycosidic linkages. The purity of each glycopeptide was demonstrated by the sequential release of single amino acid residues by Edman degradation. As six glycopeptides had unique amino acid sequences, it was concluded that the carbohydrate prosthetic group was distributed in at least six units along the protein backbone. Five glycopeptides possessed the amino acid sequence about the point of carbohydrate attachment of Asn-X (Set, Thr) where X is any amino acid. The size of the carbohydrate units ranged from 1600 to 3000 daltons. The predominant carbohydrate residues in each glycopeptide were mannose and glucosamine with lesser and varying amounts ol" fucose, xylosc, and arabinose. There was no apparent correlation of the carbohydrate composition with the amino acid ~;equence.
INTRODUCTION Horseradish peroxidase is a haem-containing glycoprotein of molecular weight 40 000. The crude extract from the horseradish plant contains seven major peroxidase isoenzymes [1] and up to 13 additional minor isoenzymes [2]. The seven major isoenzymes have been highly purified and characterized by Shannon's laboratory with respect to chromatographic behavior, electrophoretic mobility, haemin characterization, and spectrophotometric properties [1] catalytic properties [3] circular dichroism [4], amino acid composition, tryptic peptide maps, and disulfide linkages [5], carbohydrate composition [6], temperal relationship between the synthesis of the polypeptide unit and the assembly of the carbohydrate unit [7], and the properties of the N-acetylglucosaminyltransferase [8]. On the bases of these studies the seven
429 major peroxidase isoenzymes may be categorized into three groups: one group A-1 and A-2; one group A-3; and one group B, C, D and E. The present investigation was undertaken to characterize the carbohydrate prosthetic groups of the most abundant isoenzyme, isoenzyme C, with respect to the number of sites of carbohydrate attachment, the amino acid sequence about the sites of attachment, and the complement of sugar residues in each carbohydrate chain. Characterization of the carbohydrate prosthetic group is a necessary prerequisite to elucidate the structural relationships amongst the isoenzymes. An understanding of the structural relationships may provide an insight into the mode of peroxidase isoenzyme biogenesis and thereby assist in establishing the significance of multiple forms of peroxidase. While the present work was in progress Welinder et al. [9-11 ] reported sequence information and carbohydrate attachment sites in tryptic and thermolytic peptides of commercial horseradish peroxidase. The present results confirm the amino acid sequence in 3 of the 8 glycopeptides reported by Welinder. Possible reasons for the discrepancies are discussed. MATERIALS AND METHODS
Isolation ofperoxidase isoenzyme C. Peroxidase isoenzymes were isolated from horseradish roots according to the method of Shannon et al. [1]. Isoenzyme C was separated from isoenzyme B using a modified elution gradient from the CM-cellulose chromatographic column, Fig. 1. The final purification was achieved by isoelectric focusing using an LKB 8100 electrofocusing column (LKB-Produkter AB, S-161, 25 Bromma, Sweden) according to procedures outlined by Vesterberg [12]. A pH 7-10 gradient was used and 20 mg of isoenzyme C were focused at 400 V at room temperature for 8 days. Fractions of approx. 1 ml were collected. After determining the absorption of each fraction at 402 nm and measuring the pH, the appropriate fractions were pooled and extensively dialysed against deionized water. Polyaerylamide electrophoresis. Electrophoresis at pH 4.5 was carried out according to Reisfeld et al. [13]. Electrophoresis at pH 10.6 employed glycine/NaOH buffer, 0.06 M. 4 ~ gels were prepared and polymerized with (NH4)28208. Gels stained for protein were immersed in 10 ~ trichloroacetic acid for 20 min, transferred to 0.05 ~ Coomassie brilliant blue in l0 ~ trichloroacetic acid for 12 h, and destained in 10~ trichloroacetic acid. Peroxidase activity was detected with benzidine/H202 according to the method of Brewer [14]. Cellulose acetate electrophoresis. The sample was applied to cellulose acetate strips equilibrated with electrode buffer and electrophoresis was run for 2 h at 4 °C with an applied voltage of 300 V. The electrode buffers used were 0.06 M sodium acetate, pH 5.4, or 0.06 M Tris, pH 8.5. After electrophoresis peroxidase was detected by immersing the strips in a solution of O-dianisidine. The O-dianisidine was prepared by dissolving 97.6 mg O-dianisidine in 70 ml 95 ~ ethanol. To this solution was added 15.5 ml 0.3 M sodium acetate, pH 5.4, and immediately prior to staining 4.4 ml 0.1 M HzOz were added. Tryptic digestion. The haem was removed from isoenzyme C by slowly adding 10 volumes of --20 °C acidic acetone (0.5 ~ HC1, v/v). The white apoprotein pellet was resuspended in --20 °C acidic acetone and the procedure repeated. The pellet
430 was then washed twice with 20 ' C acetone to remove traces of HCI. Tile apoprotein was dissolved in I ml 0. I M NH4HCO3, pH 8.0. Trypsin which had been treated with L-(l-tosyl-amide-2-phenyl) ethyl chloromethyl ketone was dissolved in I mM HCI and the incubation carried out for 6 h at 37 C , trypsin being added at intervals such that the final ratio of trypsin:apoprotein was I:1 (w/w). Isolation of tryptic glycopeptides, Concanavalin A, twice crystaltised from Miles-Yeda Ltd., was used to isolate the glycopeptides. Alter tryptic digestion of apoenzyme C, and inactivation of the trypsin by heating, the digest was adjusted to pH 6.0 and incubated at room temperature with concanavalin A such li'_,at tile ratio concanavalin A:glycopeptide was 100:1 (w/w). Alter 12 h the incubation mixture was transferred to an Amicon Diaflo ultrafittration apparatus fitted with a PM-10 membrane, which retains material with molecular weights exceeding t0 000. The mixture was subjected to ultrafiltration and was washed with three volumes ot" water. To the material retained by the membrane was added 1 ml 0.01 M methyl~t-mannopyranoside. After 15 min the material was washed with three \olumes of water and the material passing through the membrane was lyophilised. The lyophilisate was dissolved in 1 ml water and applied to an 18 cm 1.7 cm Biogel P-2 column on which the glycopeptides were separated from excess methyl-~-mannopyranoside. The fractions containing the glycopeptides were pooled, lyophilised, and analyzed for carbohydrate by gas-liquid chromatography. Paper electrophoresis. The lyophilised glycopeptide fraction was dissolved in water and applied to Whatman 3 MM paper. The paper was thoroughly wetted by spraying with pyridine/acetic acid buffer (pyridine/acetic acid/water, 1:10:289 v/v), pH 3.6. The electrophoresis was carried out in a Savant electrophoresis tank at 2000 V, 150 mA, 5 ':'C for 75 rain. The paper was dried and the peptides were located by means of marker strips sprayed with ninhydrin. The glycopeptides were quantitatively eluted from the paper and lyophilized. Pronase digestion. The sample to be digested was dissolved m [ mJ 0.1 M Tris. HCI, pH 7.5, containing 0.03 M CaCI,, To this solution was added pronasc from Streptomyces griseus (Sigma Chemical Co., St. Louis, Mo.), such that the ratio of pronase:sample was 1:1 (w/w). The incubation was carried out at room temperature for 72 h in an atmosphere of toluene to prevent bacterial growth. Reduction and carboxymethylation. The sample was dissolved in I ml 0.1 M NH4HCOa, pH 8.0. Reduction and carboxymethylation was carried out as described by Shih et al. [5]. E&nan degradation. Semi micro-degradation was carried out according to the procedure of Peterson et al. [15]. Reagents and solvents were purchased from Pierce Chemical Co. The degradation was carried out using the manual method in an atmosphere of N> Gas-liquid chromatography was employed t\)r identification or" the phenylthiocarbamoyl amino acids. The support system was 10'},~ SP-400 Chromosorb W-H-P, 100/120 mesh packed into a 180 cm :, 2 m m glass column. The carrier gas was N2 and detection was by flame ionisation. The temperature program was: initial temperature 185 ' C for 4 min and a rise of 4 'C/rain to a maximum temperature of 285 'C. The Ptc amino acids were sitylated in the injection port with N,O-bis(trimethylsilyl)-acetamide. With the exception of glycine and threonine all the Pie amino acids could be satisfactorily resolved. Glycine and threonine were separated by thin-layer chromatography using plates of Eastman 6060 silica gel. The plates were
431 developed using 2-propanol/xylene (10:35, v/v) and the Ptc amino acid visualised under ultraviolet light and the Re compared with authentic Ptc amino acids run simultaneously. Amino acid and glucosamine analysis. The samples for amino acid analysis were sealed under Nz and hydrolyzed in 5.7 M HCI at 110 °C for 20 h. After hydrolysis the HCI was removed by lyophilisation. Amino acid analysis was performed on a Beckman 120B amino acid analyser adapted to extend the sensitivity range to 2 nmol. The site of carbohydrate attachment was determined by analysing for glucosamine those samples of Ptc amino acids which could provide sites for carbohydrate attachment, i.e. aspartic acid, asparagine, threonine and serine. The samples for glucosamine analysis were sealed under Nz, hydrolyzed in 4 M HC1 at 100 °C for 4 h and analysed using the amino acid analyser. Carbohydrate analysis. Quantitative analysis was achieved by gas-liquid chromatography of the alditol acetate derivatives of the carbohydrate. The method of Kim et al. [16] was used for the preparation of the alditol acetate derivatives. Gasliquid chromatography was carried out as described by Partridge et al. [17]. A modification of the Molisch test as described by Dubois et al. [18] was employed to monitor carbohydrate in the effluent from chromatography columns. To an aliquot was added 0.5 ml 5% phenol and 3 ml concentrated H2SO4. The solution was vigorously agitated and allowed to stand for 15 min before reading the absorbance at 490 nm. Mannose was employed as the standard for sugar estimation. RESULTS AND DISCUSSION
Purification of peroxidase isoenzyme C Separation of peroxidase isoenzyme C from B is difficult due to the close similarity of their properties [1-8]. In the present paper an improved separation was achieved using a modified elution gradient from the CM-cellulose column (Fig. 1). B
B/C
08 O075
o.o65
0.6
o.o 5 0.2
~:
I
~, 0
667
687
, 707
, 727
747
"
]0.045
767
Fraction number
Fig. 1. Elution profile of peroxidase isoenzymes B and C from a CM-cellulose chromatographic column. The column, 40 cm × 2.5 cm, was equilibrated with 0.005 M sodium acetate, pH 4.5. A gentle elution gradient was developed using the Varigrad, Buchler Instruments Co., Fort Lee, N.J. The first chamber contained 500 ml 0.01 M sodium acetate, pH 4.5; the middle chamber contained 500 ml 0.06 M sodium acetate, pH 4.5; the terminal chamber contained 500 ml 0.1 M sodium acetate, pH 4.5. Flow rate was 10 ml/h and 1.5-ml fractions were collected and monitored for peroxidase by measuring the absorbance at 402 nm. Molarity of the effluent solution was determined using a conductivity meter.
432 pH 8 8
1
10
I
Q8
r
=e 0.6
I
'l
,, ,i
04
/ ~,~
),
!
/
Q2 F
/
0
20
pH 72
'
i
~
40
60
80
Fraction
100
~, ° 120
number
Fig. 2. Elution profile of peroxJdase isoenzyme C from an isoelectric focusing column.
lsoenzyme C obtained from the CM-cellulose column was further purified using isoelectric focusing. The elution profile alter 8 days, when no further migration of the peroxidase visible bands was observed, is shown in Fig. 2. The isoelectric point of the major peak corresponds with that reported by Welinder et al. [9] and is consistent with the high lysine and arginine content of isoenzyme C and with the absence of sialic acid [1]. Subsequent work was done on protein collected from the major peak of the isoelectric t\)cusing column. Criteria
of homogeneity
The homogeneity of peroxidase isoenzyme C was demonstrated by electrophoresis on polyacrylamide gels at pH 4.5 and 10.6 and on cellulose acetate strips at pH 5.6 and 8.3 (Fig. 3). Approx. 60 fig of peroxidase isoenzyme C was applied to each polyacrylamide gel. Electrophoresis at pH 4.5 or 10.6 showed one protein band when stained with Coomassie brilliant blue and one activity band when stained for peroxidase. The protein band and peroxidase activity band were coincident. Cellulose A
Top
B
~
m
m
B C
- -
DIM 4.5
7
DIM 10,6
.
DH 5.4
B C
Origin
÷
pH 8 3
Fig. 3. (A) Acrylamide gel electrophoresis of peroxidase isoenzyme C at pH 4.5 and 10.6. (B) Cellulose acetate electrophoresis of peroxidase isoenzyme B and C at pH 5.4 and 8.3.
433 acetate strip electrophoresis at pH 5.4 and 8.3 showed peroxidase isoenzyme C was devoid of the closely related isoenzyme B, as well as other anodic and cathodic peroxidase isoenzymes. The above observations indicate that peroxidase isoenzyme C used in these studies is highly purified and appears to be devoid of other peroxidase isoenzymes as well as contaminating proteins. The R.Z. (ratio of haem absorbance at 402 nm: protein absorbance at 280 nm) of this preparation was 3.5. The carbohydrate composition of horseradish peroxidase C used in these studies is listed in Table V.
Isolation of glycopeptides In the present study trypsin was employed to obtain the glycopeptides because of two advantages: the problem of microheterogeneity of the glycopeptides with respect to their amino acid composition is avoided since trypsin is specific in hydrolyzing peptide bonds on the carboxyl side of lysine or arginine; as each glycopeptide should contain only one residue of arginine or lysine, these residues can act as internal standards to correct for preparative losses. Concurrent work in this laboratory showed that concanavalin A, a plant lectin from jack bean, would precipitate peroxidase isoenzyme C from solution. Peroxidase could be solubilized from the precipitate by the addition of 0.01 M methyl-a-mannopyranoside which competes with peroxidase for the carbohydrate binding sites on concanavalin A. These observations suggested that concanavalin A might be useful to exclusively separate glycopeptides from non-glycopeptides. To test this hypothesis a tryptic digest of isoenzyme C was incubated with concanavalin A and then subjected to ultrafiltration using a PM-10 membrane which retains material with molecular weight above 10 000. Gas-liquid chromatographic analysis of the alditol acetate derivatives of sugar residues in material retained by the PM-10 membrane showed approx. 80 ~ of the sugar initially present on the native enzyme was retained by the membrane (Table I). When concanavalin A was omitted from the mixture all the carbohydrate passed through the membrane. These observations indicate that the glycopeptides are bound to concanavalin A and can be effectively separated from peptides which, having molecular weights of less than 10 000, pass through the memTABLEI S U G A R ANALYSIS OF STARTING MATERIAL, CONCANAVALIN A GLYCOPEPTIDE RETAINED BY PM-10 MEMBRANE, AND OF GLYCOPEPTIDE WHICH PASSED T H R O U G H PM-10 M E M B R A N E AFTER TREATMENT OF RETAINED C O N C A N A V A L I N A GLYCOPEPTIDE LIGAND WITH 0.01 M METHYL-ct-MANNOPYRANOSIDE 1 mg horseradish peroxidase C was employed. The values represent the/~g carbohydrate present in the different fractions.
Starting material* PM- 10 retained PM-10 Pass after methyl-a-mannopyranoslde
Fucose
Arabinose
Xylose
Mannose
32 -- **
4 3
29 23
130 109
29
2
25
110
* Peroxidase isoenzyme C before tryptic digestion. ** Not determined as the fucose peak was obscured by an unidentified peak from concanavalin A.
434
0 o o
0 o
0 0
o 0
0
0
0
04 0
0 0
03 02
0 0 0
01
7o
Gp
A
i
B
o 0 o
~-C
Fig. 4. Paper electrophoresis of a tryptic digest of peroxidase isoenzyrne C (A), of the tryptic glycopeptides (B), and of the tryptic digest after removal of the glycopeptides (C),
brane. Having established that the glycopeptides bind to concanavalin A, an experiment was designed to determine the extent of release of the glycopeptides by methyl~-mannopyranoside. To the material retained by the membrane was added 0.01 M methyl-ct-mannopyranoside. Following ultrafiltration, the material passing through the PM-10 membrane was analyzed for carbohydrate. The results, shown in Table l, indicated that methyl-rz-mannopyranoside quantitatively released the glycopeptides from concanavalin A. Furtherlnore, over 85 o{, of the carbohydrate in native peroxidase isoenzyme C was recovered in the isolated tryptic glycopeptides. The tryptic glycopeptides were fractionated using electrophoresis. The complete tryptic digest and peptides remaining after removal of the glycopeptides were also electrophoresed for comparison. The results are shown in Fig. 4. From the electrophoretic patterns it was concluded that not only were the glycopeptides ,~:eparated from each other, but also that the glycopeptides were effectively separated front the non-carbohydrate-containing peptides. The purity of each glycopeptide was determined by N-terminal analysis and Edman degradation.
Characterization o! the glycopeptides Amino acid analysis of the tryptic glycopeptides (data not presented) showed glycopeptides Nos. 1-5 were large and contained at least 20 amino acid residues per residue of arginine. Tryptic glycopeptide No. 6 contained only five amino acid residues per arginine residue. In order to obtain peptides of a size suitable to manual
435
Q9 Q8 Q7 E c
g
q
Q6 0.5 Q4 03 Q2 0.1 O
~
~'
40
60
80
Effluent
100
120
(ml)
Fig. 5. Elution profile of pronase glycopeptides from a Sephadex 0-25 column. Elution from a 1.5 c m x 100 cm Sephadex G-25 column equilibrated with 0.2 M acetic acid. Fractions of 2 ml were collected and the glycopeptides detected by the Molisch test as modified by Dubois et al. [18] and measured at 490 nm (O---O). Pronase and non glycopeptides, 280 nm (0--(-3).
Edman degradation, the tryptic glycopeptides, with the exception of No. 6, were separately digested with pronase. The pronase glycopeptides were separated from pronase and low molecular weight material on a previously calibrated Sephadex G-25 column. The elution pattern is shown in Fig. 5. No glycopeptides were detected when TABLE II AMINO ACID COMPOSITION OF PRONASE GLYCOPEPTIDES 1, 2, 3, 3A, 3B, 4, 4R, 5 AND TRYPTIC GLYCOPEPTIDE 6 Values are normalised to that amino acid marked 1 or 2 which was shown to be present as one or two residues in the tryptic glycopeptide from which the pronase glycopeptide was derived. The results are expressed as mol amino acid/mol glycopeptide. Peptide
Asp Thr Ser Pro Gly Ala Val Cys Leu Ile Phe Arg
1
2
3*
3A
3B
4**
4R
5
6
2.0 0.7 0.8 0.8 1.0 1 0.0 0.0 0.0 0.9 0.0 0.0
1.9 0.8 0.7 0.0 1 0.0 0.0 0.0 0.0 0.0 1.1 0.0
3.1 2.6 1.1 0.7 0.9 1 0.0 0.0 1.8 0.0 0.0 0.0
2.1 0.7 0.0 0.0 0.0 0.0 0.0 0.0 0.9 0.0 0.0 0.0
0.9 1.9 0.9 1.1 0.9 1 0.0 0.0 0.8 0.0 0.0 0.0
2.0 0.0 1.9 2.1 0.0 0.8 0.0 2 0.0 0.0 0.0 1.1
2.0 0.0 1.1 0.9 0.0 0.0 0.0 1 0,0 0.0 0.0 0.0
1.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1 0.0 0.0 0.0
1.8 0.0 0.0 0.0 1.0 0.0 0.9 0.0 1.0 0.0 0.0 1
* Glycopeptide 3 was resolved into two glycopeptides, 3A and 3B, by repeating the electrophoresis step. ** Since pronase glycopeptide 4 contained half cystine, it was reduced and carboxymethylated and reapplied to the Sephadex G-25 column. The resulting glycopeptide was designated 4R.
436 pronase was incubated in the absence of tryptic glycopeptides, indicating that pronase did not contribute to the glycopeptide pool. Pronase treatment yielded glycopeptides of a size suitable to the manual Edman degradation as indicated by, tile amino acid analysis of the glycopeptides shown in Table II. All glycopeptides contained amino acids commonly associated with the N-glycosidic carbohydrate protein linkage, i.e. aspartic acid, threonine and serine. As glycopeptide No. 4 contained half cystine it was reduced and carboxymethylated and reapplied to tile Sephadex G-25 column. Pronase glycopeptide No. 4 after reduction and carboxymethylation, is designated 4R. Gas-liquid chromatography was chosen to identify the Ptc amino acids because the sensitivity of the method permitted quantities as small as I nmol to be detected. All the Ptc amino acids were adequately resolved with tile exception of threonine and glycine which could be separated by thin-layer chromatography. Upon gas-liquid chromatographic analysis threonine and serine usually appeared as two peaks, ,~serine (phenylthiocarbamoyl of
