The preparation of tritiated tunicamycin

ANALYTICAL

BIOCHEMISTRY

159,210-216

(1986)

The Preparation RUTHANN

of Tritiated

E. HUNNICUTT

Tunicamycin’

AND ROY W. KEENAN

Department of Biochemistry, University of Texas Health Science Center, 7703 Floyd Curl Drive, San Antonio, Texas 78284 Received May 27, 1986 A relatively simple and inexpensive procedure was devised for the radiolabeling of the glycoprotein biosynthesis inhibitor, tunicamycin. The procedure is based on hydrogen exchange in alkaline solutions of tritiated water. It was noted that the antibiotic was much more alkali labile than model compounds such as uridine. The alkali stability of the inhibitor was studied to determine conditions for optimum labeling and yield. The effects of alkaline incubation on the inhibitory properties of the antibiotic were also investigated and it was found that the breakdown products are not effectiveinhibitors ofthe reaction that transfers N-acetylglucosamine- 1-phosphate to dolichyl phosphate. The isolated radioactive tunicamycin homologs, however, retained all their inhibitory action. Incubation of tunicamycin in the presence of deuterated water and mass spectral analysis showed that under the conditions used for the tritiation of tunicamycin the major product exchanged six hydrogen atoms. The position of the tritium atoms in labeled tunicamycin was not determined. The radioactive label in these compounds was shown to be stable under physiological conditions and should be useful for investigations involving the action of these antibiotics. 0 1986 Academic Pms, Inc. KEY WORDS: tunicamycin;

tritium labeling; glycoprotein biosynthesis; dolichol-P-P-GlcNAc

synthesis.

INTRODUCTION

ichyl phosphate, but much higher concentrations are required to produce this effect (4). Tunicamycin was shown to be a uracil derivative that has a unique 1 l-carbon sugar called tunicamine as part of the structure (5). The tunicamycin homologs differ from each other in the type of amide-linked fatty acid, but all possess similar inhibitory power. The complete chemical synthesis of tunicamycin has been reported recently (6). Despite the large number of studies in which this compound has been used, there is relatively little information on the chemical properties of this material, particularly its stability in alkali. Due to technical problems that were encountered in our work on the preparation of radiolabeled

The antibiotic tunicamycin has been widely employed as a tool in cell physiology, cell biology, and biochemistry to investigate the role of carbohydrates in glycoprotein function (1). Tunicamycin is a potent inhibitor of the glycosylation of proteins that have N-linked oligosaccharide chains, and has been shown to inhibit the enzyme that transfers N-acetylglucosamine- 1-phosphate to dolichyl phosphate (2). Tunicamycin has been shown to inhibit nucleotide sugar transport across the golgi membrane, in addition to its direct action on the IV-acetylglucosamine- 1-I’-transferase (3). In addition, this antibiotic also inhibits the transfer of glucose from UDP-glucose’ to dol’ This investigation was supported by a research grant from the National Institutes of Health (AM 17897-l 1). ’ Abbreviations used: UDP, uridine diphosphate; TM, tunicamycin; [2H]TM, deuterated tunicamycin; Dol-P, dolichyl- 1-phosphate; Dol-P-P-GlcNAc, dolichyl-pyro0003-2697186 $3.00 Copyright 0 1986 by Academic Prers, Inc. All rights of reproduction in any form reserved.

210

phosphoryl-N-acetylglucosamine; UDP-GlcNAc, uridine diphosphate N-acetylglucosamine; Tris-HCl, tris(hydroxymethyl)aminomethane base titrated to the appropriate pH with HCl; DTT:DTE, dithiothreitol:dithioerythritol.

TRITIATION

OF

TUNICAMYCIN

211

tunicamycin, it became necessary to conduct sized as previously described (11). All other some investigations on the alkali stability of solvents and chemicals used were reagent these compounds. grade, except the HPLC solvents that were The principal reason for the present inves- specially purified for this purpose and provided tigation was to devise a procedure for the by Fisher Chemical Company. preparation of radioactive tunicamycin that Preparation of [3H]tunicamycin. The procould be utilized as a tool in studies on gly- cedure that was developed for the labeling and coprotein metabolism. The preparation of tri- isolation of labeled tunicamycin is: An aliquot of a methanolic solution of TM (0.6 pmol) tiated tunicamycin by means of the Wilzbach procedure was reported (7); however, the re- was dried in a 1.5-ml conical screw-top vial under a stream of nitrogen. Fifty microliters sulting material required extensive purification and no data were given to establish the radio- of a 3 M solution of NaOH prepared by dischemical purity of the antibiotic. The Wilzsolving approximately 6 mg of NaOH in the bath procedure is usually only applied to the appropriate amount of 3H20 was added to the labeling of compounds that cannot be labeled vial. The capped vial was incubated in a heatby any other procedure because of the diffi- ing block at 65°C for 48 h. Because of the culty in obtaining radiochemically pure prod- small volume of the reaction mixture it is esucts. sential to ensure that the entire vial is kept at Several approaches were attempted for the the same temperature to prevent changes in labeling of this compound. In our hands, direct the concentration of the solution due to concatalytic hydrogenation, photoreduction (S), densation of the water on cooler parts of the and tritium exchange reactions (9) that have vial. Following incubation, the vial was cooled been successfully exploited for the labeling of and the solution was adjusted to pH 8 by the the pyrimidine moiety of other uracil-conaddition of 250 ~1 of 3.5% acetic acid. taining compounds proved unsuccessful. A The reaction mixture was frozen and the procedure that did show some promise was water was lyophilized using a miniature lythe alkali catalyzed exchange reactions that ophilization apparatus constructed from 13 were used for isotopically labeling pyrimidines X 100 mm screw-capped tubes. The residue (10). In the course of this work, it was found resulting from the lyophilization was susthat tunicamycin was more alkali labile than pended in 200 ~1 of water and the lyophiliis generally assumed. The alkali stability of zation process was repeated an additional two the inhibitor was investigated in order to find times. Tritiated water was almost completely conditions under which tritium exchange removed from the reaction product by this would take place with a minimum loss of the procedure. The white precipitate that retunicamycin. The results of these studies led mained was taken up in 200 ~1 of 50% methto the development of a relatively simple pro- anol and filtered to remove the precipitated cedure for the labeling of tunicamycin. silicates. Pretreatment of the vials with Functional Silane (SCM Specialty Chemicals) prevented the formation of insoluble silicates. The MATERIALS AND METHODS tritiated tunicamycin was separated from the Materials. Tritiated water (1 Ci/g) and highly labeled breakdown products by adding [3H]UDP-GlcNAc (20.4 Ci/mmol) were ob- the filtrate to a small Ci8 reversed-phase coltained from DuPont/New England Nuclear. umn (Waters Sep-Pack). The breakdown Tunicamycin was purchased from Cal- products were eluted with 30 ml of 50% methBiochem-Behring. Deuterated methanol anol and the [3H]TM was eluted with 15 ml 99.5% atom excess and water 99.96% atom methanol:O. 15 M ammonium formate (3: 1, v/ excess were obtained from Aldrich Chemical v) (11). The tunicamycin-containing fraction Company. Dolichyl phosphate was synthe- was evaporated to dryness and the residue was

212

HUNNICUTT

dissolved in a small volume of 50% methanol. The partially purified material was further purified and analyzed by reversed-phase HPLC. Examples of the chromatographic separations obtained before and after the prepurification steps are given in Fig. 1. Under the conditions described above, 20-25% of the initial tunicamycin used was recovered as tritiated tunicamycin. Preparation of [‘Hltunicamycin. Deuterated TM was prepared under the same conditions as given above except that 99.96% D20 was used instead of tritiated water. The product was extracted into butanol, dried in vacua and dissolved in a small volume of 50% methanol. The methanolic solution of deuterated tunicamycin was purified as described above, except that the solution was desalted by passage over a 1 X 30-cm Sephadex G- 10 column in methanol:water (3: 1, v/v). This same solvent was used to separate the tunicamycin by reversed-phase HPLC. The [2H]tunicamycin was collected from the HPLC eluate in two fractions, the first contained tunicamycin homologs of M, 6 830, and the second fraction contained compounds with M, Z 845. The desalting procedures were required to prepare

5

10 15 Minutes

20

5

10 Minutes

15

20

FIG. 1. [3H]Tunicamycin reaction mixture analyzed by reversed-phase (C,,) HPLC and Radioanalytic Flow-l@ in-line radioactivity detecting scintillation counter, before (left) and after (right) purification through a Waters Cl8 Sep-Pack. Ultraviolet absorbance at 254 nm (top) and tritium labeling (bottom) were measured simultaneously. Solvent, MeOH:O. I5 M ammonium formate (3: 1, v/v) at 1 ml/min with scintillation fluid at 2 ml/mitt.

AND KEENAN

samples for analyses by fast atom bombardment-mass spectrometry as described below. Fast atom bombardment-mass spectrometry analysis. The mass spectra of the deuterated TM samples were obtained by Dr. Susan Weintraub using a Finnigan MAT 2 12/INCOS 2200 system equipped with an ion tech saddle field atom gun operating at 8 kV with xenon. The accelerating voltage in the mass spectrometer was 2 kV and the ion source temperature was 70°C. Following application of l-2 ~1of DDT:DTE (5: 1, v/v) to the copper probe tip, a background spectrum was acquired. The probe was then removed from the instrument and the sample was added in 2-3 ~1of methanol that was thoroughly mixed with the DTT:DTE. Sample spectra were evaluated after subtraction of the contributions from the DTT:DTE. RESULTS

Preliminary experiments indicated that tunicamycin was much less stable in alkali treatment than anticipated. For this reason, experiments were carried out in which tunicamycin stability in alkali was compared to that of uridine. The results of one experiment that are given in Fig. 2 show that tunicamycin is degraded more rapidly than uridine under similar conditions. Tunicamycin had a tip of 6 h in 1.O N NaOH at 80°C while uridine had a t,,2 of 40 h under these conditions. The disappearance of 254~nm absorption followed first-order kinetics for the first 5 1 h of the reactions. The first-order rate constants were 2.3 X 10e2 and 7.4 X 10m3 h-’ for tunicamycin and uridine, respectively, under these conditions. Uridine breakdown was followed by measuring the loss of 254~nm absorbance as a function of time, but because of the expense and solubility problems associated with TM, its breakdown was assayed by HPLC under conditions similar to those employed in Fig. 1. The assay measured the rate of disappearance of the principal HPLC peak. It is possible that the pyrimidine ring may in fact be breaking down at the same rate in both compounds,

TRITIATION

L

48

noun

26

213

OF TUNICAMYCIN

51

FIG. 2. Comparison of uridine and tunicamycin stability in basic solution versustime as measured by uv absorbance. Uridine (1.0 rmol) or tunicamycin (0.3 pmol) was incubated in a screw-top vial at 80°C in 1.ON NaOH. Aliquots were taken at selected intervals over a 51-h period and 260~nm absorbance (uridine) or peak height of the major 254~nm absorbing peak in HPLC was measured (TM). Uridine (0) TM (m).

but the HPLC assay measures other changes in the molecule as well as those due to the cleavage of the ring. It was noted that all of the TM homologs broke down at the same rate; i.e., the profile of the HPLC peaks, though smaller, remained unchanged in these experiments indicating that similar reactions were occurring to all TM homologs. As tunicamytin loss increased, there was a corresponding increase in an unidentified peak that ran near the solvent front that may be a deacylated or deacetylated compound with the uracil moiety intact. To determine if the alkaline degradation products of tunicamycin retain their inhibitory properties, the effect of alkali-treated TM on the incorporation of Wacetylglucosamine phosphate into dolichyl pyrophosphoryl Nacetylglucosamine by liver microsomes was measured in the presence of identical quantities of tunicamycin. One sample was heated in the presence of alkali under conditions that led to a 95% loss of tunicamycin, and the control sample was not heat treated. The results of this experiment, given in Fig. 3, show that the degree of inhibition is directly related to the loss of tunicamycin, clearly demonstrating that the intact structure of the antibiotic is required for its inhibitory function. More detailed studies in which the breakdown of tunicamycin and the alkali catalyzed

exchange of tritium into these compounds were measured under different conditions are shown in Fig. 4. These experiments were carried out to gain more information on the alkali stability of tunicamycin and to determine conditions under which the maximum amount of isotope incorporation could be achieved with the smallest loss of the inhibitor. As expected, both the breakdown of tunicamycin and tritium incorporation increased with increasing NaOH concentration and temperature. The reaction conditions that resulted in the best incorporation of tritium with reasonable yields of product were determined to be incubation in 3 N NaOH at 65°C for 48 h. The preparation and purification of [3H]tunicamycin using these conditions are described under Materials and Methods. Inhibitory properties of alkali-treated tunicamycin. To ascertain whether alkali treatment might cause some subtle change in the tunicamycin molecule that would not influence its chromatographic mobility, but would

,

0. 0

023

OS0

00

195

TMIng) FIG. 3. The effect of heating tunicamycin in alkaline solutions on its inhibitory properties. Two identical samples of tunicamycin (20 rmol) were mixed with 50 ~1 of 1.0 N NaOH and maintained at either 4°C (0) or 80°C (+) for 48 h. After neutralization, these samples were diluted, and aliquots were added to incubation mixtures. Each assaytube contained 0.8 mg liver microsomal protein, 9 pg Dol-P, 0.06 nmol [3H]UDP-GlcNAc (20.4 Ci/mmol), 0.15% Triton X- 100,5 mM MgClz ,0.05 M Tris-HCI buffer, pH 7.4, and the indicated quantity of inhibitor in a final volume of 0.4 ml. Following a 15-min incubation at 37°C the reactions were assayedas previously described (12).

214

HUNNICUTT

AND KEENAN

if radioactivity would be lost under conditions similar to those likely to be encountered under 0.51~ 4* D “,i’.O experimental conditions, radioactive tunicamycin was incubated in 0.5 M Tris-HCl buffer, l.O\ 6 pH 8.0, at 37°C and aliquots were removed l.ON NaOH f 2.0 at various times up to 48 h and lyophilized. Less than 0.10% of the radioactivity was recovered in the lyophilisate, demonstrating that the tritium would not be labile under conditions similar to those likely to be encountered physiologically. Location of the isotopic hydrogens in labeled tunicamycin. Deuterium-labeled tunicamycin prepared under the same conditions as those used for tritium labeling was analyzed by fast atom bombardment-mass spectrometry. A I V comparison of the spectra of the normal and 26 40 96 1.0 2.0 deuterated samples revealed that tunicamycin l.ON NoOH 1 had exchanged from four to seven hydrogens for deuterium. The major species were the homologs that had exchanged six hydrogens (‘H6y-tunicamycin. Based on the mass spectra HOUIS of the tunicamycin homologs, it is believed FIG. 4. The alkali catalyzed breakdown and tritium exthat the assignments of deuterium shown in change into tunicamycin. Tunicamycin (0.3 pmol) samples were incubated in screw-capped vials in 50 rl of either 1.O the structure given in Fig. 6 are potential sites An NMR spectrum or 3.0 N NaOH in ‘Hz0 (0.1 Ci/g). Aliquots were removed of tritium incorporation. at intervals up to I 14 h and assayedfor both tunicamycin of deuterated tunicamycin revealed that the (m) and radioactivity (0) as described in Fig. 1. N-acetyl(methy1) hydrogens were exchanged for deuterium (data not shown). “‘/---

l.ON

NaOH

j2”

alter its properties as an inhibitor, radioactive tunicamycin prepared as described above was isolated by HPLC and compared to nontreated tunicamycin for its ability to serve as an inhibitor of UDP-Glc-dolichyl-I?GlcNAc1-P transferase. Figure 5 illustrates that purified, alkali-treated, radiolabeled tunicamycin retains its original inhibitory potency. The data points shown are within experimental error of the curve drawn, due to the difficulties in quantitating TM concentrations in the nanomolar range. Stability of the tritium label on tunicamycin. The studies on the alkali catalyzed exchange of tritium shown in Fig. 4 indicated that there was relatively little isotope exchange under conditions where the base concentration and temperature were relatively low. To determine

1

3

5

7

9

II

13

TM h-19) FIG. 5. A comparison of the inhibitory properties of control and tritiated tunicamycin. Aliquots containing equal quantities (based on quantitative HPLC assays)of [3H]tunicamycin (0) or untreated tunicamycin (0) were added to incubation mixtures similar to those described in Fig. 3 and the effect on the incorporation of radioactive N-acetylglucose-l-phosphate into dolichyl-pyrophosphoryl-N-acetylglucosamine was determined ( 12).

TRITIATION

OF TUNICAMYCIN

624

750

650

950

6?6

750

650

r

950

FIG. 6. (A) Fast atom bombardment-mass spectrum (FAB-ms) of a mixture of tunicamycin homologs as described under Materials and Methods. The molecular weight of the major tunicamycin homolog is 844. (B) FABms of the major tunicamycin homolog after deuterium exchange. Inset: The structure of tunicamycin showing the tentative assignments of deuterium atoms exchanged into the major product.

DISCUSSION

The primary objective of this investigation was to develop a procedure for the radioactive labeling of the glycoprotein biosynthesis inhibitor, tunicamycin. The availability of the labeled antibiotic should make it possible to conduct many experiments that can provide information on the nature of reactions involved in glycoprotein biosynthesis. Initially, attempts were made to label the pyrimidine moiety of the inhibitor by exchange reactions that have been described for many uracil de-

215

rivatives (9). This approach was chosen because of its apparent simplicity and because technical problems were encountered in attempts to catalytically hydrogenate the tunicamycins. Catalytic reduction also alters the structures of the natural antibiotics. Despite the structural similarity between tunicamycin and the other uracil derivatives it was not possible to demonstrate tritium exchange into the antibiotics under a wide variety of conditions that led to the labeling of compounds such as uridine diphosphoglucose. The exchange reaction with uracil derivatives has been the subject of detailed studies (9). The reaction proceeds by addition-elimination across the 5,6-double bond. The effects of substituents on the ribose moiety of uridine were shown to markedly affect the exchange reaction, S-OH increased the rate of exchange of the C-5 hydrogen, presumably by the formation of a 6-membered exocyclic ring. Although tunicamycin has a 5’-OH, it is bound to a chiral carbon and may be fixed in position as opposed to the other uridine derivatives that have asymmetric carbons and free rotation about the 5’-C atom to facilitate the formation of the exocyclic ring. Also the bulky group attached to the 5’-C comprising the rest of the tunicamycin molecule may cause enough steric hindrance to interfere with the exchange reaction. Our experiments clearly demonstrated that the behavior exhibited by the uracil moiety of tunicamycin was not typical of other uracil derivatives that we used as models. An alternate approach for tunicamycin labeling was to carry out the base-catalyzed exchange reaction that was described from the labeling of the 6-position of pyrimidines (10). These studies were not successful and rapid degradation of tunicamycin was observed under the conditions used by these workers. We are unaware of any systematic investigation of the alkali stability of tunicamycin. Tkacz ( 13) stated that “tunicamycin is rapidly degraded when heated in strong acid or exposed to periodate, but there is no loss of activity during a 24 hr treatment with 2 N KOH for 24 hrs at 105 “C,” citing Takatsuki et al. ( 14). The orig-

216

HUNNICU’I-I-

inal article reported that tunicamycin was resistant to alkaline hydrolysis and that antiviral activity was detected even after a 24-h treatment with 2 N KOH at 105°C. Based on our results, it is likely that the antiviral activity that remained was due to a small amount of undegraded tunicamycin. Because of the uncertainties regarding the stability of tunicamycin under basic conditions, a series of experiments was conducted which revealed that tunicamycin is more unstable in alkali than uridine due to the presence of other alkali labile sites in the antibiotic molecule in addition to the pyrimidine ring. The nature of the breakdown products was not determined, but could result from the cleavage of glycosidic linkages or amide bonds, both of which are usually relatively alkali stable. The important point is that conditions can be chosen that result in the labeling of tunicamycin by exchange in tritiated water with acceptable losses of the antibiotic. The experiments conducted with deuterated water showed that there are a number of different sites at which hydrogen exchange could occur in addition to the anticipated incorporation of label into the 6-position of the pyrimidine ring. The greater extent of deuterium incorporation into tunicamycin compared to tritium is due to the greater abundance of deuterium atoms (99.96%) as opposed to only 0.03% ‘H atoms in the tritiated water, thus the exchange in *H20 is much more sensitive and would reveal slowly exchanging hydrogens that might not be detectable by tritium exchange. The observation that the specific activity of the tunicamycin labeled by our procedure was greater than the specific activity of the tritium in the tritiated water used in the experiments can be explained either by assuming that there is a positive isotope effect, or that more than one hydrogen exchanges. From our data, we are unable to determine the site or sites of tritium exchange.

AND KEENAN

The results of the experiments described in this communication provide data on the alkali stability of tunicamycin that could explain ambiguous results obtained with tunicamycin stored in alkaline solutions for long periods of time. Using the procedure for tritium labeling, radioactive tunicamycin can be prepared simply and at low cost. The specific activity of the product is limited only by the specific activity of the tritiated water employed and the radioactive label is stable under physiological conditions. The labeled compound retains its activity as an inhibitor of glycoprotein biosynthesis. Radioactive tunicamycin should provide an excellent tool for the study of certain important reactions in glycoprotein biosynthesis. REFERENCES Tamura, G. (ed.) (1982) Tunicamycin, Japan Scientific Societies Press, Tokyo. Tkacz. J. S., and Lampen, J. 0. (1975) Biochem. Biophys. Res. Cornmun. 43, 761-764. Yusuf, H. K. M.. Pohlentz, G., Schwarzmann, Cl., and Sandhoff, K. (1984) Adv. Exp. Med. Biol. 174, 227-239. 4. Elbein, A. D., Gafford, J. F., and Kang, M. S. (1979) Arch. Biochem. Biophys. 196, 3 I l-3 18. 5. Takatsuki, A., Kawamura, K., Kodama, Y., Teichiro, I., and Tamura, G. (1979) Agric. Biol. Chem. 43, 761-764. 6. Suami, T., Hiroaki, S., Kazuhiro, M., and Suzuki, N. (1985) Curbohydr. Res. 143,85-96. I. Kuo, S.-C., and Lampen, J. 0. (1976) Arch. Biochem. Biophys. 172, 574-581. 8. Cerutti, P., Kondo, Y., Landis, W. R., and Witkop, B. (1968) J. Amer. Chem. Sot. 90,771-715. 9. Santi, D. V., and Brewer, C. F. (1973) Biochemistry 12,24 16-2424. 10. Rabi, J. A., and Fox, J. J. (1978) J. Amer. Chem. Sot. 95, 1628-1632. 11. Keenan. R. W., Martinez, R. A., and Williams, R. F. (1982) J. Biol. Chem. 257, 14817-14820. 12. Keenan, R. W., Hamill, R. L., Occolowitz, J. L., and Elbein, A. D. (198 1) Biochemistry 20,2968-2973. 13. Tkacz, J. S. (1981) Antibiotics VI, p. 3, Springer-Verlag, New York/Berlin. 14. Takatsuki, A., Arima, K., and Tamura, G. (1971) J. Antibiot. 24,2 15-223.