preparation which is available contains an oligo-1,4-0 1,4-glucantransferase

THE PROPERTIES OF AN OLIGO-1,4 -- 1,4-GLUCANTRANSFERASE FROM ANIMAL. TISSUES* BY DAVID H. BROWN AND BARBARA ILLINGWORTH DEPARTMENT OF BIOLOGICAL CHEMISTRY, WASHINGTON UNIVERSITY SCHOOL OF MEDICINE, SAINT LOUIS, MISSOURI Communicated by Carl F. Cori, July 16, 1962 The limit dextrin produced by exhaustive treatment of glycogen or amylopectin with phosphorylase and inorganic phosphate was assumed by Cori and Larner to be asymmetric with respect to the number of glucose units remaining on the main and side branches.' The fact that amylo-1,6-glucosidase preparations acted without delay on such a limit dextrin to produce free glucose (see Fig. 2a of previous paper2) was taken to mean that the outer 1,6-linked branch point glucose residues had been exposed by prior phosphorylase degradation. Larner and Schliselfeld found that the rate curve for glucose liberation by glucosidase was predictable on the basis of the Km for the limit dextrin and the Ki for competitive product inhibition.3 Walker and Whelan have studied the structure of such a limit dextrin by other enzymatic means and have concluded that the arrangement of glucose residues about an outer branch point is symmetric and that four glucose units remain on each chain.4 In such a structure there are no exposed branch point glucose residues. To account for the action of amylo-1,6-glucosidase on a limit dextrin these authors have proposed that the enzyme preparation is contaminated by a transglycosylase whose action in exposing a branch point unit is required before the glucosidase can act. They found, in fact, that a partially purified preparation of the glucosidase of rabbit muscle acted on maltose and maltotriaose in such a way that it appeared to contain a transglycosylase of the type described by Giri et al.5 and by Stetten.6 This activity expressed itself as the transfer of one glucose residue in a-1,4 linkage from a donor oligosaccharide to a suitable acceptor. On the other hand, Illingworth et al. found that after amylo-1,6-glucosidase is purified more extensively by column chromatography, it has no action whatsoever on maltose or maltotriaose.7 Recently Verhue and Hers concluded that the symmetric structure of Whelan coexists with the asymmetric structure of Cori in the limit dextrin molecule, possibly because some branches are inaccessible to phosphorylase action.8 We have now been able to show that the most highly purified amylo-1,6-glucosidase preparation which is available contains an oligo-1,4-0 1,4-glucantransferase which has no action on maltose or maltotriaose, but which is able to move, preferentially, a chain of three maltosidically linked glucose residues, and, to a lesser extent, two such residues from a donor to an acceptor oligosaccharide or even to glucose itself. It has no significant action in transferring a single glucose unit. Thus, this enzyme has some of the properties found by Whelan and co-workers for the D-enzyme of potato,9 as well as some of those reported by Petrova for a transglycosylase from rabbit liver. 10 Because of its preference for the movement of either three or two glucose residues, the animal enzyme appears to have a sharper specificity than the D-enzyme of plants. The presence of this oligotransferase in amylo-1,6-glucosidase preparations does not itself prove that it is a required enzyme in glycogen degradation. This ques- 1783

1784 BIOCHEMISTRY: BROWN AND ILLINGWORTH PROC. N. A. S. tion can be settled only when glucosidase preparations are available without oligotransferase and vice versa. A study of the enzymatic constitution of tissues from cases of glycogen storage disease could also be decisive. Neither does the existence of the oligotransferase of itself make possible an answer to the problem of the structure of a phosphorylase limit dextrin. Despite the evidence of Hers8 that there are exposed glucose branch points in the limit dextrin, the possibility must be considered that phosphorylase preparations used to prepare limit dextrins may contain the oligotransferase whose presence would ordinarily escape detection in the absence of amylo-1,6-glucosidase which can be effectively removed from phosphorylase by recrystallization. A definitive answer to this problem requires a separate test for transferase activity. Materials and Methods.-The preparation of amylo-1,6-glucosidase has been outlined.2 The low molecular weight branched dextrins used in this work were prepared either from glycogen or from starch-u-c'4 (500,000 cpm per Amole "glucose") to which corn amylopectin was added as a diluent before a-amylase digestion.2 All reaction mixtures were deionized on Amberlite MB-3 resin columns before descending chromatography on Whatman No. 1 paper in a butanol-pyridinewater solvent (3:2:1.5). In the case of experiments involving isotopically labeled substrates, the paper chromatograms were examined for radioactive zones in a Vanguard Automatic Scanner equipped with an Automatic Data System so that the relative amount of C'4 in the different compounds could also be determined. In some cases, the radioactive substances were eluted with water so that their specific activities could be determined. Total glucose was measured microenzymatically after hydrolysis at 1000 in 1 N HCl for 3 hr. Radioactivity was measured with either a windowless gas-flow counter or a low background counter (Nuclear). Results.-The nature of the reaction catalyzed by the oligotransferase was shown in an experiment in which 4.14 Mmoles of maltotriaose-u-c14 (sp. act., 23,850 cpm/amole) was incubated in 1.0 ml of 0.003 M Na citrate and 0.01 M glycylglycine and 0.001 M mercaptoethanol buffer, ph 6.5, with 20 mg of human muscle glycogen and 25,ug of a rabbit muscle protein fraction (in which amylo-1,6-glucosidase had a sp. act. of 66,000 units/mg). Incubation was for 22 hr at 30 under toluene vapor. A control was incubated without added glycogen. The glycogen was isolated from the first reaction mixture by alcohol precipitation and purified by alkali-digestion and repeated reprecipitations. It was found to be unlabeled (2.36 mg had 14.1 cpm; background, 14.2 cpm). Thus, there was no transfer of any portion of the maltotriaose molecule to glycogen. Chromatography of the reaction mixture revealed that although it contained only a trace of radioactive glucose (present also in the control incubation), there was a prominent radioactive peak of maltohexaose and a lesser peak of maltopentaose. In contrast, maltotetraose was present as a barely detectable zone. Sufficient maltohexaose had been formed to allow its detection by a benzidine-trichloroacetic acid spray. The control incubation mixture contained no trace of any radioactive compound other than the maltotriaose which had been added. Scanning the experimental chromatogram showed that a total of 1,650 cpm of maltohexaose and 421 cpm of maltopentaose had been formed in the reaction mixture. The maltohexaose was eluted and found to have a sp. act. of 20,470 cpm/amole. The fact that this is 86% of that of the added maltotriaose shows that the molecule of maltohexaose consists of three labeled and three unlabeled glucose residues, the latter derived from glycogen. That the oligotransferase prefers to move three units is shown by the fact that maltopentaose was formed (by the movement of two units) in only 25%7 of the yield of the hexaose.

Voi.. 48, 1962 BIOCHEMISTRY: BROWN AND ILLINGWORTH 1785 That the enzyme has little or no ability to move one glucose unit is shown by the fact that maltotetraose was scarcely detectable in the reaction mixture. It should be emphasized that the oligotransferase reaction is slow and incomplete when studied in a system such as that just described. Thus, only about 4%N of the maltotriaose was converted to the two higher oligosaccharides. Maltotriaose is not the most favorable acceptor in such a system. When the singly branched compound containing five glucose residues which has been found to be acted upon directly by amylo-1,6-glucosidase2 was used instead of maltotriaose in a reaction mixture containing glycogen and the oligotransferase, 21/2 hr of incubation served to produce approximately as much of the transfer products (a branched 8-unit dextrin and a branched 7-unit dextrin) as was observed in 22 hr in the triaose experiment already described. Thus, the oligotransferase acts more readily when 90 branched dextrins are available as ac- > 80 ceptors. However, it has been found E that the branched 4-unit dextrin does m0 Fast 8 not act as an acceptor at all under con-60- ditions where the branched 5-unit com- / pound which is a glucosidase substrate v 50 is effective. Thus, the oligotransferase X appears to favor those branched oligo- a saccharides as acceptors in which at m 30 least one glucose residue is on the main 20- chain beyond the branch point, sic / Glucose If6 0 Gluctose1-4 Glucose1-4 Glucose1---- TIME OF INCUBATION (HOURS) The availability of a B7 dextrin has FIG. 1.-Rates of formation of glucose from made possible a demonstration of the Fast B1 by direct amylo-1,6-glucosidase action ability of the oligotransferase to and from move B7 by the action of the glucosidase following the action of oligotransferase. B.;, 8.1 X a maltesyl unit but not a glucosyl unit 10-i AI; B7, 8.7 X 10-4 Al. Enzyme, 29.6,4g intramolecularly as well as intermolec- mg; protein/ml; 0.01 Al Na glucosidase citrate and sp. 0.02 act., Al50,400 glycylglvcin units/- ularly. In Figure 1 are shown data and 0.001 A1 mercaptoethanol buffer, ph 6.6; on the combined action of amylo-1,6-300. glucosidase and oligotransferase in producing glucose from an oligosaccharide designated B7 (see Fig. 1 of previous paper2) which contains a single branch point glucose residue covered by two glucose units. Inasmuch as there is no exposed branch point glucose residue in this structure an action of the glucosidase on this compound (B7) would not be expected. As shown, the rate of glucose formation is about 10 times slower from B7 than from the true glucosidase substrate, Fast B5, which lacks the maltose residue covering its branch point glucose. That the slow liberation of glucose is due to the fact that the oligotransferase must first act to expose the branch point unit was shown by the following experiments in which transfer of a maltosyl residue was detected. The enzyme preparation (13 sag of protein) was added to 0.60 ml containing 1.74 Mmoles of B7, 0.01 Ml Na citrate buffer, ph 6.7, and 4.1 4moles of either maltose-

1786 BIOCHEMISTRY: BROWN AND ILLINGWORTH PROC. N. A. S. UC14 (14,760 cpm/,gmole) or maltotriaose-u-c'4 (23,850 cpm//amole). Incubation was for 42 hr at 300 under toluene vapor. At the end of this time, 0.91 mole of glucose per mole of B7 had been formed in the reaction mixture containing added maltose, and 0.82 mole of glucose per mole of B7 in that containing added maltotriaose. Control incubations containing no added B7 had less than 1% of this amount of free glucose. The reaction mixtures were subjected to paper chromatography and the compounds present were detected by spraying with benzidinetrichloroacetic acid, and by scanning for radioactivity. In nearly every instance the quantity of each substance formed as well as its specific activity were determined. Table 1 gives the data from the experiment with added maltose-u-c'4 TABLE 1 SUBSTANCES FORMED FROM 1.74,UMOLES OF B7 BY OLIGOTRANSFERASE AND AMYLO-1,6-GLUCOSIDASE ACTION IN PRESENCE OF 4.1 limoles OF MALTOSE-U-C14 Detected by Amount isolated* Specific activity Substance formed spray (Mmoles) (cpm/umole) Glucose + 1.57 60 Maltotetraose + 0.52 2,740 Maltohexaose + 0.56 136 B7 + 0.28 about 0 Maltooctaose + 0.11 56 * Except for glucose which was determined directly, all other substances were analyzed after elution from paper chromatograms. and Table 2 the data from that with added maltotriaose-u-c.14 In both of these experiments as well as in others (not shown) in which no maltose or maltotriaose were added, maltotetraose and maltohexaose were formed from B7 in approximately equal amounts. The hexaose results from intramolecular movement of a maltosyl residue from the side chain to the main chain followed by glucosidase splitting of the TABLE 2 SUBSTANCES FORMED FROM 1.74 1AMOLES OF B7 BY OLIGOTRANSFERASE AND AMYLO-1,6-GLUCOSIDASE ACTION IN PRESENCE OF 4.1 1AMOLES OF MALTOTRIAOSE-U-C'4 Detected by Amount isolated* Specific activity Substance formed spray (,moles) (cpm/pmole) Glucose + 1.42 37 Maltotetraose + 0.31 89 Maltopentaose + 0.29 14,575 Maltohexaose + 0.33 200 Maltoheptaose - 0.06 4,840 B7 + 0.24 about 0 * See note to Table 1. rearranged B7 molecule. The tetraose results from intermolecular movement of a maltosyl residue from one B7 molecule to another. The products, B5 and B9, are then acted upon by glucosidase to give the tetraose and the octaose respectively. When C14-maltotriaose was added (Table 2), it acted as an acceptor of most of the maltosyl residues which escaped intramolecular transfer. Thus, maltopentaose of high specific activity was formed in an amount equal to that of the essentially unlabeled tetraose originating from B7 itself. Because of the rather high concentration of added C14-triaose, it served in this experiment as a more effective acceptor of the maltosyl residues than B7 present at lower concentration. Hence, no detectable B9 and maltooctaose were formed. Table 1 shows the formation of labeled tetraose (18% of the total), occurring because of the transfer of maltosyl resi-

VOL. 48, 1962 BIOCHEMISTRY: BROWN AND ILLINGWORTH 1787 dues to C'4-maltose. The amount thus formed (calculated: 0.09 Mmole) taken with the amount of octaose (0.11 jsmole) is an approximate measure of the total intermolecular transfer of maltosyl units. The sum, 0.20 jsmole, compares well with the quantity of labeled pentaose (0.29 Mmole) produced in the experiment of Table 2, whose formation measures the same kind of reaction. In other experiments which are not described, oligotransferase action on B7 has been allowed to occur in the presence of C14-glucose. Under these circumstances, C14-amylotriaose was formed demonstrating that the monosaccharide can act as an acceptor. Manners has suggested on the basis of glycogen structure that a glycosyltransferase may be missing in certain subtypes of Type III glycogen storage disease.11 The information which has been obtained on the action of the oligotransferase is being used to devise a specific assay for the enzyme in human tissue as well as to facilitate its further purification from muscle extracts. Elucidation of its role in glycogen metabolism must await the success of these experiments. Summary.-An oligo-1,4 -- 1,4-glucantransferase has been discovered in rabbit muscle as a contaminant of amylo-1,6-glucosidase preparations. The transferase acts to move preferentially maltotriosyl and, to a lesser extent, maltosyl residues from donor oligosaccharides, including glycogen, to acceptor linear and branched oligosaccharides. The enzyme has no detectable action in moving single glucose residues. With a branched 7-unit dextrin of suitable structure, it has been possible to observe intramolecular transfer from the side branch to the main chain as wvell as intermolecular transfer of a maltosyl residue. The authors wish to thank Sue Hettel for expert assistance in the chromatographic procedures involved in this work and Juanita Buster for making available the amylo-1,6-glueosidase preparations. * This work has been supported in part by a research grant (RG 4761) from the U.S. Public Health Service. 1 Cori, G. T., and J. Larner, J. Biol. Chem., 188, 17 (1951). 2 Illingworth, B., and D. H. Brown, these PROCEEDINGS, 48, 1619 (1962). 3 Larner, J., and L. H. Schliselfeld, Biochim. et Biophys. Acta, 20, 53 (1956). 4 Walker, G. J., and W. J. Whelan, Biochem. J., 76, 264 (1960). 5 Giri, K. V., A. Nagabhushanam, V. N. Nigam, and B. Belavadi, Science, 121, 898 (1955). 6 Stetten, M., J. Am. Chem. Soc., 81, 1437 (1959). 7 Illingworth, B., D. H. Brown, and C. F. Cori, Fed. Proc., 20, 86 (1961). 8Verhue, W., and H. G. Hers, Arch. Int. Physiol. et Biochim., 69, 757 (1961). 9 Walker, G. J., and W. J. Whelan, Biochem. J., 67, 548 (1957). 10 Petrova, A. N., Enzymologia, 21, 23 (1959). 1 Manners, D. J., and A. Wright, Biochem. J., 79, 18-19 P (1961).