Synthesis and structural characterization of a cardioactive biotinylated digoxin analogue

Clin Biochem, Vol. 24, pp. 469--473, 1991 Printed in the USA. All rights reserved.

0009-9120/91 $3.00 + .00 Copyright © 1991 The Canadian Society of Clinical Chemists.

Synthesis and Structural Characterization of a Cardioactive Biotinylated Digoxin Analogue ANITA NUTIKKA, 1 HENRIANNA PANG, 2 STEVEN SOLDIN, 1 and CLIFFORD LINGWOOD 1,3,4 1Research Institute, Hospital for Sick Children, Toronto, Ontario, Canada, and Departments of 2Medical Genetics, 3Biochemistry and 4Clinical Biochemistry, University of Toronto, Ontario, Canada We have biotinylated the terminal glycose of digoxin by reaction of the periodate-oxidized steroid with biotin hydrazide. A biotinylated product (BD-1) was formed which retained significant digoxin receptor (Na+/K ÷ ATPase) binding activity. Sustained reaction resulted in a second biotinylated product (BD-2) which showed reduced receptor binding activity. The products were characterized by FAB mass spectroscopy and shown to be the mono- and dibiotinylated digoxin conjugates of the oxidized glycose moiety. These analogues may prove useful in determining the subcellular site of digoxin binding.

these derivatives has not been rigorously investigated. Compounds in which the carbohydrate moiety have been removed or in which the lactone ring has been removed or altered no longer retain biological activity. We have developed a simple method for the biotinylation of digoxin with the maintenance of cardioactivity. Materials a n d m e t h o d s

KEY WORDS: cardiac glycoside; periodate oxidation; mass spectroscopy; Na+/K + ATPase-binding. Introduction

igoxin is the major cardioactive glycoside used D in the m a n a g e m e n t of various cardiac arrythmias (1). It is a steroid carbohydrate conjugate isolated from the plant digitalis lanata. Its cardioactivity depends on the presence of an intact lactone ring in the steroid moiety (A2°:22/-3~, 12~,14,21-tetrahydroxynorcholenic acid lactene) and a triose oligosaccharide (2-desoxy-D-altromethylose) at the 3 position of the steroid ring. Sequential removal of the sugar residues results in the progressive loss of bioactivity. The aglycone is essentially inactive (2). Displacement (3), photoaffinity (4) and ion flux (5) experiments have shown t h a t digoxin binds to the Na+/K + ATPase of the plasma membrane and its therapeutic activity is a result of binding to and subsequent inhibition of this enzyme on cardiac muscle cells (6). While several analogues of digoxin have been synthesized to study this interaction (4), the residual cardioactivity of

Correspondence: Dr. Clifford Lingwood, Research Institute, Hospital for Sick Children, 555 University Avenue, Toronto, Ontario M5G lX8, Canada. Current address for Dr. Steven Soldin: Department of Laboratory Medicine, Children's Hospital National Medical Center, 111 Michigan Avenue, Washington, DC 20010. Manuscript received November 22, 1990; revised April 18, 1991; accepted May 31, 1991. CLINICAL BIOCHEMISTRY, VOLUME 24, DECEMBER 1991

BIOTINYLATION

500 ~LL 0.12 mmol/L sodium metaperiodate were added to 20 mmoles of digoxin (Sigma) dissolved in 5 mL methanol with stirring. The reaction was allowed to proceed for 2 h at room temperature. The reaction was monitored on high-performance thin layer chromatography [HFrLC] plates (Brinkmann Instruments, Ontario, Canada) using a methylene dichloride:ethyl acetate:acetone (1:1:1 by vol) solvent system. Carbohydrate containing species were detected by spraying with 3% aqueous chloramine T/25% trichloracetic acid in ethanol (1:4 by vol) (7). The oxidized product (RF 0.4 cf. digoxin 0.15) was dried under nitrogen and extracted using methylene dichloride and water (1:1 by vol). The organic layer was dried down and the oxidized digoxin was dissolved in 2 mL methanol. 40 mmoles biotin hydrazide [Pierce (sparingly soluble)] was added and the reaction was stirred at room temperature. After 2h, a new carbohydrate and lactone ring-containing product was detected in the reaction mixture by HPTLC (biotinylated digoxin-1; BD-1) using acetone: ethanol (1:1) as the separating solvent (RF 0.55 cf. 0.7 for oxidized digoxin). A second product (BD-2) was detected after a 24 h reaction period (RF 0.3). These two products were purified by preparative TLC. RECEPTOR BINDING

Cardiac muscle membranes were prepared as previously described (2). Briefly, membrane ATP469

NUTIKKA, PANG, SOLDIN, AND LINGWOOD

A

1

2

3

:4': :":

Figure 1 -- HPTLC separation of digoxin biotinylation products. Oxidized digoxin was incubated with biotin hydrazide as described in the methods, and aliquots were removed at the time indicated and were separated by HPTLC using acetone:ethanol (1:1, v/v). (A) After 2 h, stained for carbohydrate. Lane 1: oxidized digoxin; lane 2: reaction mixture. (B) A ~ r 24 h, stained for carbohydrate: lanes 1, 2, and stained for lactone ring: lanes 3, 4. Oxidized digoxin: lanes 1, 3; reaction mixture: lanes 2, 4. ase preparations were incubated in receptor buffer (50 mmol/L Tris-HC1, pH 7.4 containing 0.5 mmol/L EDTA, 80 mmol/L NaC1, 4 mmol/L MgSO4) in the presence of increasing concentrations of digoxin or the modified analogues. After equilibrium was established, subsequent binding of tritiated digoxin was measured by further incubation and filtration of the membranes (2).

MASS SPECTROSCOPY

Positive ion FAB mass spectroscopy was carried out on BD-1 and BD-2 using a VG Analytical ZAB-SE mass spectrometer (VG Instruments Manchester, UK) in xeon gas (8 Kv, 1 mA). Samples were dissolved in methanol/thioglycerol (1:1) for analysis.

Results The reaction between periodate oxidized digoxin and excess biotin hydrazide had gone to completion within 2 h since oxidized digoxin was no longer detectable (Figure 1A). On further incubation, how-

470

ever, a second slower-migrating product was detected with a corresponding decrease in the amount of product 1 (Figure 1B). Both these products were purified and tested for their ability to displace 3Hdigoxin from h u m a n cardiac muscle plasma membranes (Figure 2). Both compounds could displace digoxin from its receptor, but the first product (BD-1) was more effective. However, BD-1 was approximately 8-fold less effective a ligand as compared to native digoxin (Figure 2). The structures of BD-1 and BD-2 were determined by FAB-mass spectroscopy. The mass spectra are shown in Figure 3. Periodate oxidation of the terminal glycose moiety of digoxin opens the ring, generating two aldehyde functions. The molecular ions show that BD-1 (1041:M+Na) is the conjugate with one biotin hydrazide molecule, whereas BD-2 (1281:M+Na) is the conjugate with two. A fraction of the underivatized aldehyde group in BD-1 has reacted with the methanol solvent to give the corresponding hemiacetal (1073:M + Na). The structures and peak assignments are shown in Figure 4.

Discussion Digoxin is the most widely prescribed drug for the treatment of cardiac arrythmias in North Amer-

CLINICAL BIOCHEMISTRY, VOLUME 24, DECEMBER 1991

BIOTINYLATED DIGOXINANALOGUE 45(30

A

e

i'I

, dL .... ,11 .

,

.

.

.

'~

3500¸

t,]I I

21 181 197 .

.

.

.

4o7 435

385

-

1

55

1

.,,L

,

,

,

3

......

A

•

.

r~

67a,

.

e "ldit~id'kadn'~'''~ ....

@

,

.

817

•

d iln n n

,

91tl

,

1i

I ]

I

1181 I

2500c~

B 196

r~

.237

loo!

1500

56 loo,

0

20

40

60

80

6 888

1~

lloo

1~

1380

14~

Figure 3 -- Mass spectroscopy of BD-1 (A) and BD-2 (B). nmoles

Figure 2 -- Displacement of 3H-digoxin from cardiac receptor by BD-1 and BD-2. Biotinylated digoxin 1 and 2 were purified by preparative TLC from the 24 h reaction mixture and tested for efficacy in the digoxin receptor assay. Open circles: BD-1; open squares: BD-2; closed circles: native digoxin. Average of duplicates shown.

ica. Although its pharmacokinetics are a matter of some controversy (8), the site of action has been established to be the sodium/potassium ATPase in the plasma membrane of cardiac muscle cells. Increased extracellular potassium concentrations antagonizes the binding (9) and reduce the inotropic effect (10) of cardiac glycosides; decreased intracellular potassium levels potentiate the cardiac effects of digoxin (11). Our goal was to synthesize a biotinylated analogue of digoxin which retained biological activity. Such a derivative might provide a versatile tool to study the mode of action of digoxin, since various streptavidin conjugates are available for different detection systems which would be suitable for cellular and subcellular detection of digoxin binding. Despite the extensive study of the pharmacody-

namics, pharmacokinetics, therapeutic efficacy of digoxin (reviewed in 12), structure-activity relationships (13) and the site of binding to the N a + / K ÷ ATPase (4,14,15), little is known of the cellular location of digoxin binding. Although the intact carbohydrate is necessary for full receptor binding activity (8), removal of the terminal glycose moiety results in only a 50% loss in binding. Therefore, periodate oxidation of the terminal digitoxose (16) was selected as the method for conjugation. The reaction of the digoxin oxidation product with biotin hydrazide gave two reaction products, the first being detected early, and the second at later times during the incubation (Figure 1). These products were shown by mass spectroscopy to be the mono-(BD-1) and di-(BD-2) substituted digoxin derivatives (Figures 3, 4). Both products retained the ability to specifically displace 3H-digoxin in the receptor binding assay (Figure 2). As expected, both derivatives were less effective than native digoxin, with conjugation of the second biotin moiety further reducing binding efficacy, confirming the importance of the carbohydrate in ATPase binding. BD-1 should prove a useful probe in the investigation of the site and mechanism of digoxin action•

CLINICAL BIOCHEMISTRY,VOLUME 24, DECEMBER 1991

471

NUTIKKA, PANG, SOLDIN, AND LINGWOOD o

0

0

HN'~NH 3811 +l,~-H =407

C.H=

JH,

,

CH2

CH.

O~N--N-C

I

' ~

CH,IX

O ~ : : ~ ' 1 . ,~AY ~,.,

,

H

k\

I

'~

k,

o

, I

h

' 519 +Na+ H • M3

I 849+H .IS1 649+NI*H

1

I 3~9 +N$ + H I • 413

515

\

I

]

278 + Na " 301

CH 3

NS +Ns-H = 407

. rn

BD-1

R

,~.0~,1

1

R = C o - c ~ H H~ I OCH 3

hemiacetal

M-1018 MNa,.1041

(rxn with CH3OH solvent) M - 1050 MNa=107 3 885 +Na-H = 907

194s+Na-H I

I I

=9S7

I

O

fl I

I o

I

:,,s

__ ~

' c,,

I

CH 3l

(

I

J

) (CH2)4

I

755 +Na-H

I

""'""'

LJ,.J

r"c"k° O

CH31

""

I

I

I 869 +Na + H = 891

L=~ __ .

\

NH

x----

HN M = 1258 MNa =1 2 81

60g

I

I

I I

I

I (CH2) 4

BD-2

l

o.

-,.

I

\

i

"

H N

O

(I

,//

I

961 *Na+H

OHcH 3

r~,N°÷H

I

= 761

I

NH 0

F i g u r e 4 -- F r a g m e n t a t i o n of BD-1 a n d BD-2. The possible f r a g m e n t a t i o n ions of BD-1 and BD-2 were calculated a n d compared w i t h those obtained in the spectra shown in F i g u r e 3. The ions detected a r e indicated in bold.

472

CLINICAL BIOCHEMISTRY, VOLUME 24, DECEMBER 1991

BIOTINYLATED DIGOXIN ANALOGUE

Acknowledgements

1. Smith TW. Digitalis glycosides, Part i. N Engl J Med 1973; 288: 719-22. 2. Bednarczyk B, Soldin SJ, Gasinska I, D'Costa M, Perrot L. Improved receptor assay for measuring digoxin. Clin Chem 1988; 34: 393-7. 3. Brooker G, Jelliffe R. Serum cardiac glycoside assay based upon displacement of 3H-Ouabain from Na ÷ /K ÷ ATPase. Circulation 1972; 45: 20-36. 4. Deffo T, Fullerton DS, Kihara M, et al. Photoaffinity labeling of the sodium- and potassium-activated adenosinetriphosphatase with a cardiac glycoside containing the photeactive group on the C-17 side chain. Biochemistry 1983; 22: 6303-9. 5. Ozaki H, Nagase H, Urakawa N. Interaction of palytoxin and cardiac glycosides on erythrocyte membrane and Na+/K + ATPase. Eur J Biochem 1985; 152: 475-80. 6. Schwartz A, Whitmer K, Grupp G, Grupp I, Adams R, Lee S. Mechanism of action of digitalis: is the Na+/K + ATPase the pharmacological receptor? Ann N Y Acad Sci 1982; 402: 253-71.

7. Waldi D. Steroids. In: Stahl E, ed. Thin layer chromatography. Pp. 249-79. New York: Springer-Verlag, 1965. 8. Soldin SJ. Digoxin--issues and controversies. Clin Chem 1986; 32: 5--12. 9. Akera T, Brody TM. The role of Na+/K+-ATPase in the inotropic action of digitalis. Pharmacol Rev 1978; 29: 187-220. 10. Prindle K, Skelton C, Epstein S, Marcus F. Influence of extracellular potassium concentration on myocardial uptake and inotropic effect of tritiated digoxin. Circ Res 1971; 28: 337-45. 11. Lown B, Salzberg H, Enselberg CD, Weston RE. Interrelation between potassium metabolism and digitalis toxicity in heart failure. Proc Soc Exp Biol Med 1951; 76: 797--801. 12. Koren G, Soldin S. Cardiac glycosides. Clin Lab Med 1987; 7: 587--605. 13. Gunter T, Linde H. Cardiac glycosides. In: Greef K, ed. Handbook of experimental pharmacology. Pp. 1323. New York: Springer-Verlag, 1981. 14. Rodgers T, Lazdunski M. Photoaffinity labelling of digitalis receptor in the Na÷/K + activated adenosinetriphosphatase. Biochemistry 1979; 18: 135-40. 15. Hall C, Ruoho A. Ouabain-binding-site photoaffinity probes that label both subunits of Na÷/K+-ATPase. Proc Natl Acad Sci USA 1980; 77: 4529-33. 16. Lingwood CA, Soldin SJ, Nutikka A. Synthesis of a fluorescent, cardioactive analogue of digoxin. Anal Lett 1988; 21: 813-26.

CLINICAL BIOCHEMISTRY, VOLUME 24, DECEMBER 1991

473

This work was supported by the Research Development Corporation of the Hospital for Sick Children. Mass spectroscopy was performed at the Carbohydrate Research Centre, University of Toronto.

References